Data Storage Corp (DTST) 2026 Q1 法說會逐字稿

Greetings and welcome to the Data Storage Corporation first-quarter 2026 earnings conference call. (Operator Instructions) As a reminder, this conference is being recorded.

各位好，歡迎參加 Data Storage Corporation 2026 年第一季財報電話會議。(接線員指示) 提醒各位，本次會議將被錄音。

It is now my pleasure to introduce David Waldman, Investor Relations. Thank you. You may begin.

現在我很榮幸介紹投資人關係部的 David Waldman。謝謝。您可以開始了。

Thank you, and good morning, everyone. Welcome to Data Storage Corporation's 2026 first-quarter business update conference call.

謝謝，各位早安。歡迎參加 Data Storage Corporation 2026 年第一季業務更新電話會議。

On the call with us this morning are Chuck Piluso, Chairman and Chief Executive Officer; and Chris Panagiotakos, Chief Financial Officer. The company issued a press release this morning containing its 2026 first quarter financial results, which is also posted on the company's website. If you have any questions after the call or would like any additional information about the company, please contact Crescendo Communications at 2126711020.

今天早上與我們一同出席電話會議的有：董事長兼執行長 Chuck Piluso，以及財務長 Chris Panagiotakos。公司今早發布新聞稿，內容包含 2026 年第一季財務結果，亦已張貼於公司網站。若您在會後有任何問題或希望取得公司更多資訊，請聯絡 Crescendo Communications，電話 2126711020。

Before we begin, please note that today's call contains forward-looking statements within the meaning of the Private Securities Litigation Reform Act of 1995. Actual results may differ materially due to various risks and uncertainties described in the company's filings with the SEC. Except as required by law, the company assumes no obligation to update or revise forward-looking statements.

在開始之前，請注意今天的電話會議包含《1995 年私人證券訴訟改革法》所定義的前瞻性陳述。實際結果可能因公司向 SEC 提交文件中所述的各種風險與不確定性而出現重大差異。除法律要求外，公司不承擔更新或修訂前瞻性陳述之任何義務。

I'd now like to turn the call over to Chuck Piluso. Please go ahead, Chuck.

接下來我想把電話會議交給 Chuck Piluso。Chuck，請開始。

Thank you, David. Good morning, everyone. We appreciate everyone joining us today. The first quarter of 2026 marked another important milestone in the strategic transformation of Data Storage Corporation. Over the past year, we have repositioned the company following the successful sale of our cloud solution business in 2025.

謝謝你，David。各位早安。感謝大家今天加入我們。2026 年第一季是 Data Storage Corporation 策略轉型的另一個重要里程碑。在過去一年中，隨著我們於 2025 年成功出售雲端解決方案業務，我們已重新定位公司。

And today, we are operating from a position of financial strength, strategic flexibility and operational focus. As many of you know, the sale of the CloudFirst business was transformational for Data Storage Corporation. That transaction not only validated the value we created over more than 2 decades, but also provided us with the capital foundation necessary to reposition the company towards what we believe are significantly larger long-term market opportunities.

而今天，我們是在財務實力、策略彈性與營運聚焦的基礎上運作。如各位所知，出售 CloudFirst 業務對 Data Storage Corporation 具有轉型意義。該交易不僅驗證了我們在超過 20 年間所創造的價值，也為我們提供了必要的資本基礎，使公司得以重新定位，朝向我們認為長期而言規模顯著更大的市場機會前進。

Following the transaction, we completed a substantial tender offer that reduced our outstanding shares count by approximately 72%, while still maintaining debt-free balance sheet and substantial liquidity. Importantly, the period following the sale was not a pause in activity. It was a period of evaluation of analysis of strategic development. We spent considerable time assessing emerging infrastructure trends, regulatory developments, competitive positioning and areas where we believe meaningful structural market gap existed. What became increasingly clear (technical difficulty) experimentation into mission-critical software deployment environments.

交易完成後，我們完成了一項大規模要約收購，使流通在外股數約減少 72%，同時仍維持零負債的資產負債表與充足流動性。重要的是，出售後的期間並非活動停滯。那是一段評估、分析與策略發展的時期。我們投入大量時間評估新興基礎設施趨勢、監管發展、競爭定位，以及我們認為存在具意義之結構性市場缺口的領域。愈來愈清楚的是（技術困難）在關鍵任務軟體部署環境中的實驗。

Across industries such as health care, financial service, insurance, organizations are beginning to deploy Sovereign AI and AI factory environments. On-site equipment, designed to run proprietary AI models on highly sensitive data sets. These are not public AI cloud environments. These are private enterprise-grade AI infrastructures that organizations increasingly rely upon for core operating workflows, security, decision-making, compliance functions and customer-facing processes. As we study this market, we identify what we believe is a critical infrastructure gap.

在醫療保健、金融服務、保險等產業中，組織正開始部署主權 AI（Sovereign AI）與 AI 工廠環境。這些是現場設備，旨在以高度敏感的資料集運行專有 AI 模型。這些並非公共 AI 雲端環境。這些是私有、企業級的 AI 基礎設施，組織日益依賴其支撐核心營運工作流程、安全、決策、合規職能以及面向客戶的流程。在研究此市場時，我們辨識出一個我們認為至關重要的基礎設施缺口。

As these systems are deployed, today, we believe there are no widely adopted purpose-built platforms designed specifically addressing recovery, resilience, behavior validation and regulatory compliance to these AI factory environments.

隨著這些系統被部署，我們認為目前尚無被廣泛採用、且專為 AI 工廠環境設計的平台，能夠特別針對復原、韌性、行為驗證與監管合規等需求。

After two successful decades operating CloudFirst, we understand the clients' requirements as it relates to meeting their expectations surrounding business continuity. Traditional Data Storage systems focus primarily on restoring hardware or infrastructure uptime, but AI introduces an entirely different challenge that. Enterprises will require a business continuity service and will increasingly need to validate those models of behaving correctly when a situation occurs.

在成功營運 CloudFirst 的二十年後，我們了解客戶在滿足其對業務持續性期待方面的需求。傳統資料儲存系統主要著重於恢復硬體或基礎設施的正常運作時間，但 AI 帶來的是完全不同的挑戰。企業將需要一項業務持續性服務，並且在事件發生時，將愈來愈需要驗證這些模型是否表現正確。

That output remains compliant, that inference consistency is maintained and that recovery procedures themselves satisfy the client and regulatory standards. We believe this creates a significantly new category of infrastructure need. To address this opportunity, we plan to establish Sovereign AI Solutions, a wholly owned subsidiary focused on developing what we describe as an AI continuity control plane for regulated enterprises.

也就是確保輸出仍符合規範、推論一致性得以維持，且復原程序本身能滿足客戶與監管標準。我們相信這將形成一個顯著的新型基礎設施需求類別。為把握此機會，我們計畫成立 Sovereign AI Solutions，一家全資子公司，專注於為受監管企業開發我們所稱的「AI 持續性控制平面」。

Our intention is to create a platform capable of serving as resiliency, recovery, validation and compliance layer for Sovereign AI infrastructure environments. The platform we envision is designed to detect behavioral anomalies, execute validated recovery sequences and generated audit-ready documentation that regulated industries may increasingly require as AI becomes embedded into critical business operations.

我們的目標是打造一個平台，可作為主權 AI 基礎設施環境的韌性、復原、驗證與合規層。我們構想的平台旨在偵測行為異常、執行經驗證的復原序列，並產生可供稽核的文件；隨著 AI 被嵌入關鍵業務營運，受監管產業將愈來愈需要此類文件。

Importantly, we believe our approach is differentiated because it focuses not only on infrastructure restoration, but also on preserving operational integrity, compliance posture at the model and behavioral levels. We also believe the market timing is compelling. Earlier this month, several leading AI developers announced multibillion-dollar initiatives designed to integrate AI deeply into the enterprise-wide workflows, further validating large-scale AI deployment across mission-critical environments is accelerating rapidly. While this market remains early stage and rapidly evolving, we believe long-term opportunity could be substantial.

重要的是，我們認為我們的方法具差異化，因為它不僅著重於基礎設施的恢復，也著重於在模型與行為層面維持營運完整性與合規狀態。我們也認為市場時機相當具吸引力。本月稍早，數家領先的 AI 開發商宣布數十億美元規模的計畫，旨在將 AI 深度整合至企業全域工作流程，進一步驗證在關鍵任務環境中進行大規模 AI 部署正在快速加速。雖然此市場仍處於早期且快速演進，我們相信長期機會可能相當可觀。

Based on our preliminary analysis, regulatory-driven enterprise AI infrastructure spending could ultimately represent a multibillion-dollar annual market opportunity. At the same time, we are not currently aware of any other purposely built platform, targeting compliance-driven AI recovery for regulated enterprises in the manner we are pursuing.

根據我們的初步分析，受監管驅動的企業 AI 基礎設施支出，最終可能代表每年數十億美元的市場機會。同時，就我們所知，目前尚無其他專門打造的平台，以我們所追求的方式，針對受監管企業提供以合規為導向的 AI 復原。

Our focus throughout 2026 will be advancing the platform architecture, redefining our go-to-market strategy, continuing industry engagement discussions and progressing towards potential initial customer opportunities. We expect to provide additional commercial and operational updates as these initiatives advance throughout the year. At the same time, our Nexxis business continues to provide an important operational and financial foundation for DTST.

我們在 2026 年的重點將是推進平台架構、重新定義市場進入策略、持續進行產業互動討論，並朝向潛在的初始客戶機會邁進。隨著這些計畫在今年持續推進，我們預期將提供更多商業與營運更新。同時，我們的 Nexxis 業務仍為 DTST 提供重要的營運與財務基礎。

Nexxis remains a stable recurring revenue business, delivering VoIP dedicated Internet access, SD-WAN and data transport services. During the first quarter of 2026, Nexxis sales increased 10.9% year-over-year, while gross profit increased 32.1% and gross margins expanded to 53.7% compared to 45% in the prior period. We believe these results demonstrate both the continued demand for our connectivity services and operational discipline within the business.

Nexxis 仍是一項穩定的經常性收入業務，提供 VoIP、專線網際網路接取、SD-WAN 與資料傳輸服務。在 2026 年第一季，Nexxis 銷售額年增 10.9%，毛利年增 32.1%，毛利率擴大至 53.7%，相較前一期間為 45%。我們認為這些結果展現了市場對我們連線服務的持續需求，以及該業務的營運紀律。

Just as importantly, Nexxis provides us with a recurring revenue base and operating infrastructure that supports our broader strategic initiatives. Financially, we believe DTST is well positioned relative to many companies pursuing emerging technology opportunities.

同樣重要的是，Nexxis 為我們提供經常性收入基礎與營運基礎設施，以支持我們更廣泛的策略計畫。在財務面，我們認為相較於許多追求新興科技機會的公司，DTST 的定位相當有利。

We ended the year with no long-term debt, substantial working capital, significant market securities and a highly flexible balance sheet. That strength gives us the ability to remain patient, strategic disciplined on how we allocate capital while SaiS remains our primary strategic initiatives.

我們在年底時沒有長期負債，擁有充足的營運資金、可觀的有價證券，以及高度彈性的資產負債表。這樣的實力使我們能夠保持耐心並以策略紀律分配資本，同時 SaiS 仍是我們的主要策略計畫。

We are also continuing to evaluate complementary opportunities, including partnerships, strategic investments, mergers and acquisitions and other transactions that could strengthen our competitive position and enhance long-term shareholder value. Ultimately, our goal is to position DTST at the intersection of enterprise AI infrastructure, resiliency, compliance and mission-critical continuity areas where we believe demand will continue to expand significantly over the coming years.

我們也持續評估互補性機會，包括合作夥伴關係、策略性投資、併購以及其他可能強化我們競爭地位並提升股東長期價值的交易。最終，我們的目標是讓 DTST 站在企業 AI 基礎設施、韌性、合規與關鍵任務持續性等交會點上；我們相信未來數年這些領域的需求將持續顯著擴大。

We appreciate the continued support and confidence of our shareholders, and we look forward to updating everyone on our progress as we move throughout 2026.

我們感謝股東持續的支持與信任，並期待在邁入 2026 年的過程中，向各位更新我們的進展。

And I'd like to turn it over to Chris Panagiotakos for a review of the financial results. Chris?

接下來我想把時間交給 Chris Panagiotakos，請他回顧財務結果。Chris？

Thank you, Chuck. Good morning, everyone. As previously discussed on September 11, 2025, we closed the sale of our CloudFirst business for $40 million. As a result of the transaction and in accordance with auditing and reporting standards, our ongoing financial reporting now reflects only our continuing operations, specifically our Nexxis subsidiary. Sales from continuing operations were $347,000 for the three months ended March 31, 2026, an increase of $34,000 or 10.9% compared to $313,000 in the prior year.

謝謝你，Chuck。各位早安。如同先前於 2025 年 9 月 11 日所討論，我們已以 4,000 萬美元完成出售 CloudFirst 業務。由於該交易，並依據審計與申報準則，我們目前的持續性財務報導僅反映持續營運，亦即我們的 Nexxis 子公司。截至 2026 年 3 月 31 日止三個月，持續營運之銷售額為 347,000 美元，較前一年度的 313,000 美元增加 34,000 美元，增幅 10.9%。

The increase was primarily attributable to continued growth in our Nexxis Voice and Data Solutions business, driven by the addition of new customers and increased spending from existing customers.

該增長主要歸因於 Nexxis 語音與數據解決方案業務的持續成長，成長動能來自新增客戶以及既有客戶支出增加。

Revenue growth during the period reflects continued demand for our voice and data connectivity solutions and expansion of services within our existing customer base. Gross profit for the three months ended March 31, 2026, was $186,000, an increase of $45,000 or 32.1% compared to $141,000 in the prior period. Selling, general and administrative expenses for the three months ended March 31, 2026, increased $615,000 or 71.8% to $1.5 million from $857,000 for the three months ended March 31, 2025.

本期營收成長反映市場對我們語音與數據連線解決方案的持續需求，以及在既有客戶基礎內擴大服務範圍。截至 2026 年 3 月 31 日止三個月，毛利為 186,000 美元，較前期的 141,000 美元增加 45,000 美元，增幅 32.1%。截至 2026 年 3 月 31 日止三個月，銷售、一般及行政費用較截至 2025 年 3 月 31 日止三個月的 857,000 美元增加 615,000 美元（增幅 71.8%）至 150 萬美元。

The increase was primarily driven by a $425,000 or 311% increase in noncash stock-based compensation as a result of grants to certain employees during the three months ended March 31, 2026. Professional fees increased by $135,000 or 73.6% attributable to higher fees paid relating to legal and consulting services during the period.

該增加主要由非現金的股份基礎給付增加所帶動；截至 2026 年 3 月 31 日止三個月，因對部分員工授予獎勵，非現金股份基礎給付增加 425,000 美元（增幅 311%）。專業費用增加 135,000 美元（增幅 73.6%），主要歸因於本期支付較高的法律與顧問服務費用。

Net loss attributable to common shareholders for the three months ended March 31, 2026, was $631,000 compared to net income of $24,000 for the three months ended March 31, 2025. We ended the quarter with cash, cash equivalents and marketable securities of approximately $9.7 million at March 31, 2026. We used $29.5 million of the proceeds from the sales of marketable securities to repurchase common stock from our shareholders in connection with the tender offer, which closed on January 15, 2026.

截至 2026 年 3 月 31 日止三個月，歸屬於普通股股東的淨損為 631,000 美元；相較之下，截至 2025 年 3 月 31 日止三個月為淨利 24,000 美元。截至 2026 年 3 月 31 日，我們期末現金、約當現金及有價證券約為 970 萬美元。我們使用出售有價證券所得款項中的 2,950 萬美元，於要約收購（tender offer）中向股東回購普通股；該要約收購已於 2026 年 1 月 15 日結束。

Thank you. And I will now turn the call back to Chuck.

謝謝。現在我把電話會議交回給 Chuck。

Thanks, Chris. Let's open up the call for some questions.

謝謝，Chris。我們來開放提問。

Thank you. And at this time, we'll conduct a Q&A session. (Operator Instructions)

謝謝。接下來我們將進行問答環節。（接線員指示）

Matthew Galinko, Maxim Group.

Matthew Galinko，Maxim Group。

Hey, good morning. Thanks for taking my question. As you pursue the AI strategy. I'm curious how you'll pursue, I guess, developing technical solutions to support the go-to-market. Do you expect to bring developers in-house to the current structure? Or just curious how you'll approach that?

嗨，早安。謝謝讓我提問。在你們推進 AI 策略的同時，我想了解你們將如何推動、也就是說，開發技術解決方案來支援市場推廣（go-to-market）。你們預期會在現有架構下把開發人員納入內部團隊嗎？或是想了解你們會如何著手？

Good morning, Matt. What we're doing right now is that just covered across the board essentially is that we have a recruiter working on finding us someone to run the subsidiary. We are hopefully lining up CTOs that we can interview that may want to start off as a consulting basis and handle the overall project.

早安，Matt。我們目前正在做的事情，基本上是全面性地：我們已請一位招募顧問協助尋找能夠負責子公司營運的人選。我們也希望安排一些 CTO 候選人進行面談，他們可能願意先以顧問形式開始，並負責整體專案。

We're talking to three -- four other companies essentially that want to participate in everything from a subcontracting to them to partnerships for them to do the installation. We came across this because we put out a letter of intent to a company and found out a while ago about Sovereign AI and looking into this and seeing where the holes are.

我們也正在與三到四家公司洽談，基本上希望他們能參與從分包、到合作夥伴關係、再到由他們負責安裝等各個層面。之所以會走到這一步，是因為我們曾向一家公司提出意向書（letter of intent），並在一段時間前得知 Sovereign AI，於是開始研究並找出其中的缺口。

So in doing that, we started finding out, okay, who are the folks that are installing this Sovereign AI? And then as we started looking at this very seriously, we said, well, these are companies that we can use to sub out.

因此在這個過程中，我們開始釐清：好，那麼到底是哪些人、哪些團隊在安裝這個 Sovereign AI？當我們開始非常認真地看待這件事時，我們就說，這些公司是我們可以用來外包（sub out）的對象。

So from a US basis, Eastern Europe and from Indian basis, companies are looking to develop this software that today does not exist. You can do it -- we did at CloudFirst for over 20 years, protecting someone's information and having a run book to get the companies up and going because regulated companies using the cloud with proprietary data, they're pretty much building it themselves. So we're really on all fronts at this point. And so we hope to start building a statement of work probably over the next 30 days.

所以以美國、東歐以及印度為基礎的公司，都在尋求開發目前尚不存在的這套軟體。你可以自己做——我們在 CloudFirst 做了超過 20 年，保護客戶資訊，並且有一套 run book 讓企業能夠上線運作；因為受監管的公司在雲端使用專有資料時，幾乎都是自己建置。所以我們目前是多線並進。我們希望在接下來大約 30 天內開始擬定工作說明書（statement of work）。

And that might involve probably three separate companies, each one having a different discipline. But right now, a number of companies, as I've gone around talking about this, and I'm kind of be somewhat quiet to a degree because you turned them into competitors.

而這可能會涉及大約三家不同公司，各自負責不同專業領域。但目前我在外面談這件事時，有不少公司——我某種程度上也會保持低調，因為你一講出來就把他們變成競爭對手。

But for the most part, we would say there's probably going to be three companies involved with putting this together in the 2 colocation centers is what our intention were to be. But overall, we have to start with someone that's going to be project management, and that's why we have the recruiter going on because there'll be a lot going on, but we've done it before with 10 data centers in three countries.

但整體而言，我們會說大概會有三家公司參與，在兩個機房共置中心（colocation centers）把這件事組合起來，這是我們原本的意圖。不過總體上，我們必須先從一位能做專案管理的人開始，這也是我們正在進行招募的原因；因為會有很多事情同時發生，但我們以前做過——在三個國家、十個資料中心。

We're very similar to that, but the software to flip it over when there's a disaster of some sort. And even though people can say, well, Tier 3 data centers, but everybody that's in Tier 3 data centers today still has to be geographically diverse if they're going to be compliant and a whole list of other things. And that's where we're heading.

這與那時非常相似，只是這次是要在發生某種災難時，能把系統切換過去的軟體。即使有人會說「Tier 3 資料中心」，但今天在 Tier 3 資料中心的所有業者，如果要合規，仍然必須具備地理分散性，以及一長串其他要求。而這正是我們要前進的方向。

Stage 1 will be to make it look like it was almost CloudFirst, but on the GPU side and everything that goes along with GPU and storage, and the second stage of it will be building the software all along to be able to have it flip over and act behaviorally the same way, behavioral point objective, behavioral time objective. So this is very, very much similar that we did with CloudFirst, but it's GPUs, and they are different. But that's so there'll be multiple companies involved. I'm sorry, a short question, a very long answer, but there'll be multiple companies that we're talking to today.

第一階段會讓它看起來幾乎像 CloudFirst，只是改成 GPU 端，以及所有與 GPU 和儲存相關的內容；第二階段則是一路把軟體建起來，使其能夠切換（flip over），並在行為上以相同方式運作——行為復原點目標（behavioral point objective）、行為復原時間目標（behavioral time objective）。所以這與我們在 CloudFirst 做的事情非常、非常相似，只是這次是 GPU，而 GPU 確實不同。但總之會有多家公司參與。抱歉，你問的是短問題，我回答得很長，但我們今天確實正在與多家公司洽談。

Sure. No, I appreciate all the color. It's helpful to kind of conceptualize what you're doing. Maybe just as a follow-up. Obviously, I think you have a better sense of timing than we do.

了解。不會，我很感謝你提供這麼多細節。這有助於我們把你們正在做的事情具體化。也許再追問一下。顯然我想你們對時程的掌握會比我們更清楚。

But will we start to see expenses ramp up maybe in the second quarter or more in the third quarter around the initiative? And so will we see that starting to hit the P&L? Or would investments be capitalized and we won't necessarily see it on the P&L. Just curious how the participation might look or as it's looking today and if that's the right time line to think about?

但我們是否會開始看到與該項倡議相關的費用在第二季可能上升，或更多是在第三季上升？那麼我們會看到這些開始反映在損益表（P&L）上嗎？或者投資會被資本化，因此我們不一定會在損益表上看到它。只是好奇參與情況可能會是什麼樣子，或就目前看起來如何，以及這是否是正確的時間軸來思考？

Sure. Well, rounding our money, we have, let's say, $10 million in the bank. We have some escrows going on still from the Renovus sale. We just settled one on the net working capital with them and have $700,000 that we have come in or coming in over the last week or so. So we do have some cash.

當然。嗯，先把數字取整來說，我們銀行裡大概有1,000萬美元。我們仍有一些與Renovus出售相關的託管款（escrow）在進行中。我們剛就淨營運資金（net working capital）與他們結算了一筆，過去一週左右我們已收到或即將收到70萬美元。所以我們確實有一些現金。

The Board approved at a recent Board meeting for us to go out and explore this and line it completely up with all the pieces that are needed. But I think that it will hit the cash, but it won't be I don't want to use the word significant. I can't imagine us spending more than $250,000 to $300,000 on being able to get it to the point of our statement of work part before we say go.

董事會在最近一次董事會會議上批准我們出去探索這件事，並把所有需要的環節完整地對齊。但我認為它會影響現金，但不會——我不想用「重大」這個詞。在我們說「開始做」之前，我無法想像我們會在把它推進到工作說明書（statement of work）階段之前花超過25萬到30萬美元。

When we say go, it's they're going to be capital expenses. Those capital expenses will be depreciated over 5 years for the most part. So the big hit on the cash, I think most of it would be capital. The software development and all of that, we'll see how we can make arrangements, but that will probably be the part that will be just unknown at this particular point, frankly, on the software side. But there'll be capital expenditures going on.

當我們說「開始做」時，那些將會是資本性支出（CapEx）。這些資本性支出大多會在5年內折舊。所以我認為對現金的主要衝擊，大部分會是資本項目。至於軟體開發等等，我們會看看能如何安排，但坦白說，在目前這個時間點，軟體端可能會是比較未知的部分。但會有資本支出在進行。

But I think we have enough money to implement this and still have a couple of year run if revenue wasn't generated. But we're hoping to take -- hopefully taking agreements in the first quarter of '27, maybe earlier, of which I'll call reservations versus subscription, but they'll all be recurring revenue.

但我認為我們有足夠的資金來落地這個計畫，即使沒有產生營收，仍可支撐運作幾年。不過我們希望在2027年第一季簽下——希望能在2027年第一季，甚至更早——我稱之為「預留」而非「訂閱」的協議，但它們都會是經常性收入（recurring revenue）。

Yes. That makes sense. And maybe last question, and then I'll jump back in the queue. But I guess referring to that not a subscription. I guess that kind of speaks towards figuring out what capacity you need relative to how many customers you have, but and what their demands are.

是的。這很合理。也許最後一個問題，然後我就回到隊列裡。不過我想，提到那個不是訂閱。我想這某種程度上是在說，你需要根據客戶數量以及他們的需求，來判斷你需要多少產能，但——以及他們的需求是什麼。

But can you talk a little bit how you're thinking about how far ahead of demand that you need to build out capacity and how access to GPUs and data center space might look as you progress over the next few quarters?

但你能否談談你們在思考要提前於需求多少來建置產能，以及在未來幾個季度推進過程中，取得GPU與資料中心空間的情況可能會是什麼樣子？

Well, I'm going to say the next one or two quarters, we'll just be setting everything all up, hopefully, having it all in place by the end of the year. What's interesting about it is that we're -- we wouldn't be into this -- let's keep buying more and more GPUs, spending $50 billion that you're seeing that's going on. That's not the play here. The play here is essentially to use just an example, take a midsized hospital. A midsized hospital, let's say they're going to spend $1 million and set up their environment.

嗯，我會說接下來一到兩個季度，我們主要會把所有事情都建置好，希望在年底前把一切都到位。有趣的是，我們——我們不會走那種——一直買更多更多GPU、花你看到的那種500億美元的路線。這不是這裡的打法。這裡的打法本質上是用一個例子來說：以一家中型醫院為例。一家中型醫院，假設他們要花100萬美元來建置他們的環境。

They're going to run logistics for an operating room where they're pharmaceutical and they're building this critical. They might have subscribed to software.

他們會在手術室跑物流流程，涉及藥品，而且他們正在建置這個關鍵系統。他們可能已經訂閱了某個軟體。

They didn't build it. They install it and it keeps learning and becoming more intelligent. Well, now what are they going to spend to get to the other side to have the compliance in Sarbanes-Oxley, all these things that no one's talking about yet. So now are you going to double that CapEx? Or do you want to go to a service bureau, and we don't believe NVIDIA is going to build a service bureau, by the way, CoreWeave and people like that, they could do it, they're not really focused on it.

他們不是自己開發的。他們安裝後，它會持續學習並變得更智慧。那麼，為了在《薩班斯—奧克斯利法案》（Sarbanes-Oxley）等合規要求下達到另一端的合規性——這些事情目前還沒有人在談——他們還要花多少錢？所以現在你要把資本支出（CapEx）加倍嗎？還是你想去找一家服務局（service bureau）？順帶一提，我們不認為NVIDIA會去建服務局；CoreWeave之類的公司可能做得到，但他們其實沒有真正聚焦在這上面。

But for the most part, they now need to have the ability to be able to recover. And so when we talk about this recovery piece, the return on investment seems significant for them. So I would say that when we're looking at this, a midsized hospital is going to need to be able to be compliant. Their confidential information is sitting on their storage remotely, and we have runbooks.

但大多數情況下，他們現在需要具備復原（recover）的能力。因此當我們談到這個復原的部分時，對他們而言投資報酬（ROI）看起來相當可觀。所以我會說，當我們看這件事時，一家中型醫院需要能夠合規。他們的機密資訊遠端放在儲存設備上，而我們有作業手冊（runbooks）。

But at some point, it needs to flip over and act the exact same way and recover. So it's -- I don't know if I'm answering that question completely. But that's kind of the model that you're looking at. That could be insurance companies as well, financial institutions. Does that answer your question, Matt? I'm not sure.

但在某個時間點，它需要切換過去、以完全相同的方式運作並完成復原。所以——我不確定我是否完全回答了那個問題。但這大概就是你在看的模型。保險公司也可能是這樣，金融機構也是。Matt，這有回答你的問題嗎？我不太確定。

It helps. I guess to clarify, I guess when you were hosting CloudFirst and disaster recovery there, you had an idea of how much capacity you needed. But taking the $1 million environment in a midsized hospital, what would be the -- I assume you'll have enough capacity. Are you spending 1 to 5, so your environment would support 5? And how do you balance the investment of customer needs to fail over in the GPU environment versus how much overcapacity you want to build?

有幫助。我想澄清一下：我想當你們以前在CloudFirst託管並提供災難復原時，你們對需要多少產能是有概念的。但以中型醫院那個100萬美元的環境來說，會是——我假設你們會有足夠的產能。你們是投入1來支援5嗎，也就是你們的環境可以支援5個？以及你們如何在客戶在GPU環境中需要故障切換（fail over）的需求投資，與你們想建置多少超額產能之間取得平衡？

Well, the first thing I think we know by now after all these years providing business continuity is that a hospital is going to run this application or multiple applications to improve efficiency and all of that, and they're going to depreciate this equipment over three to five years that hospital is not going to be in the race to add more and more GPUs and more and more GPUs.

嗯，第一件事是，我想我們在提供業務持續性（business continuity）這麼多年之後，現在已經知道：醫院會跑這個應用或多個應用來提升效率等等，而且他們會在三到五年內折舊這些設備；醫院不會加入那種不斷加更多GPU、更多GPU的競賽。

So we don't see the growth there. So when you don't -- we don't see them continue to build upon that at the rates that we're seeing folks spending $50 billion. So we can match their equipment on our side. So let's just say, for example, that they want to recover within 15 minutes.

所以我們不認為那裡會有那種成長。因此當你不——我們不認為他們會以你看到有人花500億美元那樣的速度持續擴建。所以我們可以在我們這一端匹配他們的設備。例如，假設他們希望在15分鐘內完成復原。

Well, that's going to be a higher-level service, and that's not going to run a ratio. That's going to be 1:1 for them. And that's going to be what we would call high availability in a regular sense. Then there's another layer underneath there, like you are mentioning that, where you're going to run a 5:1 ratio, and 8:1 ratio. The one thing we learned during 9/11 with CloudFirst and then other disasters and storms that all happen is that things can happen geographically within a particular region.

那將會是更高等級的服務，而且不會用比例來跑。對他們而言會是1:1。而那會是我們在一般意義上所稱的高可用性（high availability）。然後在其下還有另一層，就像你提到的那樣，你會用5:1的比例、8:1的比例來跑。我們在9/11期間使用CloudFirst，以及後來發生的其他災害與風暴中學到的一件事是：事情可能會在特定區域內以地理方式發生。

So if you run too high of a ratio, you can't support it. So it needs to be coming from different geographies on that. But I would assume that a 5:1 ratio would be successful as long as you could probably run a 10:1 ratio as long as the 10 are in all different parts of the United States.

所以如果你的比例拉得太高，你就無法支撐。因此在這方面需要來自不同地理區域的支援。但我會假設，只要這 10 個都分布在美國各地不同地區，5:1 的比例會是可行的，同樣地，你大概也能運行 10:1 的比例。

But I would say on standby type service, where you have run books and all of that, I would say that probably 5:1 would be a good ratio.

但我會說，就待命型服務而言，也就是你有運行手冊（run books）以及所有那些配套，我認為 5:1 可能會是一個不錯的比例。

Very helpful. Thank you.

非常有幫助。謝謝。

[Ellen Lidsack, Fourth Capital].

[Ellen Lidsack，Fourth Capital]。

Yes, hi, and thank you so much for taking my question. Can you elaborate on the market opportunity you see for the Sovereign AI solutions and why you think now is the right time to enter the space?

是的，您好，也非常感謝您回答我的問題。您能否進一步說明您所看到的 Sovereign AI 解決方案的市場機會，以及為什麼您認為現在是進入這個領域的合適時機？

Sure, thanks, Ellen. The right time, it could be early on it, but if it takes like 6 months, when all of a sudden, we believe that when everyone starts -- everyone looks at AI as a general population of the world now as they go into ChatGPT and they ask a question or Claude and say, design this and design that. The fifth layer of this AI is the business process, and that's the software being developed. And so these 150 executives that OpenAI is putting in place that was in a press release is going out to actually build this software.

當然，謝謝你，Ellen。所謂合適的時機，可能現在還算早，但如果再過大約 6 個月，突然之間，我們相信當所有人——也就是全球一般大眾——現在把 AI 當作他們進入 ChatGPT 問問題，或用 Claude 說「設計這個、設計那個」的工具時。AI 的第五層是業務流程，也就是正在被開發的軟體。因此，OpenAI 在新聞稿中提到要配置的那 150 位高階主管，實際上就是要出去打造這些軟體。

As this software gets deployed, they're going to need to be compliant the same way all the CPUs have to be compliant in industry that they're using best practices. Today, that's not in existence. It might be all happening in one data center.

當這些軟體開始部署後，它們將需要符合合規要求，就像產業中所有 CPU 都必須合規、遵循最佳實務一樣。而今天，這樣的機制並不存在。它可能全部都發生在同一個資料中心。

So I think it's a matter of time before compliance and regulations start surrounding as more and more organizations regulated organizations are deploying these types of software and services to make them more efficient, to learn better, reduce staff, whatever they're thinking. But that's why these 150 people are being hired because companies are interested. The talent is lacking on it. and we're there to be able to go to Sovereign AI to say, well, you put this in place, how compliant are you? No one, I don't believe anyone is asking that question, and we've been talking to a lot of people.

所以我認為，隨著越來越多受監管的組織部署這類軟體與服務，以提升效率、學得更好、減少人力，或不管他們在想什麼，合規與監管開始形成只是時間問題。但這也是為什麼要聘用這 150 人，因為企業有興趣。相關人才是缺乏的，而我們能在 Sovereign AI 這塊提供協助，去問：你把這套建起來之後，你的合規程度如何？我不相信有人在問這個問題，而我們已經和很多人談過了。

So everyone is focused on learning, training the models, installing equipment, testing it, but they're not there on compliance and all the regulations that went on over the previous years. And that's why I believe it's a very solid business model.

所以大家都專注在學習、訓練模型、安裝設備、測試，但在合規以及過去幾年累積的各項監管要求上，還沒有跟上。這也是我相信這會是一個非常穩健商業模式的原因。

That makes sense. That kind of leads into my next question. What do you think really differentiates the Sovereign AI solutions from traditional disaster recovery, cybersecurity or any enterprise infrastructure providers currently in the market?

這很合理。這也引出了我的下一個問題。您認為，Sovereign AI 解決方案與目前市場上的傳統災難復原、資安或任何企業基礎設施供應商相比，真正的差異化在哪裡？

I think it's the same thing. Essentially, you can say it's the same thing, but none of the folks that are today in disaster recovery that we know that our research came up with are doing anything like this. Whether they're planning that, I'm not exactly sure, but there's enough room in it. Some of the ratios I've seen is that this is going to be somewhere around 5% to 10% of anyone that's putting Sovereign AI in place. So some numbers I've seen, and it is very tough when you start looking at market numbers is that it's a Sovereign AI is right around a $50 billion total addressable marketplace and 10% is what some of the numbers that I've seen for this type of thing.

我認為是同一件事。本質上你可以說它是同一件事，但就我們所知、以及我們研究所找到的結果，目前在災難復原領域的那些業者，沒有任何一家在做類似的事情。他們是否在規劃，我不太確定，但這裡面空間很大。我看到的一些比例是：這大概會占任何部署 Sovereign AI 的客戶的 5% 到 10%。我看過一些數字；而且當你開始看市場規模數據時會非常困難，但 Sovereign AI 的整體可服務市場（TAM）大約在 500 億美元左右，而我看到的某些數字顯示，這類型的需求大約占其中 10%。

But they're rough calculations, and I wouldn't hold me to it. But I know this is -- I have a solid feeling that this is coming. And I do believe that the folks that are in this business that CloudFirst competed with will eventually move into this. I think we might have a head start on it, and I think that, that's important, but there's enough room with 5 or 6 competitors. But right now, if we get this up by the end of the year, and we start talking to people in the fourth quarter, I think we'll have a little bit of a lead because of our background, we know about escalation risk.

但這些都是粗略估算，你不要拿這些數字來要求我負責。不過我知道這件事——我有很強烈的感覺它會到來。而且我確實相信，過去 CloudFirst 競爭的那些同業，最終也會進入這個領域。我認為我們可能有先發優勢，我覺得這很重要，但即使有 5 或 6 個競爭者，市場空間也足夠。但就目前而言，如果我們能在年底前把它建起來，並在第四季開始與客戶洽談，我認為我們會有一些領先，因為我們的背景讓我們了解升級風險（escalation risk）。

We know how to do that. We were doing that. We know how to have run books and all the things that went on with that. So we do understand all of that. And I think it fits in really, really well with this. But we saw the whole -- we saw that come up because we see what's going on with Sovereign AI and AI factories. I heard some numbers from Dell of proposals outstanding. They were just some large numbers. So I'm pretty excited about it.

我們知道怎麼做。我們以前就在做這件事。我們知道如何建立運行手冊（run books）以及相關的所有流程。所以我們確實理解這一切。而我認為這與這個方向非常、非常契合。但我們之所以看到整個——看到這個機會浮現，是因為我們看到 Sovereign AI 和 AI 工廠正在發生的事情。我從 Dell 那邊聽到一些尚未結案的提案數量。那些數字非常大。所以我對此相當興奮。

Definitely very exciting. And I guess in terms of the development time line and then the potential commercialization path for Sovereign AI, what does that look like over the next 12 to 24 months?

確實非常令人振奮。那我想就 Sovereign AI 的開發時間表以及潛在商業化路徑而言，未來 12 到 24 個月會是什麼樣子？

Everything is about execution. We all know that. So initially, we were going to try to do everything and then launch. And then studying it some more. We felt maybe the thing to do is to do a 2-stage approach.

一切都取決於執行。我們都知道這點。所以一開始，我們打算把所有事情都做完再推出。但在更深入研究之後，我們覺得也許應該採取兩階段的方法。

Let's get this up and going without the behavioral side of it. So that these are regulated organizations, they can be protected, but it's going to be different. It might not move over the exact same way right away behaviorally. You have the run book and all of these things. But the first stage will be to stand it up start taking reservations, which I want to call it reservations instead of subscription and get it moving so they can start testing and coming over to us.

先把它建起來並運作起來，但先不包含行為面（behavioral side）。這樣這些受監管的組織可以先受到保護，但會有所不同。在行為層面上，它可能不會立刻以完全相同的方式切換過來。你會有運行手冊（run book）以及所有這些東西。但第一階段是把它架起來、開始接受預約——我想稱之為「預約」而不是「訂閱」——並讓它開始運轉，讓客戶可以開始測試並逐步轉移到我們這邊。

And then from the very beginning, let's just say, within 60 days, software starts to get developed.

然後從一開始來看，比方說在 60 天內，軟體就會開始開發。

So by the time everything gets deployed on the hardware side, staffing is in place. Hopefully, it's not going to take more than 9 months. There's some software out there that you can work with, but a lot has to be developed. So it just doesn't exist. We dealt with this with our IBM systems with -- precisely that did a roll-up of all the software companies we used for 15-plus years.

因此等到硬體端全部部署完成、人員配置到位時，希望不會超過 9 個月。市面上有一些軟體可以搭配使用，但仍有很多必須自行開發。所以它就是不存在。我們在 IBM 系統上也遇過這種情況——完全一樣——當時我們把 15 年以上使用過的所有軟體公司做了整合彙總。

And so -- and we think there'll be very, very good value in owning the software as well. But that's kind of the timeline like.

所以——我們也認為，連同軟體一起擁有會帶來非常、非常好的價值。但大致的時間表就是這樣。

Okay. Well, that's great. And are you currently evaluating any like strategic partnerships, acquisitions or maybe even like investments that could potentially accelerate this AI infrastructure strategy?

好的。嗯，那太好了。那麼你們目前是否正在評估任何例如策略性合作夥伴關係、併購，甚至可能是投資，以加速這項 AI 基礎設施策略？

I originally wanted to do and I still -- we still made a joint venture, folks that are already set up that are installing Sovereign AI today. and to do a joint venture because they have the staff already in place, and they have the knowledge of it. And it's great for them and that becomes an automatic partner because they're installing AI factories and Sovereign AI. But we are talking to folks to be partners. One of the problems, Ellen, is that when you're small, a lot of times, you're not going to be able to get larger organizations to go with you because that credibility is not there.

我原本想做，而且我現在仍然——我們也確實成立了一個合資公司，和那些已經就位、今天正在安裝主權 AI（Sovereign AI）的人合作。之所以做合資，是因為他們已經有現成的團隊，也具備相關知識。這對他們很棒，而他們也會自動成為合作夥伴，因為他們正在安裝 AI 工廠與主權 AI。但我們也正在和一些人洽談成為合作夥伴。Ellen，其中一個問題是，當你規模很小時，很多時候你無法讓大型組織願意跟你合作，因為缺乏那種信譽。

They want to see a $1 billion company. Even though the $1 billion company can be insolvent, just for the most part, they want to see a very large scope.

他們想看到一家市值 10 億美元的公司。即使那家 10 億美元的公司可能資不抵債，但多數情況下，他們就是想看到非常大的規模。

So typically working through partners and that's how we did it at CloudFirst as well. When you get that very large deal, you bring in a partner on it. But we are looking at joint ventures, we're looking at partnerships. We're not really looking at investments at this time. We don't feel that that's necessary, frankly.

所以通常是透過合作夥伴來運作，我們在 CloudFirst 也是這麼做的。當你拿到那種非常大的案子時，你會把合作夥伴拉進來一起做。但我們正在看合資、也在看合作夥伴關係。目前我們並沒有真的在看投資。坦白說，我們覺得沒有那個必要。

I think we can do this with money in our bank and still leave a two-year run rate because the public company is expensive. It runs probably around, I'd say, $1.8 million to $2 million a year. But we have enough. I think we have enough to pull this off, but I'll know more over the next 90 days. But we're trying to move pretty fast with it.

我認為我們可以用銀行帳上的資金來做這件事，同時仍保留兩年的資金續航，因為上市公司很昂貴。我估計一年大概需要 180 萬到 200 萬美元。但我們資金足夠。我覺得我們有足夠的資源把這件事做成，不過接下來 90 天我會更清楚。但我們正試著把速度拉得很快。

No, this is super helpful. Thank you so much for taking my questions. I really appreciate it, Chuck. And if I have any other questions, I'll jump back in the queue.

不，這非常有幫助。非常感謝你回答我的問題。我真的很感激，Chuck。如果我還有其他問題，我會再回到提問隊列。

That's great. Thank you, Ellen.

很好。謝謝你，Ellen。

Matthew Galinko, Maxim Group.

Maxim Group 的 Matthew Galinko。

Hey, appreciate you taking another one from me. Just wanted to check in on Nexxis and kind of the current revenue generator for the business. I think you had decent annual growth in the first quarter here. Any opportunities to -- or how do you see that business trending over the rest of this year? Do you have an opportunity to accelerate that in any capacity?

嗨，感謝你再讓我問一題。我想了解一下 Nexxis，以及它目前作為公司營收引擎的狀況。我記得你們第一季的年增長還不錯。今年剩下的時間你怎麼看這塊業務的走勢？你們是否有機會在任何方面加速它？

And do you see it continuing to add to kind of cut into the burn rate, I guess, as it grows?

另外，隨著它成長，你是否認為它會持續貢獻、某種程度上幫助降低（我想是）現金消耗率？

Matt, the gross margins are great. We have put some money into Nexxis. They're not a large staff, John Camello, who is the President of that, he owns 20% of that company. John and his staff do an excellent job. John continues to look for business development types to accelerate it.

Matt，毛利率很棒。我們已經在 Nexxis 投入了一些資金。他們的團隊不大；該業務的總裁 John Camello 持有那家公司 20% 的股權。John 和他的團隊做得非常出色。John 也持續在尋找商務開發類型的人才來加速成長。

And I know that he's trying to recruit as we speak right now, he's trying to recruit business development folks to go. It's very, very difficult, the organic growth, but they're doing a great job with it. We looked at 2 -- 1 or 2 acquisitions to roll it into that company. and we're still looking at that. But I think if John gets successful, we're getting the right -- he is successful, we're getting the right people on to grow that.

我知道他現在正在招募——就在我們通話的同時，他正在招募商務開發人員加入。有機成長非常、非常困難，但他們做得很棒。我們曾看過 1 到 2 個併購案，想把它們併入那家公司，我們也仍在評估。但我認為如果 John 招募成功——他本身很成功——我們把對的人找進來，就能把那塊業務做大。

I also believe, Matt, that because they're very limited with manpower that getting a digital agency to start getting inbound leads going is one of the things that we've been talking about CloudFirst had a great flow of leads. Hal Schwartz did a great job with the digital agency and everything that he did on that to get significant leads coming in. And so we need that to happen and then these business development folks to work on that because no one's answering the phone, no one is letting you in the building.

我也相信，Matt，因為他們人力非常有限，找一家數位行銷代理商來開始帶動入站線索（inbound leads）是我們一直在討論的方向之一；CloudFirst 當時有很好的線索流量。Hal Schwartz 在數位代理商以及相關工作上做得很出色，帶來大量重要線索。所以我們也需要那樣的成果，然後再由這些商務開發人員去跟進，因為沒有人接電話，也沒有人讓你進大樓。

So John does a great job and his staff with association meetings and organizations and sponsorships, things like that. But that next step, I think, is Chris to free up some money for him to get the website going where he can get an inflow of the way that CloudFirst done. And I think that's the next stage, but he is trying to recruit the folks in the business development area. He needs the help there because he's got great growth margins and does a good job, has a great product -- the product is great.

所以 John 和他的團隊在協會會議、各類組織與贊助等方面做得很好。但我認為下一步是由 Chris 釋出一些資金，讓他把網站等建置起來，讓他能像 CloudFirst 那樣帶來流入。我覺得那是下一階段；但他目前正在招募商務開發領域的人才。他需要那方面的支援，因為他有很好的成長毛利，也做得很好，產品也很棒——產品真的很棒。

Great. Thank you.

很好。謝謝。

Thank you. There are no further questions at this time, so I'll hand it back to Chuck Piluso for closing remarks.

謝謝。目前沒有其他問題了，我把時間交回給 Chuck Piluso 做結語。

Thank you. Thank you for the questions. They were very deep questions, some of them. And Ellen, they were great. Hopefully, shortly, we'll be back to everyone, but thank you for the questions.

謝謝。謝謝大家的提問。有些問題非常深入。Ellen，你的問題也很棒。希望不久之後我們能再回到大家面前，但謝謝各位的提問。

In closing, we believe the foundation we've established over the decades of execution and value creation has positioned DTST to pursue a unique opportunity at the intersection of enterprise AI, resiliency and regulated infrastructure. Our strategy is supported by financial strength, operational stability and what we believe is a differentiation of long-term vision for AI continuity infrastructure.

最後，我們相信，數十年來在執行與價值創造上所建立的基礎，已使 DTST 具備條件去把握一個獨特機會——位於企業 AI、韌性（resiliency）與受監管基礎設施的交會點。我們的策略有財務實力、營運穩定性作為支撐，並且我們相信，在 AI 持續運作（continuity）基礎設施方面，我們具備長期願景的差異化。

As the market continues to evolve, our focus remains on a disciplined execution, strategic flexibility and creating substantial long-term value for our shareholders. We really do appreciate everyone's continued support in our shareholders and look forward to sharing additional updates as we progress. Thank you.

隨著市場持續演進，我們的重點仍是紀律性的執行、策略彈性，以及為股東創造可觀的長期價值。我們非常感謝各位股東持續的支持，並期待在我們推進的過程中分享更多更新。謝謝。

Thank you. And with that, we conclude today's call. All parties may disconnect. Have a good day.

謝謝。那麼，我們今天的電話會議到此結束。各方可掛線。祝各位有美好的一天。