McGraw Hill, Inc. (MH) 2026 Q4 法說會逐字稿

(video playing)

（影片播放中）

Good morning, and welcome to the McGraw Hill, Inc. earnings conference call. (Operator Instructions) As a reminder today's call is being recorded, and a written transcript and webcast replay will be made available in the Events and Presentations section of the company's Investor Relations website. Following the prepared remarks, we will open the call for questions.

早安，歡迎參加 McGraw Hill, Inc. 的財報電話會議。（接線員指示）提醒各位，今日會議將進行錄音，並將於公司投資人關係網站「活動與簡報」專區提供書面逐字稿與網路直播重播。在預先準備的發言結束後，我們將開放提問。

I would now like to turn the call over to your host, Danielle Kloeblen, Treasurer and Senior Vice President, Investor Relations. Please go ahead, Danielle.

現在我想把電話會議交給今天的主持人——Danielle Kloeblen，財務主管暨資深副總裁（投資人關係）。Danielle，請開始。

Good morning. Welcome to McGraw Hill's fiscal fourth-quarter and full-year ended March 31, 2026, earnings call. Joining me today are Simon Allen, Chair of the Board of Directors; Philip Moyer, President and Chief Executive Officer; and Bob Sallmann, Executive Vice President and Chief Financial Officer.

早安。歡迎參加 McGraw Hill 截至 2026 年 3 月 31 日的 2026 財政年度第四季與全年財報電話會議。今天與我一同出席的有：董事會主席 Simon Allen；總裁暨執行長 Philip Moyer；以及執行副總裁暨財務長 Bob Sallmann。

During today's call, we will be making forward-looking statements that are based on our current expectations and the current economic environment. These statements, estimates and projections are subject to significant uncertainties beyond the control of management. As detailed in the cautionary language in our fiscal fourth-quarter and full-year ended March 31, 2026, earnings release, the accompanying investor presentation, our Form 10-K and other SEC filings.

在今天的電話會議中，我們將發表以前瞻性陳述為主的內容，這些陳述係基於我們目前的預期以及當前的經濟環境。這些陳述、估計與預測存在重大不確定性，且超出管理階層可控制的範圍。相關內容已於我們截至 2026 年 3 月 31 日的 2026 財政年度第四季與全年財報新聞稿、隨附的投資人簡報、Form 10-K 以及其他 SEC 申報文件中的警示性語言詳述。

We will also reference certain non-GAAP measures today, which we believe provide useful supplemental insight into our financial and operational performance though they are not a substitute for GAAP measures. Definitions and GAAP reconciliations are available in our earnings release, the appendix to the investor presentation and on our Investor Relations website.

我們今天也會引用若干非 GAAP 指標；我們相信這些指標能就公司的財務與營運表現提供有用的補充洞見，但不能取代 GAAP 指標。相關定義與 GAAP 調節表可於我們的財報新聞稿、投資人簡報附錄以及投資人關係網站查閱。

For those listening to a recording of this call, please note that the remarks are as of today, June 11, 2026, and have not been subsequently updated. With that, I'll turn the call over to our Chair of the Board, Simon Allen.

對於收聽本次電話會議錄音的聽眾，請注意本次發言內容以今日——2026 年 6 月 11 日——為準，之後未再更新。接下來，我把電話會議交給我們的董事會主席 Simon Allen。

Good morning, everyone. Fiscal year 2026 demonstrated the strength of our business, our strategy and the enduring importance of education. Our mission to support learners and educators, build human intelligence and power the future of learning has never been more relevant.

各位早安。2026 財政年度展現了我們業務、策略以及教育長久重要性的實力。我們支持學習者與教育者、培養人類智慧並驅動未來學習的使命，從未像現在這般切合時宜。

In a year marked by rapid AI advancement and global uncertainty, McGraw Hill reinforced our position as the partner educators and learners trust most. That foundation fuels our trajectory. As a data-driven learning company, we made significant strides in AI-powered innovation, launching new tools, expanding existing solutions and deepening key partnerships that position McGraw Hill to lead the next era of AI-enabled learning.

在 AI 快速進展與全球不確定性並存的一年，McGraw Hill 鞏固了我們作為教育者與學習者最信賴夥伴的地位。這個基礎推動了我們的發展軌跡。作為一家數據驅動的學習公司，我們在 AI 驅動的創新上取得重大進展，推出新工具、擴展既有解決方案並深化關鍵合作夥伴關係，使 McGraw Hill 得以引領下一個 AI 賦能學習的新時代。

This execution reflects the strength of our strategy and the caliber of our leadership team, led by the right person. Having worked closely with Philip over the last 120 days, the Board and I have full confidence in his vision and his ability to drive McGraw Hill's next chapter of growth. We are fully aligned and energized by what lies ahead.

這樣的執行成果反映了我們策略的強度與領導團隊的水準，而這一切由最合適的人帶領。在過去 120 天與 Philip 密切合作後，董事會與我對他的願景以及推動 McGraw Hill 下一階段成長的能力充滿信心。我們完全一致，並對未來充滿動能。

I'll now turn the call over to McGraw Hill's President and CEO, Philip Moyer.

接下來我把電話會議交給 McGraw Hill 總裁暨執行長 Philip Moyer。

Thank you, Simon, and good morning, everyone. It's a privilege to step into the CEO role and lead McGraw Hill in this exciting moment in education and technology. My first 120 days have been filled with over 250 customer, team, industry and investor meetings, and it's clear that education is at a seminal moment.

謝謝你，Simon，各位早安。能在教育與科技這個令人振奮的時刻接任執行長並帶領 McGraw Hill，是我的榮幸。我上任的前 120 天，已進行超過 250 場與客戶、團隊、產業與投資人的會議；很明顯，教育正處於關鍵轉折點。

Student populations are growing, the learning landscape is increasingly complex, and outcomes are more important than ever. Students, teachers and institutions are looking for an enterprise partner they trust to ensure the next generation of students are prepared.

學生人口持續成長，學習版圖愈加複雜，而學習成效比以往任何時候都更重要。學生、教師與機構正在尋找一個值得信賴的企業級合作夥伴，以確保下一代學生做好準備。

What is also clear to me is that McGraw Hill is uniquely positioned to be this partner. Our trusted content, our learning data and our pace of AI innovation, along with the scale of our relationships, are unmatched, and I believe we are positioned to be the global leader in this next generation.

同樣清楚的是，McGraw Hill 具備成為這個夥伴的獨特優勢。我們值得信賴的內容、學習數據與 AI 創新速度，再加上我們關係網絡的規模，都是無可匹敵的；我相信我們已具備在下一世代成為全球領導者的條件。

McGraw Hill has a great financial foundation, and my focus for shareholders is to maintain our profit profile, reduce our debt and accelerate our growth in fiscal year 2027 and beyond. I see many opportunities to lead in innovation, streamline and scale our execution and profitably expand our TAM.

McGraw Hill 擁有穩健的財務基礎；對股東而言，我的重點是維持我們的獲利結構、降低負債，並在 2027 財政年度及其後加速成長。我看到許多機會可在創新上領先、精簡並規模化我們的執行力，並以獲利方式擴大我們的 TAM（可服務總市場）。

In fiscal year 2026, we exceeded our IPO expectations for revenue and adjusted EBITDA and achieved several important milestones. Recurring revenue grew by 6% and now represents over 73% of our total revenue. In K-12, our capture rate exceeded expectations despite the expected cyclical decline in the market. In higher education, our net dollar retention reached 114%, and our customer satisfaction reached its highest level ever.

在 2026 財政年度，我們的營收與調整後 EBITDA 均超越 IPO 時的預期，並達成數項重要里程碑。經常性收入成長 6%，目前已占總營收逾 73%。在 K-12 領域，儘管市場如預期出現週期性下滑，我們的得標率仍超出預期。在高等教育領域，我們的淨美元留存率達到 114%，客戶滿意度也創下歷史新高。

In fiscal year 2026, we also proved that we could profitably accelerate execution. We introduced a record number of curriculum offerings and AI learning tools while improving our adjusted EBITDA margin by nearly 80 basis points year-over-year. We now have launched eight AI learning tools, which serve more than 7.5 million users, and we have three additional launches planned this fiscal year.

在 2026 財政年度，我們也證明了能以獲利方式加速執行。我們推出創紀錄數量的課程內容與 AI 學習工具，同時調整後 EBITDA 利潤率年增近 80 個基點。目前我們已推出 8 款 AI 學習工具，服務超過 750 萬名使用者，並計畫在本財政年度再推出 3 項新產品。

We launched our California math curriculum and a new literacy curriculum, the latter of which is our single largest investment in curriculum in our history. We also introduced a K-12 AI curriculum designed to teach AI concepts and integrate them into the classroom learning.

我們推出加州數學課程，以及一套新的識字（讀寫）課程；後者是我們歷史上對課程內容的單筆最大投資。我們也推出一套 K-12 的 AI 課程，旨在教授 AI 概念並將其融入課堂學習。

As we enter fiscal year 2027, we see clear opportunities for growth. The market dynamics support a multiyear acceleration of revenue growth and underpins our confidence in our medium-term framework for mid-single-digit plus revenue growth.

邁入 2027 財政年度，我們看見明確的成長機會。市場動態支持營收成長在未來數年加速，並支撐我們對中期架構——營收成長達中個位數以上——的信心。

Now one of the most exciting areas I see that supports our growth expectations is the use of AI. While I'm frequently asked by investors if AI will disrupt our business, I've rarely been asked that question by our customers. Students, teachers, parents and institutions don't want to park a child in front of a generic LLM and hope for a good outcome.

而我認為最令人振奮、能支撐我們成長預期的領域之一，就是 AI 的運用。投資人常問我 AI 是否會顛覆我們的業務，但我們的客戶很少問我這個問題。學生、教師、家長與機構並不希望把孩子放在一個通用的 LLM（大型語言模型）前，然後期待會有好結果。

They know that AI is only as good as its data and its training. And in my past 120 days, I've deepened my excitement about the competitive moat we are building in AI, trained specifically for education and the outcomes we produce. Our proprietary content catalog includes tens of thousands of courses, spanning more than 500 subjects mapped to hundreds of thousands of regulated learning objectives and academic standards across roughly 12,000 school districts and approximately 3,300 higher ed institutions.

他們知道，AI 的好壞取決於其數據與訓練。在我過去 120 天的工作中，我對我們正在建立的 AI 競爭護城河更加振奮——這套 AI 是專為教育與我們所產生的學習成果而訓練。我們的專有內容目錄涵蓋數以萬計的課程，橫跨 500 多個科目，並對應到數十萬個受規範的學習目標與學術標準，覆蓋約 12,000 個學區與約 3,300 所高等教育機構。

We build our tools to meet digital accessibility requirements, ensuring broad access for learners, and we have our student outcomes measured by independent third parties. We have over 100 million active licenses by students and educators of this curriculum. From this usage, we captured over 25.6 billion learning interactions and have decades of data, over 190 terabytes, on how students learn, where they struggle and how instruction must adapt in real time.

我們打造工具以符合數位無障礙要求，確保學習者能廣泛取得資源，且我們的學生學習成果由獨立第三方衡量。已有超過 1 億個由學生與教育者使用本課程的有效授權。透過這些使用情境，我們累積了超過 256 億次學習互動，並擁有數十年的數據——超過 190 TB——用以了解學生如何學習、在哪裡遇到困難，以及教學如何必須即時調整。

This extraordinary content and data mode is what we use to build and tune our education-specific AI learning tools. It is also what we use to deliver a precision education experience, the right content and the right question at the right moment. I'm excited about our use of AI because I'm seeing how we are expanding this precision experience and creating new opportunities for TAM expansion.

我們正是運用這個非凡的內容與數據護城河，來建構並微調專為教育打造的 AI 學習工具。我們也用它來提供精準教育體驗——在正確的時刻提供正確的內容與正確的問題。我對我們運用 AI 感到興奮，因為我看到我們正在擴大這種精準體驗，並創造擴大 TAM 的新機會。

One great example is the emerging agentic AI landscape. AI is evolving from answering questions to taking actions, but customers need purpose-built agents that they trust. As a result, we are being asked to be a part of new agentic AI strategies in both education and non-education customers.

一個很好的例子是正在興起的代理式（agentic）AI 生態。AI 正在從回答問題演進到採取行動，但客戶需要他們信任、為特定目的打造的代理。因此，我們被邀請參與教育與非教育客戶的全新代理式 AI 策略。

And I'm excited to announce that we are piloting our new agentic AI tool that will make our precision education experience accessible as a trusted AI agent. This represents a new business model and one of several opportunities to expand our TAM.

我也很興奮地宣布，我們正在試點一款新的代理式 AI 工具，將以可信任的 AI 代理形式，讓我們的精準教育體驗更易取得。這代表一種新的商業模式，也是擴大我們 TAM（可服務總市場）的多項機會之一。

At the same time, we continue to enhance our core learning solutions with artificial intelligence. We are expanding our Connect platform with Learning Coach, an AI conversation tutoring tool developed with Kiron Learning. We've added new accessibility features to AI Reader, including translations across 38 languages.

同時，我們持續以人工智慧強化核心學習解決方案。我們正透過 Learning Coach 擴展 Connect 平台；Learning Coach 是與 Kiron Learning 共同開發的 AI 對話式家教工具。我們也為 AI Reader 新增了無障礙功能，包括支援 38 種語言的翻譯。

And we've created new AI literacy modules. As I mentioned earlier, the education market is not just asking for better tools, it is demanding measurable outcomes. In fiscal year 2026, McGraw Hill extended our trust with educators with meaningful improvements in student comprehension and grades, which is now validated by more than 100 independent researchers.

此外，我們也打造了新的 AI 素養模組。如我先前提到，教育市場不只是要求更好的工具，更要求可衡量的成果。在 2026 財年，McGraw Hill 透過在學生理解力與成績上的實質改善，進一步鞏固教育工作者對我們的信任；目前已有超過 100 位獨立研究人員加以驗證。

ALEKS, for example, continue to set the benchmark for adaptive learning with math students achieving topic mastery more than 90% of the time. In Pennsylvania, elementary schools that were using ALEKS Adventure had first grad students improving nearly 25 points on math assessments and second graders improving more than 55 points over the course of the school year.

以 ALEKS 為例，持續以數學領域的自適應學習樹立標竿，學生在超過 90% 的時間能達成主題精熟。在賓夕法尼亞州，使用 ALEKS Adventure 的小學，一年級學生在數學評量上提升近 25 分，二年級學生在整個學年期間提升超過 55 分。

AI Reader generated 57 million learning interactions across 2.4 million students, driving higher engagement with teacher assigned reading topics. And with our Sharpen tool, students engaged up to 4 times more with study activities than compared to generic AI tools. At Rowan College in New Jersey, for example, students who use Sharpen achieved final exam scores 47% higher than peers who didn't use the tool.

AI Reader 在 240 萬名學生中產生了 5,700 萬次學習互動，帶動學生對教師指派閱讀主題的更高參與度。而透過我們的 Sharpen 工具，學生在學習活動上的投入度最高可達一般通用 AI 工具的 4 倍。例如在新澤西州的羅文學院（Rowan College），使用 Sharpen 的學生期末考成績比未使用該工具的同儕高出 47%。

Looking ahead, our strategy focuses on four priorities: strengthening the core, expanding our cross-sell opportunities, growing our addressable market and driving operational efficiencies. In K-12, we're continuing to evolve our California math program to meet the emerging requirements of the market. This California math adoption cycle has been slower than expected, but 80% of the market remains undecided.

展望未來，我們的策略聚焦四項優先事項：強化核心、擴大交叉銷售機會、擴張可觸及市場，以及推動營運效率。在 K-12 領域，我們持續演進加州數學課程，以符合市場新興需求。這一輪加州數學採用週期比預期更慢，但市場仍有 80% 尚未做出決定。

Also, in K-12, we're entering one of the most important nationwide curriculum refresh cycles in this millennium as an unprecedented amount of schools evaluate new English language Arts programs. Since 2021, we have invested over $100 million in our new science of reading ELA program, Emerge, Summit and Soar. This is our largest curriculum investment to date. And we are seeing early district wins and a strong rubric of scores ahead of emerging opportunities.

同樣在 K-12，我們正進入本千年最重要的全國課程更新週期之一，前所未有數量的學校正在評估新的英語語言藝術（ELA）課程。自 2021 年以來，我們已在全新的「閱讀科學」ELA 課程（Emerge、Summit 與 Soar）投入超過 1 億美元。這是我們迄今最大規模的課程投資。我們已看到早期學區的勝出案例，且在評分量表上取得強勁分數，為後續新興機會奠定基礎。

In higher education, we continue to gain share as institutions adopt inclusive access, expand their use of our evergreen model and seek more integrated learning experiences. Fiscal fourth quarter marked our 40th consecutive quarter of market share gains in higher ed, with our share approaching 31% according to NPI. Our Net Promoter Score reached a record in the spring, and we saw strong retention and volume growth. We've also seen solid adoption of our new offerings in credentialing and AI professional development.

在高等教育方面，隨著機構採用包容性存取（inclusive access）、擴大使用我們的常青（evergreen）模式並尋求更整合的學習體驗，我們持續提升市占。2026 財年第四季是我們在高教市場連續第 40 個季度市占提升；根據 NPI，我們的市占率接近 31%。我們的淨推薦值（NPS）在春季創下新高，且留存與量的成長表現強勁。我們也看到在證照（credentialing）與 AI 專業發展方面的新產品獲得穩健採用。

Within our Global Professional business, medical education represents more than 75% of revenue. The World Health Organization projects a shortage of 11 million healthcare workers by 2030. And this year, we expanded our offerings into clinical training and AI-powered diagnostic simulations. We remain well positioned to serve this market with our innovation and trusted, high stakes learning solutions.

在我們的全球專業（Global Professional）業務中，醫學教育占營收超過 75%。世界衛生組織預估到 2030 年將短缺 1,100 萬名醫療照護人員。而今年，我們將產品擴展至臨床訓練與 AI 驅動的診斷模擬。憑藉我們的創新能力與可信任的高風險學習解決方案，我們仍具備良好條件服務此市場。

Internationally, we see education market conditions improving. Higher education enrollment in Canada is stabilizing and new opportunities in favorable demographic markets such as Latin America and the Middle East support a more attractive setup. ALEKS remains a key driver internationally with Calculus now available in additional markets.

在國際市場方面，我們看到教育市場環境正在改善。加拿大的高等教育入學人數趨於穩定，而拉丁美洲與中東等人口結構有利的市場帶來新機會，使整體布局更具吸引力。ALEKS 仍是國際市場的關鍵成長動能，微積分（Calculus）現已在更多市場推出。

Looking to fiscal year 2027, we see a clear path to revenue growth, expanding margins and strong free cash flow. We've made significant progress in reducing our gross debt, and we will remain disciplined in capital allocation, which Bob will speak through shortly. We're looking forward to sharing a broader update on our long-term strategy and capital allocation framework at our Investor Day later this calendar year.

展望 2027 財年，我們看到營收成長、利潤率擴張與強勁自由現金流的清晰路徑。我們在降低總負債方面已取得顯著進展，並將持續在資本配置上保持紀律；Bob 稍後會進一步說明。我們也期待在今年稍晚的投資人日，分享關於長期策略與資本配置框架的更全面更新。

But after this first 120 days as CEO, I am more excited than ever. At McGraw Hill, we build human intelligence. Our mission has never been more important, our assets have never been stronger, and our opportunity has never been greater.

但在擔任 CEO 的前 120 天之後，我比以往任何時候都更感到振奮。在 McGraw Hill，我們打造人類智慧。我們的使命從未如此重要，我們的資產從未如此強大，而我們的機會也從未如此巨大。

With that, I'll turn it over to Bob to review our financial results and provide guidance for fiscal year 2027.

接下來，我把時間交給 Bob，請他回顧我們的財務表現並提供 2027 財年的指引。

Thank you, Philip. Fiscal year 2026 were defined by market share gains, margin expansion and accelerating digital momentum. We exited the year with a stronger balance sheet, greater financial flexibility and stronger pricing power driven by deeper adoption of our AI-enabled solutions, all reinforcing our confidence in long-term value creation.

謝謝你，Philip。2026 財年的特徵是市占提升、利潤率擴張，以及數位動能加速。我們在年末擁有更強健的資產負債表、更高的財務彈性，以及因 AI 賦能解決方案更深度採用而帶來更強的定價能力，這些都進一步強化我們對長期價值創造的信心。

For the fiscal year, revenue was $2.1 billion, above the high end of our guided range and $2 million above prior year. Reoccurring revenue reached $1.5 billion, growing nearly 6% year-over-year, exceeding the high end of our guided range, reinforcing the durability of our model. The remaining performance obligation was $1.7 billion and will increase in the fiscal second quarter of 2027, reflecting the typical K-12 seasonality.

就全年而言，營收為 21 億美元，高於我們指引區間上緣，並較前一年高出 200 萬美元。經常性收入達 15 億美元，年增近 6%，同樣超過指引區間上緣，凸顯我們模式的韌性。剩餘履約義務（RPO）為 17 億美元，並將在 2027 財年第二季增加，反映 K-12 典型的季節性。

Fiscal year adjusted EBITDA was $744 million, up 2% year-over-year, exceeding the high end of our guided range, with margin expanding nearly 80 basis points to 35.4%. During the fiscal year, we reduced gross debt by $646 million, inclusive of the IPO proceeds. We remain focused on achieving our net leverage target of 2 to 2.5 times.

2026 財年調整後 EBITDA 為 7.44 億美元，年增 2%，超過指引區間上緣；利潤率擴張近 80 個基點至 35.4%。在該財年期間，我們將總負債降低 6.46 億美元，其中包含 IPO 募集資金。我們仍專注於達成淨槓桿目標 2 至 2.5 倍。

Turning to the fourth quarter. Revenue was $464 million, down 2% year-over-year, reflecting a smaller K-12 market opportunity, partially offset by continued outperformance in higher education. Recurring revenue totaled $373 million, while digital revenue reached $393 million, representing 81% and 85% of total revenue, respectively.

接著看第四季。營收為 4.64 億美元，年減 2%，反映 K-12 市場機會較小，但部分被高等教育持續優於預期的表現所抵銷。經常性收入合計 3.73 億美元，而數位收入達 3.93 億美元，分別占總營收的 81% 與 85%。

Fourth-quarter gross margin increased nearly 50 basis points, driven by operational efficiency and favorable digital mix. Adjusted EBITDA was $131 million, with margins expanding nearly 40 basis points year-over-year, reflecting operational leverage and early AI efficiency savings while we continue to invest in growth.

第四季毛利率提升近 50 個基點，主要受營運效率提升與有利的數位組合所帶動。調整後 EBITDA 為 1.31 億美元，利潤率年增近 40 個基點，反映營運槓桿與早期 AI 效率節省，同時我們仍持續投資成長。

Turning to segment performance. In Higher Education, fiscal year revenue grew 12% year-over-year to $879 million, with recurring revenue up 10% to $734 million. Growth was largely driven by market share gains, which exceeded expectations, alongside enrollment, pricing and lower sales returns and the associated release, contributing roughly 3 points of growth.

接著看各事業部表現。在高等教育方面，2026 財年營收年增 12% 至 8.79 億美元，經常性收入年增 10% 至 7.34 億美元。成長主要由市占提升所驅動，且優於預期；同時入學人數、定價，以及較低的銷售退貨與相關回沖（release）亦帶來助益，合計約貢獻 3 個百分點的成長。

Market share is now approaching 31% according to MPI. Momentum continued in the fourth quarter with Higher Education revenue increasing nearly 2% year-over-year to $258 million despite more school starting spring semesters later in the quarter compared to last year.

根據 MPI，目前市占率正接近 31%。第四季動能延續，高等教育營收年增近 2% 至 2.58 億美元，儘管相較去年有更多學校在本季較晚開始春季學期。

Our spring selling season was a success with year-over-year activation growth and share gains supported by strong commercial execution and our Evergreen delivery model. Evergreen, a continuously updated content delivery system that improves your attention and frees our sales teams to pursue competitive takeaways.

我們的春季銷售旺季表現成功，在強勁的商業執行與 Evergreen 交付模式的支持下，啟用量年增並帶動市占提升。Evergreen 是一套持續更新的內容交付系統，可提升您的專注度，並讓我們的銷售團隊得以專注於爭取競品客戶的轉換。

This remains a competitive differentiator in the market and now represents 68% of Higher Education revenue. Demand by discipline was particularly strong in business and science, which represents subject areas where we over-index.

這仍是市場上的競爭差異化優勢，目前已占高等教育營收的 68%。按學科別來看，商學與科學領域需求特別強勁，這些也是我們相對更具優勢的科目領域。

Underpinning that demand is the Inclusive Access delivery model. Inclusive Access represents 56% of revenue in fiscal year 2026 and continues to expand as existing campuses adopt additional courses. As a reminder, we typically add about 100 new campuses each year with those significantly scaling over the following two to three years. At the same time, our customer success organization continues to strengthen faculty partnerships, supporting retention and driving net dollar retention of 114% in the fiscal year.

支撐該需求的核心是 Inclusive Access 交付模式。Inclusive Access 在 2026 會計年度占營收的 56%，並隨著既有校園採用更多課程而持續擴張。提醒一下，我們通常每年新增約 100 所新校園，並在接下來兩到三年顯著放量成長。同時，我們的客戶成功團隊持續強化與教師的合作夥伴關係，支持留存，並推動本會計年度淨美元留存率達到 114%。

Turning to K-12. Fiscal year revenue was $884 million, down 9% year-over-year, reflecting the anticipated smaller market opportunity and difficult prior year comparison. Reocurring revenue grew 3% year-over-year to $620 million in the fiscal year. Despite market decline, performance came in slightly ahead of expectations due to stronger capture rates in key adoption markets.

接著談 K-12。本會計年度營收為 8.84 億美元，年減 9%，反映預期中的市場機會縮小以及去年同期比較基期偏高。本會計年度經常性營收年增 3% 至 6.20 億美元。儘管市場下滑，由於在關鍵採用市場的取得率更強，整體表現略優於預期。

In the fourth quarter, K-12 revenue was $126 million, down 10% year-over-year, with sequential improvements while reoccurring revenue declined 3% as prior year adoptions provided some durability. As we progress through the key selling season for the 2026 and 2027 school year, we are gaining real-time market intelligence.

第四季 K-12 營收為 1.26 億美元，年減 10%，但較前季有所改善；同時，因去年採用案帶來一定韌性，經常性營收年減 3%。隨著我們進入 2026 與 2027 學年的關鍵銷售季，我們正取得即時的市場情報。

In California, some districts are taking longer to finalize curriculum refresh plans amid higher volume of rubric choices, pilots and tools available. We have always expected the California math refresh cycle to take place over a handful of years. But the phasing has been recalibrated as we learn more from prospective customers.

在加州，部分學區在更大量的評分規準選項、試點與可用工具之下，完成課程更新計畫的定案時間拉長。我們一直預期加州數學更新週期會在數年內分階段進行。但隨著我們從潛在客戶獲得更多資訊，分期節奏已重新校準。

This delay has allowed us to better customize our materials to the market's rapidly evolving and fragmented needs and improve our capture rates and subsequent adoption years. Importantly, the market opportunity and opportunity to win remains intact.

這項延後讓我們能更好地依市場快速演變且碎片化的需求客製化教材，並提升我們的取得率以及後續年度的採用表現。重要的是，市場機會與勝出的機會仍然存在。

In Texas, our math program has been well received while state-sponsored curriculum has seen early successes due to upfront incentives provided by the state. Contracts have been shorter, reflecting a market in an experimentation mode. Historically, districts and educators favor high-quality integrated solutions and that pattern continues, giving us confidence in our ability to serve Texas in a very meaningful way going forward.

在德州，我們的數學課程獲得良好回響；同時，由於州政府提供前期誘因，州政府贊助的課程也在早期取得一些成功。合約期限較短，反映市場正處於試驗模式。從歷史來看，學區與教育工作者偏好高品質的一體化解決方案，而這一趨勢仍在延續，讓我們對未來以非常有意義的方式服務德州充滿信心。

We have seen this dynamic play out clearly in higher education, where our win ratio against OER and AI-generated curriculum vendors has been 6:1. OER was often adopted through top-down institutional mandates, but poor outcomes drove educators back to us, reinforcing that education is more than information. Pedagogical designs built with cognitive scaffolding, domain knowledge and alignment to thousands of learning objectives and localized standards demand deep expertise, and this is where McGraw Hill excels.

我們在高等教育領域已清楚看到這種動態：我們對上 OER 與 AI 生成課程供應商的勝率為 6:1。OER 往往透過由上而下的機構指令被採用，但不佳的成果使教育工作者回到我們這裡，進一步印證教育不僅僅是資訊。以認知鷹架、領域知識為基礎，並與數千項學習目標及在地化標準對齊的教學設計需要深厚專業，而這正是 McGraw Hill 的強項。

In Florida ELA, the market opportunity remains attractive, and our capture rate is trending in line with expectations. Looking ahead, we have visibility into several years of potential market expansion in K-12, driven largely by a nationwide ELA refresh cycle aligned with the science of reading.

在佛羅里達州的 ELA（英語語言藝術）方面，市場機會仍具吸引力，我們的取得率走勢符合預期。展望未來，我們可見 K-12 在未來數年具備潛在的市場擴張空間，主要由全美 ELA 更新週期所驅動，且該週期與「閱讀科學」一致。

McGraw Hill has been preparing for this opportunity with our ELA programs, Emerge, Summit and Soar, which integrates McGraw Hill Plus, AI-enabled capabilities and supplemental and intervention solutions. Early momentum in ELA is encouraging. Colorado's Department of Education score emerged as the highest rated curriculum and the latest instructional material review for K-3.

McGraw Hill 已透過我們的 ELA 課程 Emerge、Summit 與 Soar 為此機會做好準備，該方案整合 McGraw Hill Plus、AI 賦能能力，以及補充與介入解決方案。ELA 的早期動能令人鼓舞。在科羅拉多州，教育部評分結果顯示其在最新的 K-3 教學材料審查中獲評為最高等級課程。

We are also seeing early wins in urban open territory markets such as Seattle, where McGraw-Hill has historically had lower penetration. The first major state adoption opportunity for ELA will occur in California starting in fiscal year 2028, with the state expected to publish its list of approved vendors later this year. As a reminder, ELA is our largest subject area, representing approximately 40% of our historical K-12 revenue. So early signals from our new ELA program position us well.

我們也在西雅圖等都會區的公開競爭市場看到早期勝利，這些市場過去 McGraw-Hill 的滲透率較低。ELA 的第一個重大州級採用機會將自 2028 會計年度起在加州展開，預期該州將於今年稍晚公布核准供應商名單。提醒一下，ELA 是我們最大的科目領域，約占我們歷史 K-12 營收的 40%。因此，我們新 ELA 課程的早期訊號讓我們處於有利位置。

The Global Professional business remained steady in fiscal year 2026, with 4% year-over-year digital revenue growth supported by medical content and early traction from our AI-enabled solutions, which was partially offset by continued transition away from noncore print products.

全球專業（Global Professional）業務在 2026 會計年度維持穩健，數位營收年增 4%，受醫療內容與 AI 賦能解決方案的早期進展所支持，但部分被持續轉離非核心紙本產品所抵銷。

And in International, while revenue declined 7% year-over-year due to macro pressure in select markets, we are better positioned across multiple key markets heading into fiscal year 2027, supported by new commercial opportunities. Fourth-quarter performance was also impacted by Middle East project delays that have shifted into the second half of fiscal year 2027.

在國際市場方面，雖然因部分市場的總體環境壓力導致營收年減 7%，但在新的商業機會支持下，我們在進入 2027 會計年度時於多個關鍵市場的布局更佳。第四季表現亦受到中東專案延遲影響，相關收入已遞延至 2027 會計年度下半年。

While we recognized a $39 million impairment charge in the fourth quarter, driven by geopolitical and macroeconomic factors, we remain confident in the underlying fundamentals and long-term strategic importance of the global markets we serve with the outlook improving in fiscal year 2027 as headwinds ease and positioning strengthens.

儘管我們在第四季認列 3,900 萬美元的減損費用（受地緣政治與總體經濟因素驅動），我們仍對基本面以及所服務全球市場的長期策略重要性保持信心；隨著逆風緩解與布局強化，2027 會計年度展望將有所改善。

We ended the year with $254 million in cash and $704 million in total liquidity, with our revolving credit facilities remaining undrawn. During the year, we repaid $646 million of gross debt, reducing net leverage by approximately 80 basis points and lowering annualized cash interest expense by nearly $45 million.

我們在年末持有現金 2.54 億美元，總流動性為 7.04 億美元，循環信貸額度仍未動用。本年度我們償還 6.46 億美元的總債務，使淨槓桿約下降 80 個基點，並將年化現金利息費用降低近 4,500 萬美元。

Despite entering our seasonal cash trough through June, we reduced debt by an additional $50 million in the quarter, including buying $40 million of our highest coupon notes in the open market at a discount, and we will continue optimizing our capital structure in the future. Cash from operations was $331 million for the year, while CapEx and product development was just over $200 million. We generated $335 million of unlevered free cash flow in the fiscal year.

儘管進入至 6 月的季節性現金低谷期，我們本季仍額外減債 5,000 萬美元，其中包括在公開市場以折價回購 4,000 萬美元票息最高的票據；未來我們將持續優化資本結構。本年度營運活動現金流為 3.31 億美元，而資本支出（CapEx）與產品開發支出略高於 2.00 億美元。本會計年度我們產生 3.35 億美元的無槓桿自由現金流。

Turning to fiscal year 2027 guidance. We expect revenue to be in the range of $2.115 billion to $2.175 billion, remaining consistent with the dollar expectations we established at the time of our IPO. We expect approximately 55% of the full-year revenue in the first half of the fiscal year weighted towards the second quarter.

接著談 2027 會計年度指引。我們預期營收將介於 21.15 億至 21.75 億美元，與我們在 IPO 時所建立的金額預期一致。我們預期全年約 55% 的營收將落在上半年，且比重偏向第二季。

Reoccurring revenue is expected to be in the range of $1.587 billion to $1.627 billion, ahead of our IPO expectations, demonstrating a growing high-quality and predictable revenue stream. Adjusted EBITDA is expected to be in the range of $750 million to $790 million, ahead of expectations set at the time of the IPO. The implied adjusted EBITDA margin of 35.9% at the midpoint represents a year-over-year increase of 50 basis points.

經常性營收預期介於 15.87 億至 16.27 億美元，優於我們 IPO 時的預期，顯示高品質且可預測的營收來源持續成長。調整後 EBITDA 預期介於 7.50 億至 7.90 億美元，亦優於 IPO 時所設定的預期。以中位數計算的隱含調整後 EBITDA 利潤率為 35.9%，代表年增 50 個基點。

Unlevered free cash flow is expected to reach approximately $400 million, up roughly 20% year-over-year, with further growth anticipated in fiscal year 2028. CapEx and product development are expected to remain approximately 10% of revenue.

無槓桿自由現金流預期將達約 4.00 億美元，年增約 20%，並預期在 2028 會計年度進一步成長。資本支出（CapEx）與產品開發支出預期將維持約占營收的 10%。

Our capital allocation approach remains balanced and disciplined as we continue to prioritize reinvestment in the business, debt reduction and selective tuck-in acquisitions. Consistent with this approach, our Board has authorized a $50 million share repurchase plan, reflecting confidence in our sustained growth, strong profitability and meaningful cash generation.

我們的資本配置方法仍維持均衡且嚴謹；我們持續優先將資金再投資於業務、降低負債，並進行選擇性的補強式併購。與此一方針一致，我們董事會已核准一項5,000萬美元的庫藏股回購計畫，反映我們對持續成長、強勁獲利能力與可觀現金創造力的信心。

To summarize, fiscal year 2026 performance exceeded expectations, and our fiscal year 2027 revenue outlook in dollar terms remains consistent with the trajectory we outlined at the time of our IPO. At the same time, our guidance for reocurring revenue and adjusted EBITDA is above our IPO expectations, and we expect continued growth from these levels beyond fiscal year 2027.

總結而言，2026會計年度的表現超出預期，而我們對2027會計年度以美元計的營收展望，仍與我們在IPO時所勾勒的軌跡一致。同時，我們對經常性收入與調整後EBITDA的指引高於IPO時的預期，且我們預期在2027會計年度之後仍將在此基礎上持續成長。

The structural drivers of our business remain strong. In Higher Education, we expect continued share gains supported by durable pricing while taking a measured approach on enrollment assumptions ahead of the start of the academic year.

我們業務的結構性驅動因素依然強勁。在高等教育方面，我們預期在具韌性的定價支持下持續提升市占；同時，在學年開始前，對入學人數假設將採取審慎的作法。

In K-12, we expect improving trends as the market expands, supported by ELA momentum and the upcoming adoption cycle. In Global Professional, we will continue to prioritize growth in medical solutions. And in the International business, we are positioned more advantageously into fiscal year 2027.

在K-12方面，隨著市場擴大，我們預期趨勢將改善，並受益於ELA動能與即將到來的採用週期。在全球專業領域（Global Professional），我們將持續優先推動醫療解決方案的成長。而在國際業務方面，我們在邁入2027會計年度時具備更有利的布局。

Finally, a few modeling items for fiscal year 2027. Depreciation and amortization is expected to decline modestly versus fiscal year 2026, including $208 million of intangible asset amortization related to the Platinum acquisition in 2021. Stock-based compensation is expected to be approximately $15 million.

最後，提供幾項2027會計年度的模型假設。折舊與攤銷預期將較2026會計年度小幅下降，其中包含與2021年Platinum收購相關的2.08億美元無形資產攤銷。以股票為基礎的薪酬預期約為1,500萬美元。

Cash and GAAP taxes are expected to be in the range of $20 million to $40 million. And GAAP interest expense is expected to be below $180 million. These assumptions support GAAP net income in fiscal year 2027, increasing significantly compared to fiscal year 2026.

現金稅與GAAP稅負預期介於2,000萬至4,000萬美元。GAAP利息費用預期低於1.8億美元。這些假設支持我們在2027會計年度實現GAAP淨利，且相較2026會計年度將大幅增加。

With that, I'll turn the call back to the operator for questions.

接下來，我把電話交回給接線員進行提問。

(Operator Instructions) Stephen Sheldon, William Blair.

（接線員指示）Stephen Sheldon，William Blair。

First just on the revenue guide for fiscal 2027, can you help frame, (inaudible) framing at a high level what you're assuming in terms of segment level revenue growth and especially as we think about K-12 and higher ed?

首先關於2027會計年度的營收指引，你能否協助說明（聽不清）在高層次上，你們對各事業部營收成長的假設是什麼，特別是K-12與高等教育？

Sure thing. Yeah. So as I think about '27, how are we sort of positioning the business, and certainly, we're going to have continued share gains in higher ed, and you saw exceptional performance there. We see that continuing. You heard that I said we took a view of 1% enrollment gains in higher education for next year. We'll have to wait until we see students come on campus.

當然可以。是的。就我對2027年的看法，我們如何定位業務；而且可以確定的是，我們在高等教育將持續提升市占，你也看到那裡的表現非常出色。我們認為這種趨勢會延續。你也聽到我提到，我們對明年高等教育入學人數成長採取1%的假設。我們得等到看到學生回到校園後才能確認。

And then when we think about our K-12 business, we've established what we see as a slightly smaller than our original estimates in the overall market by about $100 million. As well as we're coming in at the lower end of our historical capture rates, about 25% to 30%. We can walk through that in greater detail, but we're seeing a couple of different dynamics playing out there as well.

接著談到K-12業務，我們判定整體市場規模可能比我們原先估計略小，約少1億美元。同時，我們的市占擷取率也落在歷史區間的較低端，約25%到30%。我們可以更詳細地說明，但我們也看到那裡正出現幾個不同的動態。

Really strong performance in open territory. We're continuing to take share. Supplemental intervention continues to be beneficial. We're seeing some benefits in terms of retention rates. We're seeing our ability to see the funnel grow.

在開放區域（open territory）的表現非常強勁。我們持續取得市占。補充性介入（supplemental intervention）仍持續帶來助益。我們在留存率方面看到一些改善。我們也看到漏斗（funnel）擴大的能力提升。

And then we highlighted in the prepared remarks around Texas and California, both of those markets are pretty dynamic. And while we're early in that cycle, we're performing slightly below our historical average there. So when you take that all together, you'll see that we're at the lower end of that 25% to 30% range.

另外，我們在事先準備的發言中也特別提到德州與加州，這兩個市場都相當動態。雖然我們仍處於該週期的早期階段，但我們在那裡的表現略低於歷史平均。因此綜合以上因素，你會看到我們落在25%到30%區間的較低端。

And then, nicely, you'll see return to growth in the International segment. We talked about some of the opportunities in select markets. We feel really good about where that business is heading. And then the underlying core Medical segment in Global Professional is that mid-single-digit grower, which gets offset by the tail of that print.

此外，你也會看到國際事業部恢復成長。我們談到在部分特定市場的一些機會。我們對該業務的發展方向感到非常樂觀。而在全球專業領域中，醫療（Medical）這個核心事業的基本面是中個位數成長，但會被印刷業務的尾端影響所抵銷。

And we believe we're about the final year of the wind down of that non-strategic core print business. So when I roll all that up together, Steven, that's where you get that midpoint of our guide around 2%. We feel good about it, and we think it's been a prudent guide.

我們認為這個非策略性的核心印刷（print）業務正處於縮減的最後一年左右。所以把這些全部加總起來，Steven，這就是你看到我們指引中點約2%的原因。我們對此感覺良好，也認為這是一個審慎的指引。

Very helpful. Maybe just drilling down on the Higher Ed segment. Obviously, you've had phenomenal growth there as we look back over the last couple of years. Maybe as we think about bridging expectations around fiscal 2026 growth to fiscal 2027, I mean you talked about enrollment. I guess maybe how is that growth algorithm changing as you think about the next year relative to what you've been delivering over the past couple?

非常有幫助。也許再深入談談高等教育事業部。顯然，回顧過去幾年你們在那裡有非常亮眼的成長。若要把2026會計年度的成長預期銜接到2027會計年度，你提到了入學人數。我想問的是，相較於過去幾年你們所交出的表現，明年你們對成長算法（growth algorithm）的看法會如何改變？

Yeah, you're right. It's been an exceptional year and continued exceptional performance by the team. So really pleased there. That 12%, let me break it down for you. The biggest driver of that is that ongoing sustained share gains. That will continue. We also benefited from enrollment; 3% to 4% of the growth of was enrollment related.

是的，你說得對。這是一個非常出色的一年，團隊也持續交出卓越表現。所以我們對此非常滿意。那個12%，我拆解給你。最大的驅動因素是持續且穩健的市占提升。這會延續。我們也受益於入學人數；其中約3%到4%的成長與入學人數相關。

You may have seen National Student Clearinghouse called 1.3% growth. We had 3% to 4%. The reason for that is the strength of our business and science courses as well as over-indexing to the two-year colleges. So when we look at our results, we had greater enrollment rates than the National Student Clearinghouse stated rates.

你可能看到National Student Clearinghouse提到成長1.3%。我們是3%到4%。原因在於我們在商業與科學課程的業務強勁，以及我們在兩年制學院的比重更高（over-indexing）。因此從我們的結果來看，我們的入學成長率高於National Student Clearinghouse所公布的數據。

The other thing is I've historically called out about a 1% net price, which is inflationary price offset by the movement into inclusive equitable access models. This year, we actually realized closer to 2%. And that's largely given all the additional features, all the value. As a reminder, we go through our value selling process, and as we add value to our customers, we're able to realize greater price.

另外一點是，我過去常提到約1%的淨價格（net price），也就是通膨式漲價被轉向包容性的公平可及（inclusive equitable access）模式所抵銷。今年我們實際實現的更接近2%。這主要是因為新增了許多功能與價值。提醒一下，我們會走價值銷售流程；當我們為客戶增加價值時，就能實現更高的價格。

And lastly, and this would be the component that's not reoccurring. This would be the 3% benefit we had from a sales returns. As we continue to move more digitally, we saw less returns come in. And then the corresponding reserve, we also lowered. So that had about a 3-percentage-point increase year-over-year, and we don't expect that to continue. So that's where we land at that mid-single-digit growth as we think about the outlook.

最後一點，而這會是非經常性的部分。這是我們因銷售退貨（sales returns）而獲得的3%助益。隨著我們持續更數位化，我們看到退貨減少。相對應的準備（reserve）我們也下調。因此年對年約增加3個百分點，而我們不預期這會持續。所以，這就是我們在展望中落在中個位數成長的原因。

Ryan MacDonald, Needham & Company.

Ryan MacDonald，Needham & Company。

Maybe shifting the conversation to the K-12 segment. Obviously, it sounds like a nice strong year there in the segment and sort of some good momentum sort of closing the year. Can you just talk about what exceeded our expectations the most within K-12 or pockets of strength?

也許把話題轉到K-12事業部。顯然，聽起來該事業部今年表現相當強勁，且在年底收官時有不錯的動能。你能否談談在K-12之中，哪些部分最超出我們的預期，或有哪些強勢的細分領域？

And then how is this informing your view a little bit more? Or how much of a shift change are we seeing within a few of those key states, Florida, Texas, California, as we think about how you're looking at the landscape in fiscal '27 versus maybe future years?

那麼，這如何進一步影響你們的看法？或者說，當我們思考你們對財年’27相較於未來年度的市場版圖判斷時，在幾個關鍵州（佛羅里達、德州、加州）我們看到的變化幅度有多大？

Sure, Ryan. Yeah, '26, we did exceed our expected capture rates. And as I highlighted, we are at the high end of that 25% to 30% range. Where is it coming from, right? We have continued strength in our science program. As we talked about last year, exceptional performance in Florida and Texas in Science program.

當然，Ryan。是的，’26年我們確實超出原先預期的拿下率（capture rates）。而且如我先前強調的，我們位於25%到30%區間的高端。那成長來自哪裡呢？我們的科學（Science）課程持續展現強勁動能。如同去年我們談到的，佛羅里達與德州在科學課程上的表現非常出色。

Now we saw that in other adoption states like Tennessee and Alabama as well as into the open territory. So science continues to perform above our historical rates. We also saw success in Florida ELA and some other open territory areas. So really pleased with that performance.

我們也在其他採用州（adoption states），例如田納西與阿拉巴馬，以及在開放市場（open territory）看到同樣的情況。因此，科學課程的表現持續高於我們的歷史水準。我們也在佛羅里達的ELA以及其他一些開放市場區域取得成功。所以我們對這樣的表現真的非常滿意。

Moving into sort of the '27, how do we think about it? As I mentioned, open territory continues to perform well. And I do believe we bottomed out with respect to our supplemental intervention, meaning the Astro dollars are behind us, and we're actually seeing some opportunity for meaningful growth. The pipeline is increasing in a meaningful way in supplemental intervention, and our retention rates are also growing. So I think we're very well positioned there.

接著談到’27年，我們怎麼看？如我提到的，開放市場持續表現良好。而且我確實相信，就補充性介入（supplemental intervention）而言我們已經觸底，也就是說Astro的那筆資金已經過去，現在反而看到有機會出現有意義的成長。補充性介入的案源管線（pipeline）正在以顯著幅度增加，我們的留存率（retention rates）也在提升。所以我認為我們在那一塊的布局非常到位。

We're watching closely Texas math as well as California math. We did see the California math market shift a little bit while the overall size of the market remains intact. And I always knew it was a multiyear adoption cycle. The overall timing has just shifted a little bit because of a few large districts moving into fiscal year 2028. And then we just watch -- as we move through the selling season, we're watching closely our win rates. And right now, they're falling a little behind our expectations, but we're still in that selling season.

我們也密切關注德州數學以及加州數學。我們確實看到加州數學市場出現一些結構性移動，但市場整體規模仍然維持不變。而且我一直知道那是一個跨多年的採用週期（multiyear adoption cycle）。整體時程只是因為幾個大型學區把採用推遲到財年2028而略有延後。另外，隨著我們進入銷售季（selling season），我們也在密切觀察我們的得標率（win rates）。目前它們略低於我們的預期，但我們仍在銷售季當中。

I would just add one thing, and it's related to both math and ELA literacy. Literacy for us is approximately 40% of our revenues, math, approximately 20%. And the two markets are very different right now. The way you teach math is very undecided. The pedagogy is not agreed upon by districts, by teachers and certainly not by states. It's a very fragmented market.

我再補充一點，這同時與數學和ELA讀寫（literacy）相關。對我們而言，讀寫約占營收的40%，數學約占20%。而這兩個市場目前差異非常大。數學的教法目前仍未定論。無論是學區、教師，當然也包括各州，對教學法（pedagogy）都尚未形成共識。這是一個非常分散的市場。

And that's what we're seeing in California. There's a lot of confusion. We've talked about the sheer number of companies that are bidding in California, and it's because it's so fragmented, and there's such a lot of discord still how you teach math.

這也是我們在加州看到的狀況。存在很多混亂。我們談過在加州投標的公司數量之多，正是因為市場非常碎片化，而且對於數學該怎麼教仍有很大的分歧。

Literacy, I don't think we've ever seen a market like this. There's now 42 states in the United States as well as a bill in the Senate that is mandating the way to teach literacy, which is the science of reading. That is an extraordinary legislative backdrop. We also think it can lead to a potential super cycle in the area of literacy.

至於讀寫，我想我們從未見過像現在這樣的市場。美國目前已有42個州，以及參議院也有一項法案，正在強制規定讀寫的教學方式，也就是「閱讀科學」（science of reading）。這是非常不尋常、也非常強力的立法背景。我們也認為這可能帶來讀寫領域的潛在超級週期（super cycle）。

We've made the largest single investment in our history in literacy curriculum, Emerge, Summit and Soar. And we've been bringing that curriculum out. We're kind of in year zero of it. And the exciting thing about it is that we scored the highest score on the Colorado rubric. We won large districts like Seattle. We just won in Kansas as an example.

我們在讀寫課程上做了公司史上最大的一筆單項投資：Emerge、Summit與Soar。而且我們一直在推出這套課程。我們目前算是它的第零年。令人振奮的是，我們在科羅拉多評分量表（Colorado rubric）上拿到最高分。我們贏下了像西雅圖這樣的大型學區。舉例來說，我們也剛在堪薩斯得標。

We're winning in districts that we were not even expected to be in the literacy market and they're big districts. And so one of the areas that we're excited about that momentum is building is in, as I mentioned, 40% of our business, which is the literacy market. And so it's a big area for us that we are excited about in 2027 and some really early optimistic wins.

我們正在一些原本外界甚至不預期我們能進入的讀寫市場學區中勝出，而且那些都是大型學區。因此，我們對這股動能的建立感到非常興奮；如我提到的，讀寫市場占我們業務約40%。所以這對我們而言是2027年非常重要、也令人期待的一個領域，而且我們已經看到一些非常早期、但樂觀的勝利。

I appreciate all that color there, both Bob and Philip. But maybe as a following up on the literacy opportunity because it does feel like a huge opportunity over the next three to five years here. Given there's that sort of state-level mandate and almost nearly linked it across the country, how are the states viewing sort of the opportunity versus sort of companies and vendors like McGraw Hill versus maybe rolling out additional state-owned OER kind of curriculum like you saw within Texas math? Like is that potential risk as we think about how this market evolves over the next few years?

我很感謝你們兩位（Bob與Philip）提供這麼多細節。不過我想就讀寫機會再追問一下，因為未來三到五年看起來確實是個巨大的機會。考量到有州層級的強制要求，而且幾乎在全國連動，各州如何看待這個機會？相較之下，像McGraw Hill這類公司與供應商的角色，或是各州是否可能推出更多州政府自有的OER（開放教育資源）式課程（就像你在德州數學看到的那樣）？隨著未來幾年市場演變，這會不會是潛在風險？

Yeah. I really haven't seen a contagion or similar movements in other states as it relates to literacy as Texas and Bluebonnet math. I'll start with that, first of all. The other thing -- a few things that you should know. With our literacy program, we're one of the few programs in the entire United States that is using the science of reading and is multilingual for most of the grades.

是的。就讀寫而言，我其實沒有看到像德州與Bluebonnet數學那樣，在其他州出現擴散效應（contagion）或類似的動作。我先從這點說起。另外還有幾件你應該知道的事。就我們的讀寫方案而言，我們是全美少數在多數年級同時採用「閱讀科學」且具備多語（multilingual）能力的方案之一。

And so you have to have a commitment not just around the program, but also around multi-languages. And then the investments that we make in accessibility for these programs is pretty extraordinary.

因此，你必須不只是對課程本身投入承諾，也要對多語能力投入承諾。此外，我們在這些方案的無障礙（accessibility）方面所做的投資也相當驚人。

The third thing is that the balance of screen time versus paper. It's a really delicate balance. Early on, you actually -- most districts are asking for less than 20% -- 20 minutes per week of screen time, which is something that we do. And then as you get a little bit more self-regulation, you can introduce more screens.

第三點是螢幕時間與紙本之間的平衡。這是一個非常微妙的平衡。在早期，多數學區其實要求每週螢幕時間少於20%——或每週20分鐘——而這正是我們能做到的。之後當學生的自我管理能力更成熟時，你才可以引入更多螢幕使用。

But literally, the balance between a curriculum that is holistic, literally starting and kindergarten all the way up to 12th grade, doing it in a multilingual capacity, making sure it's accessible and then making sure that you've got the right balance of screening times, it's a really, really hard thing to do. And as mentioned, it's one of the largest investments. We've made over $100 million investment in this literacy program.

但坦白說，要在一套從幼兒園一路到12年級的整體性課程中，以多語方式提供、確保無障礙，同時又要拿捏好螢幕使用時間的平衡，這是一件非常、非常困難的事。而且如前所述，這也是我們最大的一筆投資之一。我們在這個讀寫方案上已投資超過1億美元。

And as mentioned, we're scoring the highest in the rubrics. And that Colorado rubric, by the way, something that other states actually depend on, in addition to third-party assessments. And so you have to have your program assessed by third parties. You have to have it evaluate it by state. So you have to make sure that you've built it the right way. So we're just not seeing states move into that in the same way that we saw Texas experimenting with Bluebonnet.

並且如前所述，我們在各項評分量表（rubrics）中拿到最高分。順帶一提，科羅拉多的那套rubric其實也被其他州所依賴，此外還會參考第三方評估。因此，你必須讓第三方對你的方案進行評估。你也必須讓各州來評量它。所以你必須確保你是用正確的方式把它建構出來。因此，我們並沒有看到各州以同樣方式投入，就像我們看到德州在Bluebonnet上做的試驗那樣。

And the other thing I would just mention as it relates to Bluebonnet. One of the challenges, I think that other states are looking at is that we just recently announced there was over 4,000 errors in that curriculum. And you just can't take that kind of risk with the literacy program.

另外，談到Bluebonnet，我還想補充一點。我認為其他州正在關注的一個挑戰是：我們最近才宣布那套課程中有超過4,000個錯誤。而在讀寫方案上，你根本承擔不起這種風險。

Alexey Philippov, JPMorgan.

Alexey Philippov，摩根大通。

Yeah, a question to Philip. Maybe a bit more to elaborate on the science of rating program and the opportunity you see has to capture that. Is there a way to quantify that? Is it like a time expansion for you? And which years do you expect this to fall?

是的，想請教 Philip 一個問題。也許可以再多闡述一下「閱讀科學」評量計畫的科學基礎，以及你們看到可把握的機會。有沒有辦法把它量化？對你們來說，這像是時間上的延展嗎？你預期會落在哪些年度？

I think it's -- we look at the TAM expansion as an international opportunity. The need for science of reading is not just here in the United States. It is global, as I would say, the previous ways of teaching literacy around the world. was highly variable. As I mentioned, it's 40% of our revenue.

我想是——我們把 TAM 的擴張視為一個國際性的機會。對閱讀科學的需求不只是在美國。它是全球性的；我會說，過去世界各地的識字教學方式差異很大。如我提到的，這占我們營收的 40%。

Our program in the past wonders actually generated over $1.5 billion for us over the course of 10 years. And so we look at it as both potential global TAM expansion and then also, I'll call it, TAM acceleration here in the United States. I mentioned it before, I don't think we've ever seen a legislative environment like this before, where 42 states are now mandating this is the way to teach.

我們過去的 Wonders 課程在 10 年期間實際上為我們帶來超過 15 億美元的收入。因此我們把它視為一方面是全球 TAM 擴張的潛力，另一方面——我稱之為——在美國本土的 TAM 加速。我之前提過，我不認為我們曾見過像現在這樣的立法環境：目前有 42 個州正在強制規定必須用這種方式來教。

There's been a lot -- one of the biggest challenges right now across the educational institutions across the United States is the challenge in literacy scores. And it's disappointing to see that 83% of states had moved backwards. We're at roughly about a 33% reading on grade level right now. It is, in a lot of ways, a national crisis and an international crisis.

目前美國各教育機構面臨的最大挑戰之一，是識字成績的下滑。令人失望的是，有 83% 的州出現倒退。目前大約只有 33% 的學生閱讀能力達到年級水準。在很多方面，這是一場全國性的危機，也是國際性的危機。

And so the fact that it has now been settled, that science of reading is the way to teach, it's really what's called a phonics based approach as opposed to a whole language approach, really important. And so we're seeing, as I mentioned, a lot of districts that quite frankly, were not in our pipeline at the start of last year come into our pipeline.

因此，既然現在已經有共識：閱讀科學就是教學的方法——它其實是所謂以自然發音（phonics）為基礎的方法，而不是整體語言（whole language）的方法——這點非常重要。所以我們看到，如我提到的，很多學區坦白說在去年年初並不在我們的銷售管線中，但現在已經進入了我們的管線。

We've landed relationships that we weren't expecting to land. I was just recently up at an elementary school up in Chicago that reached out to me personally to start using our literacy program. So really extraordinary momentum that we're seeing in that space. So we see, I'll say, TAM acceleration. We're just in the early stages of quantifying it. So we'll come back to you with more details as we're able to start assessing that.

我們也拿下了一些原本沒預期能拿下的合作關係。我最近才去芝加哥的一所小學，那邊的校方親自聯絡我，希望開始使用我們的識字課程。所以我們在這個領域看到非常驚人的動能。因此我會說，我們看到 TAM 的加速。我們才剛開始量化它。等我們能開始評估之後，會再提供更多細節給你。

Yeah. And a quick technical on this international impairment. Can you elaborate on that? It sounds like you expect international to kind of accelerate from here. How to square that is the impairment that you recognized?

好的。另外一個快速的技術性問題，關於這次國際業務的減損。你能再多說明一下嗎？聽起來你們預期國際業務會從這裡開始加速。那要如何與你們認列的減損相互對照、解釋？

We look at international, it's really important to kind of get a sense. I think that I spoke about this on the last call that we look at the education market is about a $7.3 trillion market on a global basis. The middle class is growing in areas like Latin America, the Middle East. We've got some great traction in those marketplaces.

我們看國際市場時，先掌握整體輪廓很重要。我想我在上一次電話會議提過，我們認為教育市場在全球規模大約是 7.3 兆美元。拉丁美洲、中東等地的中產階級正在成長。我們在那些市場也取得了很好的進展。

As the middle class grows, middle class tends to suspend approximately 2.5% on education. And so those markets has just recently done in Latin America, and I was just simply -- I'll just say, so impressed with the work that was going in K-12 and high school is completely dedicated to science, to universities. I spent some time with the university that had over 300,000 people enrolled in it.

隨著中產階級成長，中產階級通常會把大約 2.5% 的支出用在教育上。因此我最近才剛在拉丁美洲走訪那些市場，我只能說——我對當地在 K-12 與高中所做的工作印象非常深刻，完全投入於科學教育，延伸到大學端也是如此。我也花了一些時間拜訪一所大學，該校註冊人數超過 30 萬人。

And so we're really excited about those opportunities internationally, and we are seeing growth in those markets. And so internally, inside of the company, we're better aligning to be able to be more reactive to local requirements, language requirements, cultural requirements, pedagogy.

因此我們對國際上的這些機會感到非常興奮，而且我們確實看到那些市場在成長。所以在公司內部，我們也在更好地調整配置，以便能更快速回應在地需求、語言需求、文化需求以及教學法。

And we do view it as a great opportunity for growth for this company. And that's across literally our K-12 segment, our University segment and then also our Medical segments. And so you'll see us focusing on how we are part of that really exciting growth worldwide.

我們確實把它視為公司一個很好的成長機會。這涵蓋我們的 K-12 事業、我們的大學事業，以及我們的醫療事業。因此你會看到我們聚焦於：我們如何成為全球這波令人振奮的成長的一部分。

Steve Koenig, Macquarie Group.

Steve Koenig，麥格理集團。

It's Pete Koenig from Macquarie. Congratulations on the solid quarter and solid year. A question here for Philip. We noted your AI blueprint fill posted with the earnings materials, super helpful in understanding your AI strategy. Maybe can you help us understand how that strategy compares with your competitors?

我是麥格理的 Pete Koenig。恭喜本季與全年表現穩健。我有個問題想請教 Philip。我們注意到你們在財報資料中附上了你們發布的 AI 藍圖，對理解你們的 AI 策略非常有幫助。也許你能協助我們理解，這個策略與競爭對手相比如何？

And maybe the second part of my question, if I may, is about open educational resources, which you mentioned in your remarks, and you did reference Bluebonnet math in your Q&A. Just curious where -- how do you think about where AI and open educational resources are headed?

如果可以的話，我問題的第二部分是關於你在談話中提到的開放教育資源（OER），而且你在問答中也提到 Bluebonnet math。我只是好奇——你如何看待 AI 與開放教育資源未來的發展方向？

Thank you very much for the question. And thanks for reading the AI blueprint. We're really excited about it. I think I mentioned that we talked about the fact that we have over 7.5 million of users of our AI solutions. And these are tools that just came out over the past 18 to 24 months. And so we're pretty excited about the early traction that we're seeing and there's a lot more ahead.

非常感謝你的提問。也謝謝你閱讀 AI 藍圖。我們對此非常興奮。我想我提過，我們的 AI 解決方案使用者已超過 750 萬。而這些工具是在過去 18 到 24 個月內才推出的。因此我們對目前看到的早期進展感到相當振奮，而且後面還有更多。

I think a lot of individuals that come into the education industry have a lot of misconceptions and they vastly underestimate what's necessary in this marketplace. And so you'll see a lot of startups that come in. You'll see some organizations that have some curriculum or some tools. And the thing that differentiates McGraw Hill is four things.

我認為很多進入教育產業的人有不少誤解，並且嚴重低估了在這個市場中所需要的東西。所以你會看到很多新創公司進來。你也會看到一些組織有一些課程內容或工具。而 McGraw Hill 的差異化在於四件事。

First of all, the deep pedological curriculum. From the outside, most people don't realize that you're taking sequences of information, something like in liters, you have to weave together 1,000 separate learning requirements. In mathematics from K-12, over 4,000 learning requirements. You have to sequence this content.

第一，是深厚的教學法與課程內容。從外部看，多數人不會意識到，你是在把一連串資訊做成序列；例如在識字領域，你必須把 1,000 個獨立的學習要求編織在一起。在 K-12 數學則有超過 4,000 個學習要求。你必須把這些內容做出順序安排。

So it's not just about just putting out a textbook or a piece of curriculum, you literally have to sequence it right down to the week. What are the third grader doing in their seventh week? And when they have an assembly, how do you kind of squeeze in addition and subtraction and multiplication tables?

所以這不只是推出一本教科書或一套課程而已，你必須把它排到非常細，甚至細到每一週。三年級學生在第七週要做什麼？如果遇到集會活動，你要如何把加法、減法以及乘法表擠進去？

So that sequence of content is hard. We have literally hundreds of thousands of titles that we've been doing over the course of our history, tens of thousands that are sequenced across 500 different subjects where we sequence how a learner learns. That's literally in many cases from -- through 12 years of their life. So that content is a huge moat versus other organizations.

因此，內容的序列化很難。我們在公司歷史中做過數十萬個品項，其中有數萬個是跨 500 個不同科目所編排的學習序列，去安排學習者如何學習。在很多情況下，這實際上涵蓋——貫穿他們人生中的 12 年。因此，相較其他組織，這些內容形成了一道非常深的護城河。

The second thing is that we have these tools that do personalized delivery. Sharpen is a great example. A study guy that we use in Rowan College, a professor there analyzed it and said that the 60% of the students that were using in his class got 47% higher grade in their final exam, a 21% higher total grade.

第二點是，我們有能進行個人化傳遞的工具。Sharpen 就是一個很好的例子。我們在 Rowan College 使用的一個學習工具，那裡的一位教授分析後表示，他班上有使用該工具的 60% 學生，期末考成績高出 47%，總成績高出 21%。

We had children and ALEKS Adventure where we saw 25-percentage-point increases in their grades and second grader is 55 percentage points. And so personalized delivery, whether or not you want to listen to a podcast, you want to see it visually or you want to gamify it. So that personalized delivery is the second thing.

我們有兒童使用 ALEKS Adventure，我們看到他們的成績提升了 25 個百分點，而二年級學生更是提升了 55 個百分點。因此，個人化的傳遞方式——不論你想聽 Podcast、想用視覺方式觀看，或是想把它遊戲化。所以，個人化傳遞是第二件事。

The third, and this is critical in an era of outcomes is our data. We have over 190 terabytes of data. We get about 25.6 billion learning interactions per year. Now when you're training AI, you need to know when the AI is on task or off task. And we sense this at a very fine grain level. We can say for a subject like algebra, out of the two to the 500th power of how -- where you're at in your learning journey, precisely where you're at. So that data is critical.

第三點，而且在成果導向的時代這一點至關重要，是我們的數據。我們擁有超過 190 TB 的數據。我們每年大約有 256 億次學習互動。現在當你在訓練 AI 時，你需要知道 AI 是在專注於任務還是偏離任務。而我們能以非常細緻的粒度感知這一點。以代數這樣的科目為例，在 2 的 500 次方種「你在學習旅程中的位置」之中，我們可以精準判斷你到底在哪裡。因此，這些數據至關重要。

The last is being able to integrate this into workflows, and we have one of the largest and most, I'll say, impressive -- since I got here, I'm seeing it, customer success organizations and go-to-market teams, where they're actually working to integrate this into the classroom. Every school that I visited, every teacher that I visited in K-12, they've asked, it's not just enough to produce this, show us how to use this in the class with different types of learners.

最後一點是能夠把這整合進工作流程中；而我們擁有規模最大且我會說非常令人印象深刻的——自從我來到這裡看到的——客戶成功組織與市場推進團隊，他們實際上正在努力把這整合進教室。我走訪的每一所學校、每一位 K-12 教師都問：不只是把這些做出來就夠了，請示範如何在課堂上、面對不同類型的學習者使用它。

So the four things that we have are content that is ordered pedologically, the second thing is personalized delivery, third is data and the fourth is this ability to be able to close the last mile. Now one of the things that's most interesting, and I mentioned this in my comments, out of the dozens and dozens and dozens of schools and teachers and administrators that I met, investors have asked me, will AI disrupt our business?

所以我們具備的四件事是：依教學法（pedologically）排序的內容；第二是個人化傳遞；第三是數據；第四是能夠補上最後一哩路的能力。現在最有意思的一點——我在評論中也提到——在我見過的數十所、數十所、數十所學校以及教師與行政人員之中，投資人問我：AI 會不會顛覆我們的業務？

But like our customers are not. They're thrilled that we are adopting AI and that we're making it simple to them. And they see the value of what we produce. And this is really borne out in the OER market. I think I've heard from a lot of people question, well, can I just take OER like a Bluebonnet and couple it together with AI and boom, there's my curriculum.

但我們的客戶並不這麼想。他們很興奮我們正在採用 AI，並且把它做得對他們而言很簡單。他們也看見我們產出的價值。而這在 OER 市場上也確實得到印證。我想我聽到很多人質疑：那我是不是只要拿 OER（例如 Bluebonnet）再搭配 AI，砰，就有我的課程了？

The reality is that we're winning almost 6 times as many times as we're not with OER. So for every unit of OER that we lose to, we win six terms of our outcomes. And we're seeing a lot of organizations, in a world of AI tools in a world of OER, we're actually seeing a lot of win backs where organizations and individuals are coming back to us because they're realizing the complexity of this.

現實是：在與 OER 的競爭中，我們的勝率幾乎是失利的 6 倍。也就是說，每輸掉 1 個 OER 單位，我們就能以成果贏回 6 個學期（terms）。而我們也看到許多組織——在 AI 工具與 OER 並存的世界裡——其實出現很多「回流」案例：組織與個人回到我們這裡，因為他們意識到這件事的複雜度。

You have to tie together assessment. You have to tie together pedagogy, and you have to make it work inside of the workflow and how the student learns. So a really exciting time for us.

你必須把評量串起來。你必須把教學法串起來，而且你必須讓它在工作流程中運作、符合學生的學習方式。所以對我們來說是非常令人振奮的時刻。

Marvin Fong, BTIG.

Marvin Fong，BTIG。

Great. Congratulations on the results. Maybe start with a bigger picture on guidance. And just sort of how you formulated that, what's your philosophy behind the guidance you provided, both revenue and EBITDA for not only 2027, but just structurally, how do you think about constructing your guidance?

很好。恭喜這次的成果。也許先從更宏觀的指引（guidance）談起。想請教你們是如何制定的、你們提供的指引背後的理念是什麼——不論是營收還是 EBITDA，不僅是 2027 年，從結構上你們如何看待指引的建構？

Sure. Thanks, Marvin. Consistent with how we've done this in the past and even if we rewind back to the forecast models we provided during our IPO and the road show, we've built this process, and we're consistently following that. And what we do is we understand what the information we have at this point in time.

當然。謝謝你，Marvin。我們的做法與過去一致；即使回溯到 IPO 與路演期間我們提供的預測模型，我們建立了這套流程，並且持續一致地遵循。我們所做的是先理解在此時此刻我們掌握了哪些資訊。

And again, we're early in the selling season with K-12. We've provided some insights around enrollment for higher ed. But based on information we have at this point in time, we provide that as our -- as the midpoint of our guide, both on revenue and EBITDA.

而且再次強調，K-12 目前仍在銷售季的早期階段。我們也就高等教育的入學人數提供了一些觀察。但基於我們此刻掌握的資訊，我們把它作為指引的——作為指引中點（midpoint），同時涵蓋營收與 EBITDA。

And then what we do is we provide a range. And consistent with what we've done in the past, that range provides for different outcomes, right? Should enrollments increase, should capture rates change, all of those things are reflected in that range.

接著我們會提供一個區間。而且與過去一致，這個區間反映不同的結果情境，對吧？例如入學人數是否增加、取得率（capture rates）是否變動，所有這些因素都會反映在該區間中。

And then ultimately, what we'll do is we'll revisit this as we move through that and ultimately once we start seeing students on campus, which, as you know, is after our second quarter that allow us to have greater insight as to the remainder of the year. So we're following a consistent approach that we've done in the past.

最後，我們會在推進過程中重新檢視；尤其是一旦我們開始看到學生回到校園——如你所知，那是在我們第二季之後——這會讓我們對全年剩餘期間有更高的能見度。所以我們遵循的是過去一貫的做法。

Perfect. And my follow-up, maybe just something that you mentioned, Philip, in your prepared remarks just about growing your addressable markets. I think you mentioned non-educational as well as educational markets. Maybe with respect to like AI and agentic tools, can you just kind of expand on that? That seems like an interesting point as you try to grow the top line, what are these kind of new markets you're thinking about?

完美。那我的追問，可能是你在事先準備的發言中提到的，Philip，關於擴大可服務市場（addressable markets）。我記得你提到非教育以及教育市場。就 AI 與代理式（agentic）工具而言，你能否再多展開談談？這似乎是你們在推動營收成長時一個很有意思的點：你們正在思考哪些新市場？

One of the most impactful visits that I had was at a medical school, and I was meeting with a professor, and he just come out of the American Medical Association. He talked about this concept of precision education. If a particular doctor have not studied or not worked with them a particular patient population, how do you really make sure that that doctor, when they're walking into a situation, is able to have education just in time right at the right moment?

我最具影響力的一次拜訪是在一所醫學院；我與一位教授會面，而他剛從美國醫學會（American Medical Association）回來。他談到一個概念：精準教育（precision education）。如果某位醫師沒有研究過、或沒有接觸過某一特定病患族群，你要如何真正確保這位醫師在走進情境時，能在正確的時刻即時獲得所需的教育？

And this is an actual -- is a trend across lots of organizations, just-in-time education. And so -- and one of the most interesting things with the American Association of Medicine has just given out 12 grants for the precision education for addressing this exact concept of really making sure that doctors in the medical profession is able to get the right education at the right moment, especially in a changing world of education.

而這其實是一個跨許多組織的趨勢：即時教育（just-in-time education）。因此——而且最有意思的是，美國醫學協會（American Association of Medicine）剛剛針對精準教育發放了 12 項補助，正是為了解決這個概念：確保醫療專業中的醫師能在正確的時刻獲得正確的教育，尤其是在教育快速變動的世界裡。

I read a study recently that in both computer science and medicine, that the number of medical papers is doubling every single year. So it's an extraordinary amount of knowledge that's coming out. I've talked in the past about the multitrillion dollar industries that are emerging and the sheer amount of knowledge that humans have to now learn, not just in K-12, but throughout their entire lifetime.

我最近讀到一項研究指出，在電腦科學與醫學這兩個領域，醫學論文的數量每一年都在翻倍。因此，新知的產出量非常驚人。我過去也談過正在浮現的數兆美元級產業，以及人類如今必須學習的龐大知識量——不只是 K-12，而是貫穿整個一生。

What we announced today that we're starting to pilot is this concept of agentic curriculum. And that is the ability to be able to consult our curriculum using agentic technologies, the MCP as an example. One of the organizations -- actually two of the organizations are not traditional education companies that are actually piloting that are working with us and starting to pilot our content.

我們今天宣布開始試點（pilot）的，是「代理式課程」（agentic curriculum）這個概念。也就是能夠使用代理式技術來諮詢我們的課程內容，例如以 MCP 為例。其中有一些組織——其實是兩個組織——並非傳統教育公司，但正在與我們合作試點，並開始試用我們的內容。

These are organizations that are in the healthcare industry that actually don't educate, but they just have professionals that are working in medicine and working in pharmaceuticals that are actually -- want to be able to consult our content. And so I look at the lifelong journey of a learner, increasingly, McGraw Hill wants to connect that journey from the earliest stages to any stage in your life, whether or not continuing education or whether or not it's on professional education, and agentic technologies are going to allow us to do that.

這些組織來自醫療保健產業，本身其實不做教育，但他們有在醫療與製藥領域工作的專業人士，並且希望能夠查詢、使用我們的內容。因此我看待學習者的終身旅程時，McGraw Hill 越來越希望把這段旅程從最早期階段一路連結到人生任何階段——不論是繼續教育或專業教育——而代理式技術將使我們能做到這一點。

One of the things I've talked about in the past is that where I expect AI to go in companies like us, so there's going to be a knowledge economy where you'll be able to plug in knowledge. You'll be able to plug in knowledge agents. Right now, we are experimenting with a knowledge agent.

我過去談過的一點是，我預期 AI 在像我們這樣的公司會往哪個方向發展，因此將會出現一個知識經濟，在那裡你可以把知識「接上去」。你也能把知識代理（knowledge agents）接上去。目前我們正在試驗一個知識代理。

We view that in the same way that there's these token-based models in artificial intelligence, the prompt-based business models, we view this as an opportunity to expand the accessibility and the -- I'll say, the business model for McGraw Hill over time with this new knowledge economy that's emerging.

我們看待這件事的方式，和人工智慧中的代幣（token）模型、以提示（prompt）為基礎的商業模式很類似；我們把它視為一個機會，能隨著這個正在形成的新知識經濟，逐步擴大 McGraw Hill 的可及性，以及——我會說——長期的商業模式。

George Tong, Goldman Sachs.

George Tong，高盛。

I wanted to dive deeper into the agentic curriculum opportunity. Can you share more color on the timeline to launch as well as the pricing model and wallet opportunity?

我想更深入探討代理式（agentic）課程的機會。你能否分享更多關於推出時程、定價模式以及可爭取的錢包份額（wallet opportunity）的細節？

Sure. George, Thanks very much as well for your -- the AI framework they put out. I really enjoyed reading it. The -- and I really agree with the core concepts around the opportunity ahead. For agentic, for us, we just -- over the course of the past few months since arriving, I've spent a lot of time with the team, and the team has really moved quickly.

當然。George，也非常感謝你——你們發布的 AI 架構。我讀得很享受。我也非常認同其中關於未來機會的核心概念。就代理式（agentic）而言，對我們來說——在過去幾個月我到任以來，我花了很多時間和團隊在一起，而團隊的推進速度真的很快。

I talked in my last call about the excitement that I have around innovation. Since I arrived, we've been able to actually build and start to internally pilot the agentic technologies, and we're just starting to move into agentic pilots with external customers. This has been something that's been asked for by lots of universities, lots of non-education organizations. It's been asked -- it's being requested.

我在上一次電話會議中談到我對創新的興奮。自從我加入以來，我們實際上已經建置並開始在內部試點代理式技術，而我們也才剛開始與外部客戶進行代理式試點。這是許多大學、許多非教育機構一直在要求的。這一直被提出——也正在被要求。

A safe way, a, I'll call it, an auditable way of being able to deliver our content. I think that there's been a choice up to now where either you take all of your content and throw it inside of a large AI model and you lose control of it, or you don't do anything in AI. And this agentic technologies allows us to safely be able to answer any questions and be able to monitor, be able to secure it in the same way that you would with an API.

一種安全的方式，我稱之為一種可稽核（auditable）的方式，來交付我們的內容。我認為到目前為止一直存在一個選擇：要嘛你把所有內容丟進一個大型 AI 模型裡，然後你就失去對內容的控制；要嘛你在 AI 上什麼都不做。而這些代理式技術讓我們能夠安全地回答任何問題，並且能夠監控、能夠保護它，就像你使用 API 一樣。

I view it almost like you call it a cloud for knowledge or an app store for knowledge. And so it's exciting technology. It's -- technology only came out over the past roughly about 12 or so, 18 months, and these standards are just starting to emerge in the AI world, MCP and agentic technologies. And so really excited about the fact that the team is able to build this this quickly and we're moving into pilot.

我幾乎把它看作是「知識的雲端」或「知識的 App Store」。所以這是一項令人興奮的技術。這——這項技術大約在過去約 12 到 18 個月才出現，而這些標準也才剛開始在 AI 世界中浮現，包括 MCP 與代理式技術。因此，我對團隊能這麼快建置出來、並且我們正在進入試點階段，感到非常興奮。

Now if you look at what's happening in the world of AI pricing, I think you're seeing that pricing is all over the map. I saw some stories over the past 24 hours that OpenAI may drop its token prices. We're seeing SaaS companies have a whole variety of pricing models. In some cases, it's simply an up charge, in other cases, you're actually getting charged by the token.

現在如果你看 AI 定價世界正在發生的事，我認為你會看到定價非常分歧。我在過去 24 小時看到一些報導，OpenAI 可能會下調其 token 價格。我們也看到 SaaS 公司有各式各樣的定價模式。有些情況只是加價（up charge），另一些情況則是按 token 計費。

What we're headed into pilot to do is to really understand how our customers use this. And in education, as you can imagine, they want predictability. They don't want some AI agent that's going to charge them potentially by the questions they have no control over, and certainly, that becomes an extraordinarily expensive proposition for the university, for the K-12 school, for the medical school and for the student.

我們接下來要進入試點要做的，是確實了解客戶如何使用它。而在教育領域，如你所想，他們希望可預測性。他們不希望有某個 AI 代理可能按提問次數向他們收費，而這些提問他們無法控制；而且這對大學、K-12 學校、醫學院以及學生而言，肯定會變成極其昂貴的方案。

And I would tell you that I had mentioned before that there is a tool fatigue where you have as many of the 25 different log-ins for students, 40 log-ins for professors. They want simplicity. And so we're going to experiment with the business models over the course of between now and the end of the year to determine precisely what's the best model for education. And as a result, we think that we're going to be able to go after kind of new wallet share and as well simplify this.

我也想告訴你，我之前提過存在一種工具疲勞（tool fatigue）：學生可能要登入多達 25 種不同系統，教授可能要 40 個登入。他們想要簡單。因此，我們會在現在到年底這段期間試驗不同的商業模式，以精確判斷哪一種最適合教育市場。因此，我們認為我們將能夠爭取某種新的錢包份額（wallet share），同時也把這件事簡化。

So I'm seeing in the world of -- when I was in cloud, I watched a lot of organizations start adopting the cloud and then costs went out of control and they had to quickly get control of it. We're seeing the same thing in the world of AI, where costs are quickly getting out of control and now financial FinOps, as we call it, is coming to AI.

所以我看到在——當我在雲端領域時，我看過很多組織開始採用雲端，然後成本失控，接著他們必須迅速把成本管控回來。我們在 AI 世界也看到同樣的情況：成本正在快速失控，而現在我們所稱的財務營運（FinOps）正走向 AI。

And so we're going to really make this simple. We're going to make it streamlined for the user, and we're going to have a business model that is probably better than a student or a university could get on their own. And so we'll bring you more information on that as the business model emerges over the next six months.

因此我們會把它做得非常簡單。我們會讓使用者端更精簡、更順暢，而且我們會採用一種商業模式，可能比學生或大學自己去取得的條件更好。因此，隨著未來六個月商業模式逐步成形，我們會提供你更多資訊。

Got it. That's helpful. And then in K-12, can you discuss where in the historical 25% to 30% capture rate range you expect to land in fiscal 2027?

了解。這很有幫助。另外在 K-12 方面，你能否談談在歷史上 25% 到 30% 的取得率（capture rate）區間中，你預期在 2027 會計年度會落在哪個位置？

Yeah. George, we anticipate we're going to be at the low end of that range.

是的。George，我們預期會落在該區間的低端。

Any color on how that evolves over the course of the year and opportunities for improvement by state?

能否補充一下這在一年當中會如何演變，以及各州有哪些改善機會？

Yeah. I think the areas that we are most focused on for upside will continue to come in the open territory and supplemental intervention, and so we'll continue to watch that. And then like I mentioned, Texas and California, as we continue to see more calls, we'll monitor that closely. But the areas that provide me the most excitement is around supplemental intervention where we have 5% share, and that represents about 15% of our overall K-12 revenues as well as open territory.

可以。我認為我們最聚焦、最有上行空間的領域，將持續來自開放領域（open territory）與補充性介入（supplemental intervention），因此我們會持續觀察。另外如我提到的，德州與加州，隨著我們持續接到更多來電／詢問，我們會密切監控。但最讓我感到興奮的領域是在補充性介入，我們在那裡只有 5% 的市占，而它約占我們整體 K-12 營收的 15%，以及開放領域（open territory）。

(Operator Instructions) Faiza Alwy, Deutsche Bank.

(接線員指示) Faiza Alwy，德意志銀行。

I just wanted to put a finer point on your messaging around K-12 because from where I sit, it seems like there's been a pretty significant backlash from parent groups around digital tools and technology in the classroom, and a lot of parents are wanting to go back to paper and pen.

我只是想把你們對 K-12 的訊息再問得更精準一些，因為就我所見，家長團體對教室裡的數位工具與科技似乎出現了相當明顯的反彈，很多家長希望回到紙筆。

And is that hurting you? And is that kind of -- is that why you're kind of at the lower end of that capture rate? Or do you see that as an opportunity? Just would love to get your perspective on that.

這是否正在傷害你們？而這是否——是否就是你們取得率落在較低端的原因？或者你們把這視為一個機會？很希望聽聽你們對此的看法。

Thank you very much for the question. It's a question I get asked a lot is screen time. And this actually is a significant strength for McGraw Hill because we have such a strong history of being able to deliver education in whatever way a school or a student requires. And today, we're one of the few organizations that is in the world that can deliver everything from AI to paper and vice versa.

非常感謝你的提問。這是一個我經常被問到的問題：螢幕使用時間（screen time）。而這其實是 McGraw Hill 的一項重要優勢，因為我們在以學校或學生所需要的任何方式提供教育方面，有非常深厚的歷史。而今天，我們是全球少數能夠提供從 AI 到紙本、以及反過來亦然的所有形式的機構之一。

And what we're seeing, first and foremost is, yes, screens, especially for earlier learners, there's a negative correlation between screen time and comprehension. There has been a lot of studies around this topic. And what we're seeing is this idea of really self-regulation. The ability to self-regulate when you're on a screen is simply lower in a lot of cases, and you don't have as much, I'll say, dwell time on the subjects.

而我們看到的，首先是：是的，螢幕——尤其對較年幼的學習者而言——螢幕使用時間與理解力之間呈現負相關。針對這個主題已經有很多研究。而我們看到的是一種真正的自我調節（self-regulation）的概念。在螢幕上進行自我調節的能力，在很多情況下就是比較低，而且你在各個主題上的——我會說——停留時間（dwell time）也沒有那麼長。

In paper, we see as much as 3 to 4 times more dwell time and rereading of content. That's engagement, engagement for a learner drives comprehension. And so that balance, I think, is something that's really people are waking up to.

在紙本上，我們看到內容的停留時間與重讀次數高出多達 3 到 4 倍。那就是投入度；學習者的投入度會帶動理解。因此我認為這種平衡，確實是大家正在逐漸意識到的。

And over the course of the past few years, as social media has really been driving up a lot more screen time, this has become an increasing focus, and we're seeing a lot of schools that are banning screens right now in the classroom or I should say, mobile phones as an example. I think we're up to over 30 states in that space. And so distraction is real. Comprehension -- lack of comprehension is real when there's more screens.

而在過去幾年中，隨著社群媒體確實推升了更多螢幕使用時間，這已成為愈來愈受關注的焦點；我們也看到許多學校現在在教室裡禁止使用螢幕，或者我應該說，以手機為例。我想在這方面我們已經超過 30 個州。因此分心是真實存在的。當螢幕更多時，理解力——理解不足——也是真實存在的。

Now with that being said, there is a place and a time for screens. When I take something like our Emerge product, as I mentioned, we are down at roughly about 7% or 20 minutes per week screen time for Emerge. It was one of the most important questions that was asked in Seattle by the Board when we were -- when they were deciding to adopt our product.

話雖如此，螢幕也有其適當的時機與用途。以我們的 Emerge 產品為例，如我提到的，Emerge 的螢幕時間大約降到約 7%，或每週約 20 分鐘。當時西雅圖的教育委員會在——他們決定採用我們產品時——提出的最重要問題之一就是這個。

What screens are really good at is actually assessment. And so you take a product like our ALEKS product for algebra and within 25 questions, we can tell you precisely out of the 500 skills that a student needs to have, how many they have with just 25 questions.

螢幕真正擅長的其實是評量。因此以我們的 ALEKS 代數產品為例，只要 25 題，我們就能精準告訴你：在學生需要具備的 500 項技能中，他們具備了多少——只用 25 題就能做到。

And so as an organization, we've invested -- we have third parties that assess how we do -- how we develop our curriculum and also how we deliver. We have school Boards that evaluate our outcomes, and then we also do a tremendous amount of testing. We have over 7,000 teachers in the network and over 2,000 students that we use when we are -- we use for testing that we tap into that network before we release a product.

因此以組織的角度，我們已投入——我們有第三方來評估我們的表現——我們如何開發課程，以及我們如何交付。我們有學校董事會評估我們的成果，另外我們也做大量測試。我們的網絡中有超過 7,000 名教師，以及超過 2,000 名學生；在我們發布產品之前，我們會在——我們會用於測試——先連結並運用該網絡來進行測試。

And so McGraw Hill is one of these rare companies that can start all the way at kindergarten with paper and workbooks and reading passages on paper, move into bite-size assessments and then all the way up into learning tools, study tools like Sharpen and higher education and even things like clinical reasoning where we're simulating parents in the medical field. So we have maybe one of the most -- the widest and most diverse ways of delivering reading -- I'm sorry, learning, and it really is a strength that's unmatched, I would say.

因此，McGraw Hill 是少數能從幼兒園階段就以紙本、作業本與紙本閱讀文章起步的公司，接著進入小份量的評量，再一路延伸到學習工具、像 Sharpen 這類的讀書工具、高等教育，甚至到臨床推理這類我們在醫療領域模擬家長的情境。所以我們可能擁有最——最廣且最多元的方式來提供閱讀——抱歉，是學習；而我會說，這確實是一項無可匹敵的優勢。

And Faiza, just to reiterate, should that movement strengthen, we see that as a benefit to us and it which should increase our capture rate over time.

而 Faiza，再重申一次，如果那個趨勢持續增強，我們認為這對我們是利多，並且應會隨時間推進提高我們的取得率。

Understood. And then if I could just follow up on capital allocation and your expectations for cash flow generation in 2027 that -- I know you announced a new buyback, so just any more color there would be helpful.

了解。另外如果我可以再追問一下資本配置，以及你們對 2027 年現金流生成的預期——我知道你們宣布了新的庫藏股回購，所以若能再多提供一些說明會很有幫助。

Sure. So I'd say we're being consistent with a disciplined approach as we think about our capital allocation. And I've always said the first place we deploy capital is on our organic internal opportunities. Those have the greatest ROI. And then we take that excess surplus cash flow and determine where do we place it.

當然。我會說，在思考資本配置時，我們一貫維持紀律性的做法。我一直都說，我們部署資本的第一優先是我們的有機內生成長機會。那些的投資報酬率（ROI）最高。接著我們會把多餘的剩餘自由現金流拿來決定要放到哪裡。

We put it into deleveraging, which we've demonstrated last year with $646 million of paydown and our commitment to the 2 to 2.5 times. We remain committed to that, and we're going to continue to -- as we've demonstrated, to pay down. And then we balance that with M&A.

我們會用於去槓桿化；去年我們已用 6.46 億美元償還債務，並承諾將槓桿倍數維持在 2 到 2.5 倍。我們仍然致力於此，並且會持續——如同我們已經展現的——繼續償還。然後我們也會在此與併購（M&A）之間取得平衡。

And the funnel looks really good. As I look at a couple of opportunities as of Philip and I just yesterday, we think there's some meaningful acceleration of our funnel, of our product development funnel. We think there's new adjacent products that we can acquire into. So I'm optimistic we'll get a couple of deals done this year.

而我們的案源漏斗看起來非常不錯。就我和 Philip 昨天才剛看過的幾個機會而言，我們認為我們的漏斗、我們的產品開發漏斗，正在出現一些有意義的加速。我們認為有一些新的相鄰產品可以透過收購納入。所以我樂觀地認為今年能完成幾筆交易。

And then ultimately, we look at the opportunity for share buyback, and we see it as a way to create value for our shareholders. We have conviction in terms of our long-term outlook, the cash generation of the business as well as our Board. So we'll balance that, and you see that with the opportunistic open market purchase of $50 million that the Board approved. So I would say it remains consistent with that disciplined and balanced approach.

最後，我們會看股票回購的機會，並將其視為為股東創造價值的一種方式。我們對長期展望、業務的現金創造能力，以及我們董事會，都有信心。因此我們會在各項用途之間取得平衡；你也看到董事會核准了 5,000 萬美元的機動性公開市場買回。所以我會說，這仍然符合那種有紀律且均衡的做法。

And then from a free cash flow perspective, I think you should be looking at the K-12 market. That really dictates that cash conversion, 50% to 100% where we sit in that adoption cycle. And given where we see the overall size of K-12 in '27, we're at the lower end of that range.

另外就自由現金流的角度，我認為你應該關注 K-12 市場。那才是真正決定現金轉換率的因素：在我們所處的採用週期位置不同，轉換率會在 50% 到 100% 之間。而考量我們對 2027 年 K-12 整體規模的判斷，我們會落在該區間的較低端。

Jeff Mueller, Baird.

Jeff Mueller，Baird。

As I look at slide 27, the K-12 core purchasing schedules by year, can you give us any sense of how ELA as a percentage of mix evolved by year? And then just with the math, any initiatives or adjustments that you can talk to improve the capture rate?

我看第 27 頁投影片，K-12 核心採購時程按年度排列；你能否讓我們了解一下，ELA 在組合中的占比是如何逐年演變的？另外就數學而言，有沒有任何你可以談的措施或調整，以提升取得率？

Sure. ELA becomes more meaningful as we move on. And I think that also not just as we see California ELA, but if you're looking at lighter gray bar section, as we see that super-cycle being rolled out, ELA becomes a bigger component. And to highlight for us, that's been 40% of our revenue.

當然。隨著我們往後推進，ELA 會變得更具意義。而我認為不只是我們看到加州的 ELA；如果你看較淺灰色長條的區段，隨著那個超級週期逐步推出，ELA 會成為更大的組成部分。特別要強調的是，對我們而言，這部分已占我們營收的 40%。

So we see that becoming more and more important over time. As I highlighted -- as we've said before, 80% of the California math opportunity remains. And so the overall opportunity stays about the same size, about $1 billion market, with 20% of that being called in the first year.

因此我們看到它隨時間變得愈來愈重要。如我強調——也如我們先前所說，加州數學的機會仍有 80% 尚未開發。因此整體機會的規模大致維持不變，約為 10 億美元的市場，其中第一年大約會有 20% 被釋出（招標/採購）。

But what are you doing to improve your capture rate? I get that there's still a lot of opportunity, but it sounds like your capture rate for California math is also lower than you expected. So I guess what are the messaging or curriculum or tech adjustments? Or just how are you going to improve it?

但你們要如何提升取得率？我理解仍有很大的機會，但聽起來你們在加州數學的取得率也低於你們原先的預期。所以我想問，你們在訊息傳達、課程內容或技術上會做哪些調整？或者說，你們要如何改善？

In California, there's a number -- as I mentioned at the start, there's a number of techniques for teaching math. I'll give you an example; open inquiry is one technique where you can talk about walking into a pizza shop and needing to be able to serve enough people with the pieces of pizza. And then there's kind of a linear or spiral approach where you start off talking about fractions and then you really explain fractions and then you work your way into inquiry.

在加州，正如我一開始提到的，數學教學有多種——有多種教學技巧。我舉個例子；開放式探究是一種技巧，你可以談到走進披薩店，需要能用披薩切片來服務足夠多的人。另外也有比較線性或螺旋式的方法：你先從分數開始談，然後把分數講得很清楚，再逐步帶入探究。

The market is very fragmented in California. It's very fragmented around the world right now about how we teach math. In the very first year, what we've seen is that there were approximately 26 different organizations that were bidding on California math and it really represented a wide variety of ways that you can teach math.

加州的市場非常分散。目前全世界在數學教學方式上也非常分散。在第一年，我們看到大約有 26 個不同的機構參與加州數學的投標，這確實代表了各式各樣的數學教學方式。

And so districts -- we watched a lot of districts push off decisions because there was so much confusion and there was not necessarily a consensus of the school board level or among the teachers about which tool is going to represent the way that they wanted to represent their pedagogy. And so while seeing that, we also were able to get a lot of feedback around some of the materials that we had just rolled out. We had accelerated our California math program.

因此各學區——我們看到許多學區延後決策，因為混亂太多，而且在學校董事會層級或教師之間，未必對哪一套工具能代表他們想呈現的教學法（pedagogy）形成共識。而在觀察到這些情況的同時，我們也就一些剛推出的教材收到了大量回饋。我們加速推進了我們的加州數學課程。

I think I also mentioned that over the past, I'd call it 12 to 18 months, we have produced more curriculum than we ever have in our history with some of the biggest programs rolling out, things like California math as well as our ELA program all simultaneously, also having to support multilingual. And so with that being the case, when we rolled out the product, we've been getting feedback among the districts that have been selected us and also the districts that haven't selected us, and we're quickly responding to all of that feedback.

我想我也提到過，在過去我稱之為12到18個月的時間裡，我們推出的課程內容比公司歷史上任何時期都多，且一些最大的專案同時上線，例如加州數學，以及我們的ELA（英語語言藝術）課程，同時也必須支援多語言。因此在產品推出後，我們從選用我們的學區以及未選用我們的學區都收到回饋，我們也在快速回應所有這些回饋。

We're really making sure we're accentuating. If a district wants open inquiry, we're accentuating it even more. If a district does not on open inquiry, we can accentuate it slightly less the way that we deliver our -- the materials. And so we're in -- and this is one of the strengths that I think that really should dwell on with McGraw Hill.

我們確實在確保把重點更凸顯出來。如果某個學區想要開放式探究（open inquiry），我們就更強化這一點。如果某個學區不想要開放式探究，我們也可以在我們交付教材的方式上把它弱化一些。因此我們正在——而這也是我認為麥格羅希爾（McGraw Hill）真正應該著墨的一項優勢。

We're one of the few companies in the world that can actually deliver pedagogy down to the ZIP code or down to the professor level. And our ability to be able to respond at scale, multilingual offerings with pedagogy, with materials, whether or not it's paper, whether or not it's screen time, we're one of the few companies that can respond at this scale.

我們是全球少數幾家能夠把教學法（pedagogy）落實到郵遞區號層級，甚至到教授層級的公司之一。而我們能夠以規模化方式回應、提供結合教學法與教材的多語言方案——不論是紙本、或是螢幕使用時間——我們是少數能以這種規模回應的公司之一。

And so we're in the process of demonstrating our strength as an organization and our strength of understanding the local requirements and then also our ability to be able to produce them. And we're in that process. As 80% is still to be decided, our teams are out working with the districts that haven't decided. We're making sure that our materials are reflective of the exact pedagogy that the local districts need. And so we're excited for the next selling year.

因此我們正在展現我們作為組織的實力、我們理解在地需求的能力，以及我們把它們生產出來的能力。我們正在進行這個過程。在仍有80%尚待決定的情況下，我們的團隊正在與尚未做出決定的學區合作。我們確保我們的教材能反映在地學區所需要的精確教學法。因此我們對下一個銷售年度感到興奮。

And Jeff, I'll highlight, we've demonstrated the ability to course correct in the past. We saw that in Florida social studies a few years back, where we had our first year was slightly below expectations, and we had a nice recovery in year two. So we're following that same playbook, and we've done this before.

Jeff，我也想強調，我們過去已經證明我們有能力修正方向。幾年前在佛羅里達社會科，我們第一年的表現略低於預期，但第二年有不錯的回升。所以我們正在沿用同樣的作法，我們以前做過。

Toni Kaplan, Morgan Stanley.

Toni Kaplan，摩根士丹利（Morgan Stanley）。

I feel like there's a few different factors here that are leading to the lower win rates and capture rates. The California delay, the Texas Bluebonnet and then I guess -- and maybe there are others. But I guess when you think about it, which is the most meaningful -- which is the one easiest to sort of see a reversal of, how long does it take for a state to maybe make a decision and then sort of turn around and say like, well, maybe we should be trying something else instead? Or does it just have to play out?

我覺得這裡有幾個不同因素導致較低的得標率（win rates）與拿下率（capture rates）。加州的延遲、德州的Bluebonnet，然後我想——也許還有其他因素。但我想問的是，當你們思考這些因素時，哪一個最具影響力——哪一個最容易看到反轉？一個州做出決定通常需要多久，然後又要多久才可能回頭說，例如「也許我們應該改試別的」？還是說只能讓它自然發展？

Well, interestingly, what we are seeing is the reaction from teachers in Texas and the response to our product has been very positive. And we're also seeing shorter duration contracts, which means that they're still in that experimentation mode. So we think that that market can turn around in the next couple of years. Fortunately, California still has 80% of the market to be called. So we still have opportunity to course correct. So we think those are two areas.

有趣的是，我們看到德州教師的反應，以及對我們產品的回饋都非常正面。我們也看到合約期限更短，這表示他們仍處於試驗模式。所以我們認為那個市場在未來幾年內有機會扭轉。幸運的是，加州仍有80%的市場尚待定案。因此我們仍有機會修正方向。所以我們認為這是兩個面向。

But I do want to highlight the strength that we are seeing in open territory. We're gaining share. We're winning there. as well as in our supplemental intervention. So I really want us to have a balanced view there and then the short-term opportunity to course correct in both Texas and California math.

但我也想強調我們在開放市場（open territory）看到的強勁表現。我們正在提升市占。我們在那裡也在贏。同時在我們的補充型介入（supplemental intervention）產品上也是如此。

I would add in one of the interesting things -- I would just add one minor thing that I think that some of the districts are starting to work with tools and very quickly realizing they've got diverse learning individuals inside of their district and the lack of accessibility in some cases, whether or not it's with OER materials or whether or not it's in particular tools.

我再補充一個有趣的點——我只補充一個小點：我認為有些學區開始使用一些工具後，很快就意識到他們學區內有多元學習需求的學生，而在某些情況下，無論是OER教材或特定工具，都存在無障礙（accessibility）不足的問題。

Our investment in accessibility is becoming an increasing strength in how we deliver, and so we're seeing organizations that are turning around and saying, we didn't realize this tool or this curriculum didn't have the accessibility that we required. And so, to Bob's point, we see one- to two-year contracts, that, to us, feels like experimentation, and we're going to continue to deliver right materials for the right types of learners.

我們在無障礙方面的投資，正逐漸成為我們交付方式上的一項更強優勢；因此我們看到一些組織回頭說：我們沒意識到這個工具或這套課程沒有我們所要求的無障礙能力。所以，呼應Bob的觀點，我們看到1到2年的合約，對我們而言這像是在試驗，我們會持續為不同類型的學習者提供合適的教材。

Great. And then looking at slide 27, we've got our rulers out trying to do the estimates of the market for the coming years just given no numbers there. So we -- my main question is, it looks like in '28, the all other states outside of the Big 3, pretty flattish with '26 and '27 and maybe even '29 a little flattish as well a little bit of improvement. But like I guess, I think we were expecting before maybe a little bit of an improvement in those other states in '28, '29, and just wondering if there's a reset of expectations for those states as well?

很好。接著看第27頁投影片，因為上面沒有數字，我們都拿尺在估算未來幾年的市場規模。所以——我主要的問題是，看起來在'28年，除了三大州（Big 3）以外的其他州，與'26、'27年相比相當持平，甚至'29年也可能有點持平，只是略有改善。但我想，我們之前可能預期那些其他州在'28、'29年會有更明顯的改善；想請問對那些州的預期是否也有下修？

No, there was no meaningful reset, any meaningful change. The most notable change we highlighted was that the $100 million primarily in California math move from '27 into '28, but no major change.

沒有，沒有任何實質性的下修或重大變動。我們強調的最顯著變化是主要在加州數學的1億美元，從'27年移到'28年，但沒有其他重大變化。

Yeah. The only area that we're watching closely as we head into this buying season is the ELA market, as mentioned. Literacy is greater focus than ever in the United States. And so we're watching that, and we'll convey that what we see.

是的。我們在進入這個採購季時，唯一密切關注的領域是如先前提到的ELA市場。識字／讀寫能力（literacy）在美國受到的重視程度前所未有。因此我們會持續觀察，並把我們看到的情況對外說明。

And Toni, just to reiterate, we're in a growing market. We have clear visibility. This is a predictable market. We see that growth in the near and medium term, and that really excites us.

Toni，我再重申一次，我們身處一個成長中的市場。我們有清楚的能見度。這是一個可預測的市場。我們看到短期與中期的成長，而這讓我們非常振奮。

Shlomo Rosenbaum, Stifel.

Shlomo Rosenbaum，Stifel。

Bob, I just want to start with the housekeeping one. I wasn't clear as to how to think about the levered free cash flow expectation for this year. I think you said something like unlevered $400 million. Should I take $400 million minus $180 million to get to $220 million, which sounds a little bit light given the one-time items. I thought there were a lot more like one-time items due to the IPO that we should, in theory, see better free cash flow than that in fiscal year '27. Am I thinking about that properly?

Bob，我先從一個例行性的問題開始。我不太清楚今年槓桿後自由現金流（levered free cash flow）的預期該怎麼理解。我記得你說過類似未槓桿自由現金流（unlevered）4億美元。我是否應該用4億減去1.8億得到2.2億美元？考量到一次性項目，這聽起來有點偏少。我原本以為因IPO帶來的一次性項目更多，因此理論上在'27財年我們應該會看到比這更好的自由現金流。我的理解正確嗎？

Shlomo, you are. That's consistent with how you should be thinking about it.

Shlomo，你的理解是正確的。你應該就是用那樣的方式來思考。

Okay. So $220 million is a levered free cash flow that that we should be thinking about?

好的。所以我們應該把2.2億美元視為槓桿後自由現金流的水準？

Yes. Correct.

是的。沒錯。

Okay. Then just a little bit broader. I just wanted to ask in the context of AI and the protections that you have around your data. There's that article where the educational curriculum providers are suing Meta for training the LLMs on your data on copyright infringement.

好的。那再問一個更宏觀的問題。我想在AI以及你們對資料的保護措施脈絡下提問。有一篇文章提到教育課程供應商因Meta以你們的資料訓練大型語言模型（LLM）而提起訴訟，主張其構成著作權侵害。

Is that like a longer-term issue for you guys in your opinion in terms of the data actually being out there and it's actually accessible so it's not like behind certain paywalls that other information services providers have to have a little bit more protection on that side? How should we just think about that kind of lawsuit in the context of your ability to control your data?

就你們的看法而言，這對你們來說是否像是一個較長期的問題：也就是資料實際上已經在外面，而且確實可被取得，因此不像某些其他資訊服務供應商那樣被特定付費牆擋住、需要在那一端做更多保護？我們應該如何在你們掌控資料能力的脈絡下，來看待這類訴訟？

Great question. I want to really be clear about this, our position, first and foremost, is that great AI and copyrights are not counter to each other. You can enforce and support copyrights and respect copyrights and have great AI for the world. And the agentic technologies I talked about are really, really great examples technically of being able to do this.

好問題。我想把這點說得非常清楚：我們的立場首先是，優秀的 AI 與著作權並不互相對立。你可以執行並支持著作權、尊重著作權，同時也能為世界帶來優秀的 AI。而我提到的 agentic 技術，從技術面來看就是非常、非常好的例子，說明如何做到這一點。

Yes, our content has been and always will be behind paywalls. And the cases where it's not, it's content that's been pirated, was sitting out on pirate sites, and we are very active in litigation against pirate websites. We really protect our content.

是的，我們的內容一直、也將永遠會在付費牆之後。而那些不在付費牆後的情況，是內容遭到盜版、被放在盜版網站上；我們也非常積極對盜版網站提起訴訟。我們確實非常保護我們的內容。

And when the content gets out there, we expect that other organizations are not going to use these pirated content, basically illegal content. And so part of that lawsuit that we were really clear about is that we intend to continue to protect our content. There are many, many, many ways for organizations, large AI companies to license content.

而當內容流出到外面時，我們期待其他組織不會使用這些盜版內容，也就是基本上屬於非法的內容。因此，我們在那起訴訟中非常清楚地表明：我們打算持續保護我們的內容。對各種組織、尤其是大型 AI 公司而言，有非常非常多方式可以去取得內容授權。

There's been great examples of Getty Images and New York Times. When I was at Vimeo, we were approached in the early days about potential delight about using all of our content, the video content, the 8 billion minutes of video we had, and we turned down the organizations originally. And then by the end of my tenure of Vimeo, people were coming back and asking just for videos of beach balls to make sure their video model bounce the beach balls right or the hands clap properly.

Getty Images 和《紐約時報》就是很好的例子。我在 Vimeo 任職時，早期就有人來接觸我們，對方對使用我們所有內容（影片內容、我們擁有的 80 億分鐘影片）表達了很大興趣，但我們一開始拒絕了那些組織。而在我任期結束前，人們又回來詢問，但只想要像是海灘球的影片，用來確認他們的影片模型能否把海灘球彈跳做對、或把拍手的手勢做得正確。

So there's a well-established market licensing content to train AI models. It's both technically very easy and financially well-established business models for this. One of the things that we were stating in that Meta lawsuit is that we really want AI companies to use these legal and technically simple ways of licensing content.

因此，市場上早已存在一個成熟的內容授權市場，用於訓練 AI 模型。這在技術上非常容易，在財務上也有成熟、可行的商業模式。我們在那起對 Meta 的訴訟中所表達的一點是：我們非常希望 AI 公司採用這些合法且技術上簡單的方式來取得內容授權。

The work we're doing with agentic, we're really excited about. And quite frankly, even if existing content might be out on pirate sites, human's knowledge and the way to teach changes every single day. And so we intend to be this great place that takes human knowledge, turns it into a pedagogy and delivers it the way you need to.

我們正在做的 agentic 相關工作，讓我們非常興奮。坦白說，即使既有內容可能出現在盜版網站上，人類的知識以及教學方式每天都在改變。因此，我們打算成為一個很棒的平台：把人類知識轉化為教學法，並以你所需要的方式交付。

And with agentic technologies, we can meter that, we can secure it, we can build an interface, we can audit to make sure that we're charging the right way. We can actually explain to every student or every professor how much content is being used at any point in time.

而透過 agentic 技術，我們可以進行用量計量（meter）、可以加以保護、可以建立介面，也可以稽核以確保我們的收費方式正確。我們甚至可以向每位學生或每位教授說明，在任何時間點使用了多少內容。

And so the technology models are evolving quickly. We're going to be at the forefront of that in the knowledge economy. And we're going to be really, really, I'll say, aggressive about making sure we're protecting and ensuring that AI does -- that AI companies are working with us and all publishers in the right way.

因此，技術模型正在快速演進。我們會站在知識經濟的最前沿。而我們也會非常、非常——我會說——積極地確保我們在保護自身權益的同時，也確保 AI 的運作——確保 AI 公司以正確的方式與我們以及所有出版商合作。

Henry Hayden, Rothschild & Co Redburn.

Henry Hayden，Rothschild & Co Redburn。

I firstly wanted to continue the conversation around kind of agentic curriculum, MCP integration and licensing agreements. This feels like a shift in kind of your philosophy around ownership of user interface, content distribution. So just curious as to any incremental color you could give around that.

我首先想延續剛才關於 agentic 課程、MCP 整合與授權協議的討論。這讓人感覺你們在使用者介面所有權、內容分發方面的理念出現了某種轉變。所以想請教你能否就此提供一些額外的說明。

And then on the agentic side, as you think about kind of the ramp-up of additional partnerships, how should we think about what cost that would present to end customers and what the margin profile for McGraw Hill looks like?

另外在 agentic 這一端，當你們思考新增合作夥伴的加速導入時，我們應該如何看待這會為終端客戶帶來的成本，以及 McGraw Hill 的毛利結構會是什麼樣子？

Yeah. We're really -- thanks very much for the question. So on the agentic side, we live in a trusted position with professors, with students, with teachers and with administrations. When we take something like Sharpen -- our Sharpen tool, not just takes content and delivers podcasts or quizzes, but we also are turning around and able to tell the professor what content is being engaged with. Professors are going to be able to deliver, I'll call it, play lists of content.

是的。我們非常——謝謝你的提問。在 agentic 這一端，我們在教授、學生、教師以及行政單位之間處於一個被信任的位置。當我們使用像 Sharpen 這樣的東西——我們的 Sharpen 工具不只是把內容轉成 Podcast 或測驗並交付出去，我們也能反過來告訴教授：哪些內容正在被使用、被互動。教授也將能夠提供——我姑且稱之為——內容播放清單（play lists）。

And so it's not just one-sided where we're serving to an AI agent. In some cases, we are the AI agent the student trusts or the professor trusts. And so we intend to really, really support that trust and be an agent of the student and the teacher from a first-party perspective uses.

因此，這不只是單向地把內容提供給某個 AI agent。在某些情況下，我們就是學生信任的 AI agent，或是教授信任的 AI agent。因此，我們打算非常、非常地維護這份信任，並以第一方（first-party）的角度，成為學生與教師的代理者（agent）。

Now in some cases, those students or those professors want us to join a course of agents or course of knowledge. And so this ability to be MCP, to provide our content via MCP is really exciting. One of the organizations wants to use us with Notebook LM, with Google Notebook LM, but they want a trusted source because today, if you just use AI, it can hallucinate on all kinds of content or quite frankly, it can hallucinate -- it can come off -- it can bring answers out at the professor hasn't even introduced into the class.

而在某些情況下，這些學生或教授希望我們加入一個 agents 的課程體系或知識體系。因此，具備 MCP 能力、能透過 MCP 提供我們的內容，這點非常令人興奮。有個組織希望在 Notebook LM、Google Notebook LM 中使用我們，但他們需要一個可信任的來源，因為今天如果你只是使用 AI，它可能會在各種內容上產生幻覺；或者坦白說，它可能會——它可能會把教授甚至還沒在課堂上介紹的答案先帶出來。

There's the right time to introduce the chemistry behind photosynthesis. And it's not the start of a class or give you the full answer before a student really truly understands biology and the chemistry behind it. And so our ability to be able to do that in a pedagogical way is also part of MCP.

例如，介紹光合作用背後的化學原理是有適當時機的。不應該在一堂課一開始就講，或在學生尚未真正理解生物學及其背後的化學之前，就把完整答案直接給出來。因此，我們能以教學法（pedagogical）的方式做到這件事，也是 MCP 的一部分。

And so this is an evolution. It's not a big shift from anything that we've already been doing. And what's exciting about it is, again, MCP gives us an ability to be able to monitor it to be able to interface. We're excited about the ability to be able to be a continuing source of education.

所以這是一種演進。它並不是對我們既有作法的重大轉向。而令人興奮的是，再次強調，MCP 讓我們能夠監控、能夠建立介面。我們也很期待能夠成為持續性的教育來源。

I like many people actually carried my textbooks with me throughout really early years as a computer scientist. I was referring back to them for a long time. I think you'll hear that from legal scholars as well as accounting scholars, medical scholars. There's a lot of reference that you do, especially early in your career and quite frankly, in some cases, even beyond the early years.

我跟很多人一樣，其實在我作為電腦科學家很早期的那些年，會把教科書一直帶在身邊。我很長一段時間都會回頭查閱它們。我想你也會從法律學者、會計學者、醫學學者那裡聽到同樣的情況。尤其在職涯早期，你會做大量的參考；坦白說，有些情況甚至在早期之後也仍然如此。

Now when I think about kind of where this goes longer term, I do think that this is an exciting opportunity for us to, as I mentioned, to expand TAM in the marketplace. And I do think that pricing is going to evolve pretty dramatically.

而當我思考這件事長期會走向哪裡時，我確實認為這對我們而言是一個令人興奮的機會——如我所提——去擴大市場中的 TAM（總可服務市場）。我也確實認為定價將會相當劇烈地演變。

The area that I'm really, really excited about -- I think that there's a misnomer and I've described this to a lot of people. We are in the era of big models. These are -- I call this peak inefficiency in AI. I don't think in the history of computing, I've ever seen that we need to literally use the United States GDP to train an algorithm. It's not a good business model right now, I would tell you.

我真正、真正感到興奮的領域是——我認為這裡存在一個迷思，我也跟很多人這樣描述過。我們正處於大型模型的時代。這些——我把這稱為 AI 的效率高峰期（peak inefficiency）。我不認為在運算史上，我曾看過我們竟然需要字面上動用美國的 GDP 來訓練一個演算法。我會告訴你，這在目前並不是一個好的商業模式。

And what we're able to do, I'm a big believer that we're going to move into an era of smaller models, more purpose-built models that are much more efficient, can be much more reactive, can actually hit the performance requirements that are necessary and also lower cost.

而我們能做到的是——我非常相信——我們將進入一個小型模型的時代，也就是更具特定用途、效率更高的模型；它們能更快速反應，實際上也能達到必要的效能要求，同時成本更低。

As an example, we'll take our agentic technologies where we're creating vectorization of our actual content, coupled together with a small lightweight model and we'll get a fraction of the cost of having to run things through a large model. And that's where we expect it to move to.

舉例來說，我們會採用我們的代理式（agentic）技術，把我們實際內容做向量化，並搭配一個小型輕量模型，如此一來，成本只需要把事情丟進大型模型去跑的一小部分。而這就是我們預期會走向的方向。

I really encourage you to look at technologies like knowledge graphs where organizations are getting much better performance, especially there's one example in the financial services industry where it's literally 1% of the cost of a large model and it's achieving way higher than expertise level than the big models.

我也非常鼓勵你去看看像知識圖譜（knowledge graphs）這類技術，許多組織能取得更好的效能；尤其在金融服務業有一個案例，成本只有大型模型的1%，但達到的專業水準卻遠高於大型模型。

And so this idea of cost, what we do inside of McGraw Hill, is that we use the right model along with our vectorized content and that we deliver the right question, the right content at the right time at a much lower cost than throwing all the content in with, I call it, everyone at Taylor Swift songs and all the F1 drivers and the predictions for the next Spur series to have to wait through just to be able to describe what photosynthesis is.

因此，談到成本這個概念——我們在McGraw Hill內部的做法是：使用合適的模型搭配我們向量化的內容，在正確的時間提供正確的問題與正確的內容；其成本遠低於把所有內容都丟進去——我把它稱作，還得把泰勒絲的所有歌曲、所有F1車手，以及下一部《Spur》系列的預測都一起篩過——就只是為了能描述什麼是光合作用。

So cost is high with big models. Purpose-built lightweight models combined together with vectorized content, I think, is going to be a way more efficient delivery mechanism. And then we do the cost management on behalf of the student as well as the university. And our buying capacity for AI is much higher than most universities and certainly students. And so we also think we have an opportunity to be able to really be a, I'll say, a more effective, more cost-effective and more pedagogically effective solution.

所以大型模型的成本很高。我認為，特定用途的輕量模型結合向量化內容，將會是效率高得多的交付機制。接著，我們也會代表學生與大學進行成本管理。而我們在AI方面的採購能力遠高於多數大學，當然也高於學生。因此我們也認為，我們有機會成為——我會說——更有效、更具成本效益，且在教學法上更有效的解決方案。

That's very helpful. And then my follow-up question was just a quick one on K-12. Have you seen any kind of -- you mentioned that your discussions around supplemental intervention, you think it's bottomed out. How is the tenor of those sales discussions going? Like as you move through the renewal cycle, are you seeing any pressure on funding kind of impacting you on a go-forward basis?

這非常有幫助。那我接著的追問是關於K-12的一個簡短問題。你提到你們對補充性介入（supplemental intervention）的討論，你認為已經觸底。這些銷售討論的氛圍目前如何？例如在續約週期推進的過程中，你們是否看到資金方面的壓力，可能在未來對你們造成影響？

We have not seen any pressure from funding, as we've highlighted. We know that learning loss is real. What we have experienced is the connected classroom, that is resonating. There's too many tools; there's tool fatigue. And so as we're out in the market, being able to sell core and supplemental intervention solutions together is really resonating.

我們沒有看到任何來自資金面的壓力，如同我們先前強調的。我們知道學習流失（learning loss）是真實存在的。我們所看到的是「連結式教室」（connected classroom）這個概念正在引起共鳴。工具太多了；大家出現工具疲乏（tool fatigue）。因此當我們在市場上推進時，能夠把核心（core）與補充性介入解決方案一起銷售，確實非常能引起共鳴。

Josh Chan, UBS.

Josh Chan，瑞銀（UBS）。

I'm just going to stick to one just for the sake of time. So I guess, I know Texas is a completely different story. So maybe set that aside. But if you look at the California math dynamic, how do you think about that issue being strictly limited to California and strictly limited to the math, I guess?

為了節省時間，我就只問一個問題。我想，我知道德州完全是另一回事。所以也許先把那個放一邊。但如果你看加州數學的動態，你怎麼看待這個問題是否嚴格只限於加州、也嚴格只限於數學呢？

I think, I would say the fragmentation is not limited to California or just a math. When you get into teaching social studies in the United States on a global basis, extraordinarily fragmented, science as well. So there is fragmentation in the education industry. And I'll say again, one of the most exciting things since I've arrived here is understanding that we're one of the few companies in the world that can respond to this fragmentation at scale.

我認為——我會說——這種碎片化並不只限於加州或只限於數學。當你在美國以全球視角來看社會科教學時，碎片化程度極高；科學也是如此。所以教育產業確實存在碎片化。我再說一次，自從我來到這裡以來，最令人興奮的一件事就是理解到：我們是全球少數能以規模化方式回應這種碎片化的公司之一。

And to respond to it, you have to understand it, first of all. You've got to understand down at the ZIP code level, the teacher level, even the professor level inside of the higher ed. In the medical industry, in some countries, we teach body parts; in other cases, we teach symptoms; in some cases, we're teaching full systems. So fragmentation in education is I'll say, it's been around as long as I would say education has been around.

而要回應它，首先你必須理解它。你必須在郵遞區號（ZIP code）層級、教師層級，甚至在高等教育體系內的教授層級去理解。在醫學領域，在某些國家我們教的是身體部位；在其他情況下，我們教的是症狀；有些情況下，我們教的是完整系統。所以教育的碎片化——我會說——它存在的時間幾乎和教育本身一樣久。

What I would say is that as we look across the spectrum, our ability to be able to respond to that is unmatched. And so I look at this opportunity in California as another opportunity in the same way that we've responded and demonstrated our ability to respond to the different ways of teaching science. You're seeing us in real time, be able to respond to California with the strength of McGraw Hill's relationships, our tools, our variety of distribution means and then also just our ability to be able to do different pedagogies for different students. So it's our strength.

我想說的是，當我們放眼整個光譜，我們回應這件事的能力是無可匹敵的。因此我把加州的這個機會視為另一個機會，就像我們過去回應並證明我們能回應不同科學教學方式一樣。你正在即時看到我們能夠回應加州：憑藉McGraw Hill關係網的強度、我們的工具、我們多元的發行／分銷方式，以及我們能為不同學生採用不同教學法的能力。所以這是我們的強項。

We have reached the end of the Q&A session. This concludes today's call. Thank you for attending. You may now disconnect.

我們已到問答環節的結束。今天的電話會議到此結束。感謝各位參與。您現在可以斷線。