Upstart Holdings Inc (UPST) 2023 Q1 法說會逐字稿

Good day, and welcome to the Upstart First Quarter 2023 Earnings. Today's conference is being recorded.

美好的一天，歡迎來到 Upstart 2023 年第一季度收益。今天的會議正在錄製中。

At this time, I would like to turn the conference over to Jason Schmidt, Head of Investor Relations. Please go ahead.

此時，我想將會議轉交給投資者關係主管 Jason Schmidt。請繼續。

Good afternoon, and thank you for joining us on today's conference call to discuss Upstart's first quarter 2023 financial results. With us on today's call are Dave Girouard, Upstart's Chief Executive Officer; and Sanjay Datta, our Chief Financial Officer.

下午好，感謝您參加今天的電話會議，討論 Upstart 2023 年第一季度的財務業績。 Upstart 的首席執行官 Dave Girouard 和我們一起參加了今天的電話會議；和我們的首席財務官 Sanjay Datta。

Before we begin, I want to remind you that shortly after market closed today, Upstart issued a press release announcing its first quarter 2023 financial results and published an Investor Relations presentation. Both are available on our Investor Relations website, ir.Upstart.com.

在我們開始之前，我想提醒您，今天收市後不久，Upstart 發布了一份新聞稿，宣布了其 2023 年第一季度的財務業績，並發布了投資者關係介紹。兩者均可在我們的投資者關係網站 ir.Upstart.com 上找到。

During the call, we will make forward-looking statements, such as guidance for the second quarter of 2023 related to our business and plans to expand our platform in the future. These statements are based on our current expectations and information available as of today and are subject to a variety of risks, uncertainties and assumptions. Actual results may differ materially as a result of various risk factors that have been described in our filings with the SEC. As a result, we caution you against placing undue reliance on these forward-looking statements. We assume no obligation to update any forward-looking statements as a result of new information or future events, except as required by law.

在電話會議期間，我們將做出前瞻性陳述，例如與我們業務相關的 2023 年第二季度的指導以及未來擴展我們平台的計劃。這些陳述基於我們目前的預期和截至今日可用的信息，並受到各種風險、不確定性和假設的影響。由於我們在提交給美國證券交易委員會的文件中描述的各種風險因素，實際結果可能存在重大差異。因此，我們提醒您不要過分依賴這些前瞻性陳述。除非法律要求，否則我們不承擔因新信息或未來事件而更新任何前瞻性陳述的義務。

In addition, during today's call, unless otherwise stated, references to our results are provided as non-GAAP financial measures, and are reconciled to our GAAP results, which can be found in the earnings release and supplemental tables. (Operator Instructions)

此外，在今天的電話會議中，除非另有說明，否則我們的業績參考將作為非 GAAP 財務指標提供，並與我們的 GAAP 結果進行核對，這可以在收益發布和補充表中找到。 （操作員說明）

Later this quarter, Upstart will be participating in the Barclays Emerging Payments and Fintech Forum on May 17, and the Bank of America Global Technology Conference, June 7.

本季度晚些時候，Upstart 將參加 5 月 17 日舉行的巴克萊新興支付和金融科技論壇，以及 6 月 7 日舉行的美國銀行全球技術大會。

Now I'd like to turn it over to Dave Girouard, CEO of Upstart.

現在我想把它交給 Upstart 的首席執行官 Dave Girouard。

Good afternoon, everyone. Thank you for joining us on our earnings call covering our first quarter 2023 results. I'm Dave Girouard, Co-Founder and CEO of Upstart. Despite the headwinds facing our industry in early 2023, I'm pleased with the progress we made against the objectives that I set out for you previously. I'm hopeful that as we move through the year, you will come to see Q1 as a transitional quarter for Upstart.

大家下午好。感謝您加入我們的財報電話會議，了解我們 2023 年第一季度的業績。我是 Upstart 的聯合創始人兼首席執行官 Dave Girouard。儘管我們的行業在 2023 年初面臨逆風，但我很高興我們在實現我之前為您設定的目標方面取得了進展。我希望隨著我們度過這一年，您會把第一季度視為 Upstart 的過渡季度。

While the economic environment remains turbulent, there are many reasons to be optimistic about our future. I assure you we aren't waiting around for the economy to improve. First, our development teams made giant leaps forward in each of our main product areas. Innovation in AI is the primary source of Upstart's competitive advantage, and we continue to break new ground in this area. I'll share more about these wins shortly.

儘管經濟環境依然動盪，但我們有很多理由對我們的未來感到樂觀。我向你保證，我們不會坐等經濟好轉。首先，我們的開發團隊在我們的每個主要產品領域都取得了巨大的飛躍。 AI 創新是 Upstart 競爭優勢的主要來源，我們將繼續在這一領域開闢新天地。我將很快分享更多關於這些勝利的信息。

Second, we accomplished this while taking significant fixed costs out of our business. Last quarter, I told you that I'm committed to running an operationally and fiscally tight ship. Given our concerted efforts in Q1 to reduce both payroll and operational expenses, I'm confident that Upstart is now a more streamlined and efficient company, setting us up to return to profitable growth soon. I'll share more about our cost reduction efforts later.

其次，我們在實現這一目標的同時從我們的業務中削減了大量的固定成本。上個季度，我告訴過你我致力於運營一艘運營和財政緊張的船。鑑於我們在第一季度共同努力減少工資和運營費用，我相信 Upstart 現在是一家更加精簡和高效的公司，使我們能夠很快恢復盈利增長。稍後我將分享更多關於我們降低成本的工作。

And finally, I'm pleased to tell you that we secured multiple long-term funding agreements. Together, expected to deliver more than $2 billion to the Upstart platform over the next 12 months. This is a critical first step toward building resiliency and predictability into our business. Together, I believe these efforts put us in a stronger position regardless of which direction the economy turns.

最後，我很高興地告訴你，我們獲得了多項長期資助協議。預計在未來 12 個月內，將為 Upstart 平台提供超過 20 億美元的資金。這是在我們的業務中建立彈性和可預測性的關鍵的第一步。總之，我相信無論經濟轉向哪個方向，這些努力都會使我們處於更有利的地位。

Our own analysis, which we launched publicly in March in the form of the Upstart Macro Index, or UMI, suggests that the financial health of the mainstream American consumer deteriorated rapidly through the first 9-or-so months of 2022, but has since stabilized, if not improved for the last several months.

我們自己的分析在 3 月份以新貴宏觀指數 (Upstart Macro Index, UMI) 的形式公開發布，表明美國主流消費者的財務狀況在 2022 年前 9 個月左右迅速惡化，但此後趨於穩定，如果最近幾個月沒有改善的話。

The personal savings rate, which may be the most relevant predictor of UMI, bottomed out mid last year at 2.7%, and has increased to 5.1% since then. Since late 2022, loans on our platform have been priced conservatively relative to UMI. So our bank and credit union partners can feel confident that recent vintages are today performing at or above expectations.

個人儲蓄率可能是 UMI 最相關的預測指標，去年年中降至 2.7% 的谷底，此後已升至 5.1%。自 2022 年底以來，我們平台上的貸款相對於 UMI 的定價較為保守。因此，我們的銀行和信用合作社合作夥伴可以確信最近年份的葡萄酒今天的表現達到或超過預期。

And while banks are certainly treading carefully in the current environment, many lenders and institutional investors appreciate the combination of high yield and short duration that Upstart-powered loans offer. Irrespective of the environment, I push our team very hard to make sure Upstart improves constantly in 4 critical dimensions: First, best rates for all. We founded Upstart to improve access to credit. So delivering the best rates possible to all consumers will always be our true north.

儘管銀行在當前環境下肯定會謹慎行事，但許多貸方和機構投資者都看好 Upstart 貸款提供的高收益和短期限的結合。無論環境如何，我都非常努力地推動我們的團隊確保 Upstart 在 4 個關鍵方面不斷改進：首先，為所有人提供最好的價格。我們成立 Upstart 是為了改善獲得信貸的機會。因此，為所有消費者提供最優惠的價格將永遠是我們真正的方向。

Given the breadth and diversity of competition, we'll never 100% achieve this goal. But in a fierce effort to do so, we can become the market leader in a vitally important segment of our economy. Better rates are unlocked first and foremost, by a more accurate and predictive credit model, one that excels at separating good risk from bad, and AI is the key to this.

鑑於競爭的廣度和多樣性，我們永遠不會 100% 實現這一目標。但通過努力做到這一點，我們可以成為我們經濟中一個至關重要的領域的市場領導者。更好的利率首先是通過更準確和更具預測性的信用模型解鎖的，該模型擅長將好風險與壞風險區分開來，而人工智能是實現這一點的關鍵。

Our models are today trained on more than 100 billion cells of performance data. And now with an average of 90,000 new loan repayments due each day across all our bank partners, the system is learning and adjusting in near real time to actual loan performance. We pushed 23 new and improved versions of our AI models into production during the first quarter alone, about 1 every 3 days.

今天，我們的模型接受了超過 1000 億個性能數據單元的訓練。現在，我們所有銀行合作夥伴平均每天要償還 90,000 筆新貸款，該系統正在近乎實時地學習和調整以適應實際貸款績效。僅在第一季度，我們就將 23 個新版本和改進版本的 AI 模型投入生產，大約每 3 天一個。

We're confident that our AI has never been as sophisticated or as accurate as it is today. But in order to deliver the best rates for all, we also need the diversity of bank and credit union partners, each with different priorities, business objectives and balance sheet issues to solve. Today, we have almost 100 such partners. In order of magnitude more than the 10 we had when we went public in December 2020.

我們相信我們的 AI 從未像今天這樣複雜或準確。但為了為所有人提供最優惠的利率，我們還需要銀行和信用合作社合作夥伴的多樣性，每個合作夥伴都有不同的優先事項、業務目標和資產負債表問題需要解決。今天，我們有近 100 個這樣的合作夥伴。比我們在 2020 年 12 月上市時的 10 個數量級還要多。

We additionally need a strong presence and reputation in institutional and capital markets because the limited risk appetite of bank balance sheets will never serve the needs of the entire U.S. credit market. This quarter, we expanded our roster of institutional partners in ways that should help us deliver quality offers to consumers through all parts of the cycle.

我們還需要在機構和資本市場上擁有強大的影響力和聲譽，因為銀行資產負債表的有限風險偏好永遠無法滿足整個美國信貸市場的需求。本季度，我們擴大了機構合作夥伴名單，以幫助我們在整個週期的各個階段為消費者提供優質服務。

Second, more efficient borrowing and lending. Every quarter, we aim to make our platform more efficient for consumers and bank partners. This improvement comes from better AI, which enables more sophisticated risk models, which in turn, enable a faster and more efficient experience for consumers and lenders.

第二，提高借貸效率。每個季度，我們的目標是提高我們平台對消費者和銀行合作夥伴的效率。這種改進來自更好的 AI，它支持更複雜的風險模型，從而為消費者和貸方提供更快、更高效的體驗。

In the first quarter, we achieved a record level of automation with 84% of Upstart-powered loans fully automated across all our bank partners. By this, I mean the loans were approved and verified instantly with 0 human intervention from rate request through loan funding. We know of no other lending marketplace with this level of automation.

在第一季度，我們實現了創紀錄的自動化水平，我們所有銀行合作夥伴中 84% 的 Upstart 貸款完全自動化。通過這個，我的意思是貸款被立即批准和驗證，從利率請求到貸款資金的人為乾預為 0。據我們所知，沒有其他借貸市場具有這種自動化程度。

This instant and automated process, which, by the way, 70% of consumers access via a mobile phone, creates an unparalleled wow moment for the borrower, which is -- who is often surprised, if not shocked, to realize that there are no more steps in the approval process. Because consumers rightfully value their time, this delightful moment is often more impactful than the specific rate our bank partners offer.

這個即時和自動化的過程，順便說一下，70% 的消費者通過手機訪問，為藉款人創造了一個無與倫比的驚喜時刻，他們經常驚訝，如果不是震驚，意識到沒有批准過程中的更多步驟。因為消費者理所當然地珍惜他們的時間，所以這個愉快的時刻通常比我們的銀行合作夥伴提供的特定利率更具影響力。

Efficiency is obviously important for the lender as well. Our bank partners' Upstart-powered lending programs are open for business 24 hours a day, 7 days a week. In addition to providing the modern all-digital experience their customers expect, banks can tailor their Upstart lending programs to precisely target their business objectives, risk appetite and balance sheet needs. When it comes to routine transactions like offering a loan, branch hours are inconvenient, human intervention is costly and consumers just want what they want, when they want it.

效率顯然對貸方也很重要。我們的銀行合作夥伴的 Upstart 貸款計劃每週 7 天、每天 24 小時開放。除了提供客戶期望的現代全數字體驗外，銀行還可以量身定制 Upstart 貸款計劃，以精確瞄准其業務目標、風險偏好和資產負債表需求。當涉及到提供貸款等常規交易時，分支機構營業時間不方便，人工干預成本高昂，消費者只需要他們想要的東西，當他們想要的時候。

Third, more resilient. After all we have experienced in the last year, I'm keenly aware that we need to build more resilience into Upstart's business. Lending is inherently cyclical, but we aim to build a platform that largely mitigates that cyclicality, ensuring that credit continues to be available and slowing when it's needed, albeit accurately priced to prevailing conditions.

三是韌性更強。在經歷了去年的所有經歷之後，我敏銳地意識到我們需要為 Upstart 的業務增強彈性。貸款本質上是周期性的，但我們的目標是建立一個平台，在很大程度上減輕這種週期性，確保信貸繼續可用，並在需要時放慢速度，儘管根據當前條件準確定價。

Central to this resilience is securing a baseline supply of long-term capital that we can depend on through the market's ups and downs. As I mentioned earlier, we've completed agreements with multiple strategic partners as of today, and we'll continue to explore additional partnerships. Our primary goal is to have loan funding capital committed at a level that allows us to remain cash flow positive through typical market cycles.

這種彈性的核心是確保長期資本的基準供應，我們可以依靠它來度過市場的起伏。正如我之前提到的，截至今天，我們已經與多個戰略合作夥伴完成了協議，我們將繼續探索更多的合作夥伴關係。我們的主要目標是將貸款資金投入到一定水平，使我們能夠在典型的市場週期中保持現金流為正。

Resilience also comes from a flexible business model, with lower fixed costs, proven pricing power and durable unit economics. We've always been a capital-efficient business, raising and spending a fraction of what peer companies spend. But there are always more ways to drive efficiency. Between Q4 and Q1, we took the necessary steps to reduce Upstart's headcount by almost 30%.

彈性還來自靈活的商業模式，具有較低的固定成本、成熟的定價能力和持久的單位經濟效益。我們一直是一家資本效率高的企業，籌集和支出的資金只是同行公司支出的一小部分。但總有更多方法可以提高效率。在第四季度和第一季度之間，我們採取了必要措施將 Upstart 的員工人數減少了近 30%。

While clearly a gut-wrenching decision, it will allow us to return to profitability at a significantly lower loan volume and has led us to be more focused and nimble in delivering our product road map. We also identified opportunities to reduce our technical infrastructure costs by as much as $10 million annually and sublet some unnecessary office space, both of which will go directly to the bottom line.

雖然這顯然是一個令人痛心的決定，但它將使我們能夠以顯著降低的貸款量恢復盈利，並使我們更加專注和靈活地交付我們的產品路線圖。我們還確定了每年將技術基礎設施成本降低多達 1000 萬美元並轉租一些不必要的辦公空間的機會，這兩者都將直接影響利潤。

On the revenue side, we flexed up our take rates, delivering a record contribution margin of 58% in Q1. Our prior record contribution margin was 54% in Q3 of 2020. Our strong and flexible unit economics are a byproduct of the competitive advantage provided by AI. All these together position us well to navigate whatever twists and turns lie ahead while strengthening our position for the inevitable sunnier markets to come.

在收入方面，我們提高了收入率，在第一季度實現了創紀錄的 58% 的邊際貢獻率。 2020 年第三季度，我們創紀錄的邊際貢獻率為 54%。我們強大而靈活的單位經濟效益是人工智能提供的競爭優勢的副產品。所有這些共同使我們能夠很好地應對未來的任何曲折，同時加強我們在不可避免的陽光市場到來時的地位。

The last important lever on resiliency is our product offering itself, which I'll speak to now. Fourth, a broader range of products. Moving beyond our core personal loan product is critically important to reaching our potential as a business. Offering a wider range of solutions makes us more relevant to more consumers and also more valuable to our bank and credit union partners. Product and borrower diversification can also provide greater resilience through future market cycles.

彈性的最後一個重要槓桿是我們的產品本身，我現在將談到這一點。四是產品種類更加豐富。超越我們的核心個人貸款產品對於發揮我們作為企業的潛力至關重要。提供更廣泛的解決方案使我們與更多消費者更相關，對我們的銀行和信用合作社合作夥伴也更有價值。產品和借款人的多元化也可以在未來的市場週期中提供更大的彈性。

We continue to make progress in our second big bet, the auto lending market. This market has had what may be the most tumultuous 3 years in its 100-plus year history. Despite this, we've made rapid progress with our products and couldn't be more excited about our potential. Since our last earnings call, we announced that both Acura and Mercedes-Benz approved Upstart as a digital retail provider, becoming our eighth and ninth OEM partners. We recently launched a new and improved AI model for our auto retail lending product that builds off our existing auto refinance model. We now consider both of these models calibrated and performing on target.

我們在第二大賭注——汽車貸款市場——繼續取得進展。這個市場經歷了可能是其 100 多年曆史中最動蕩的 3 年。儘管如此，我們在產品方面取得了快速進步，並且對我們的潛力感到無比興奮。自上次財報電話會議以來，我們宣布謳歌和梅賽德斯-奔馳都批准 Upstart 作為數字零售提供商，成為我們的第八和第九個 OEM 合作夥伴。我們最近為我們的汽車零售貸款產品推出了一個新的和改進的 AI 模型，該模型建立在我們現有的汽車再融資模型的基礎上。我們現在考慮這兩個模型都經過校準並按目標執行。

Our footprint of dealers piloting our lending product expanded to 39 since last I updated you, and we expect this rollout to continue throughout the year. We also signed our first external funding agreement for auto retail, which is an important milestone for us. And lastly, we're making rapid improvements to our servicing and collections of auto loans, which accrues directly to model performance.

自上次更新以來，我們試行貸款產品的經銷商已增至 39 家，我們預計這一推廣將在全年繼續進行。我們還簽署了第一份汽車零售外部融資協議，這對我們來說是一個重要的里程碑。最後，我們正在快速改進我們的汽車貸款服務和催收，這直接增加了模型性能。

I'm also pleased to let you know that we expect to launch a home equity product later this year. This is a great fit for Upstart for a few reasons. First, 95% of HELOCs are financed by banks and credit unions, so it's an asset our lending partners know and value. Second, HELOCs naturally serve a very prime consumer, namely homeowners. We expect Upstart HELOCs to have annual loss rates less than 1%. And third, home equity products tend to be countercyclical to refinance products. We know this because in Q4 2022, HELOC volumes grew 32% year-on-year even while mortgage refinances plummeted. Importantly, there's a lot of opportunity to improve the process of originating HELOCs. The average HELOC today takes 36 days to fund, while we are aiming for online approval in 10 minutes, and funding within 5 days.

我也很高興地告訴您，我們預計將在今年晚些時候推出房屋淨值產品。出於幾個原因，這非常適合 Upstart。首先，95% 的 HELOC 由銀行和信用合作社提供資金，因此這是我們的貸款合作夥伴了解和重視的資產。其次，HELOCs 自然服務於非常主要的消費者，即房主。我們預計 Upstart HELOC 的年損失率將低於 1%。第三，房屋淨值產品往往與再融資產品具有逆週期性。我們之所以知道這一點，是因為在 2022 年第四季度，HELOC 的交易量同比增長了 32%，儘管抵押貸款再融資大幅下降。重要的是，有很多機會可以改進 HELOC 的原始流程。今天的 HELOC 平均需要 36 天才能獲得資金，而我們的目標是在 10 分鐘內獲得在線批准，並在 5 天內獲得資金。

Lastly, I have to mention the incredible progress made by our small dollar loan team. By way of context, banks feel pressure from regulators to eradicate overdraft fees and instead to provide affordable relief loans to consumers who have short-term cash needs. But they have struggled for years to do this both meaningfully and profitably. Short-term loans of a few hundred dollars to existing customers or even to random lock-up consumers at rates within the APR limits for nationally chartered banks has seem beyond the reach of the banking industry.

最後，我不得不提一下我們小額貸款團隊取得的令人難以置信的進步。就背景而言，銀行感受到來自監管機構的壓力，要求他們取消透支費，轉而向有短期現金需求的消費者提供負擔得起的救濟貸款。但多年來，他們一直在努力做到既有意義又有利可圖。以國家特許銀行的 APR 限額內的利率向現有客戶甚至隨機鎖定消費者提供幾百美元的短期貸款似乎超出了銀行業的能力範圍。

Well, we're building it for them, and we believe it has the potential to eradicate more than 70% of payday loans in the next 5 years. Our small dollar product already has 90% automation rates, far beyond even our core personal loan product. And in Q1, we launched a new AI model for this product that delivered the largest single accuracy improvement measured in our history. We've also expanded the offering to include loan terms as short as 3 months, which drove a 37% increase in approval rates. I'm not sure we've ever delivered a product with as much impact and as much alignment with our mission as we're seeing from the small dollar team.

嗯，我們正在為他們建造它，我們相信它有可能在未來 5 年內消除超過 70% 的發薪日貸款。我們的小額美元產品已經擁有 90% 的自動化率，甚至遠遠超過我們的核心個人貸款產品。在第一季度，我們為該產品推出了一個新的 AI 模型，實現了我們歷史上最大的單一準確度改進。我們還擴大了產品範圍，包括最短 3 個月的貸款期限，這使批准率提高了 37%。我不確定我們是否曾經交付過像我們從小美元團隊看到的那樣具有如此大的影響力和與我們的使命一致的產品。

To wrap things up, I want to acknowledge that the financial industry is not out of the woods yet. But even in this challenging economy, I believe Upstart is in position to grow reasonably through our typical run rate of model and technology improvements. And when the banking and credit markets eventually normalize as surely they will, the true strength of our platform will become clear to all.

總結一下，我想承認金融業尚未走出困境。但即使在這個充滿挑戰的經濟中，我相信 Upstart 也有能力通過我們典型的模型運行率和技術改進實現合理增長。當銀行和信貸市場最終正常化時，我們平台的真正實力將公之於眾。

Upstart's success will continue to be built on excellence and leadership in AI. By this, I mean the speed with which we can develop, deploy and calibrate new and more accurate AI models. Together with the amazing quality of our team, it is this that makes me optimistic about Upstart's future.

Upstart 的成功將繼續建立在人工智能領域的卓越和領導地位之上。我指的是我們開發、部署和校準新的、更準確的人工智能模型的速度。加上我們團隊的驚人素質，正是這一點讓我對 Upstart 的未來感到樂觀。

Thank you, and I'd like now to turn it over to Sanjay, our Chief Financial Officer, to walk through our Q1 2023 financial results and guidance. Sanjay?

謝謝，我現在想把它交給我們的首席財務官 Sanjay，讓我們了解一下我們 2023 年第一季度的財務業績和指導。桑傑？

Thanks, Dave, and thanks to all of you for joining us today. At a headline level, over the past quarter, we have observed continuing stability in consumer repayment trends on the borrower side of our platform, counterbalanced by heightened volatility in the banking and institutional funding markets. In the U.S. Consumer Personal Finance, a couple of virtuous trends continue to unfold in the aggregate statistics.

謝謝戴夫，感謝大家今天加入我們。在總體層面上，在過去一個季度，我們觀察到我們平台借款人方面的消費者還款趨勢持續穩定，但銀行和機構融資市場波動加劇抵消了這一趨勢。在美國消費者個人理財方面，總體統計數據繼續呈現出一些良性趨勢。

The percent of adults participating in the workforce continues to climb steadily as our population returns to work. And the average personal tax burden for each individual has fallen considerably since last year, together, leading to a rising level of real disposable income per adult.

隨著我們的人口重返工作崗位，參與勞動力的成年人的百分比繼續穩步攀升。自去年以來，每個人的平均個人稅負大幅下降，共同導致每個成年人的實際可支配收入水平上升。

On the expenditure side, real consumption per capita continues to moderate and creep back into line with disposable income. Rising disposable income and moderating consumption expenditures are resulting in a personal savings rate that has now risen in each of the past 6 months since bottoming out last September. This particular indicator has demonstrated a strong correlation to borrower repayment health since the pandemic, and we are seeing this reflected in our Upstart Macro Index, which peaked last October and is showing signs of early recovery.

在支出方面，實際人均消費繼續放緩並逐漸回到與可支配收入的水平。可支配收入的增加和消費支出的放緩導致個人儲蓄率自去年 9 月觸底以來，在過去 6 個月中的每個月都在上升。自大流行以來，這一特殊指標已證明與借款人的還款健康狀況密切相關，我們在我們的新貴宏觀指數中看到了這一點，該指數於去年 10 月達到頂峰，並顯示出早期復甦的跡象。

The ongoing improvement in the consumer fiscal condition, together with the recalibration of our own underwriting models, is resulting in loan performance that as of our Q4 vintages, we believe is on track to deliver unlevered gross returns of approximately 11% as a blended average across the banks, credit unions and institutional buyers on our platform.

消費者財政狀況的持續改善，加上我們自己的承銷模式的重新校準，導致我們認為截至第四季度的貸款業績有望實現約 11% 的無槓桿總回報率，作為整個行業的混合平均值我們平台上的銀行、信用合作社和機構買家。

Conversely, the funding side of the ecosystem remains challenging in the current climate. Increased conservatism among existing lending partners in the wake of the recent bank failures has enhanced our headwinds, even as we have brought additional volume onto the platform through implementation of new bank partnerships.

相反，在當前氣候下，生態系統的資金方面仍然充滿挑戰。在最近的銀行倒閉之後，現有貸款合作夥伴之間更加保守，這增加了我們的阻力，即使我們通過實施新的銀行合作夥伴關係為平台帶來了更多的交易量。

The banking sector turbulence has also contributed to wider market spreads on high-yield debt after a brief period of moderation in January and February, which is in turn causing institutional investors who rely on functioning ABS markets to be increasingly cautious in their deployment of capital.

在經歷了 1 月和 2 月的短暫緩和之後，銀行業的動盪也導致高收益債券的市場利差擴大，這反過來又導致依賴 ABS 市場運作的機構投資者在部署資本時越來越謹慎。

We have managed to more than offset these headwinds by securing multiple longer-term funding agreements, which we expect to deliver more than $2 billion in capital over the coming 12 months. We believe that these deals, as well as others in the pipeline, will provide us with a stronger and more resilient capital supply over the coming quarters.

我們通過獲得多項長期融資協議成功地抵消了這些不利因素，我們預計這些協議將在未來 12 個月內提供超過 20 億美元的資金。我們相信，這些交易以及其他正在籌備中的交易將為我們在未來幾個季度提供更強大、更具彈性的資本供應。

With these items as context, here are some financial highlights from the first quarter of 2023. Revenue from fees was $117 million Q1, coming in above our guided expectations as a result of some increased funding secured through our longer-term capital arrangements as well as ongoing take rate optimization.

以這些項目為背景，以下是 2023 年第一季度的一些財務亮點。第一季度的費用收入為 1.17 億美元，高於我們的指導預期，這是由於通過我們的長期資本安排以及持續優化。

Net interest income was slightly below guidance at negative $14 million, largely a result of unrealized fair value adjustments, informed by the economics of a balance sheet transaction that was expected to close after the end of the quarter.

淨利息收入略低於指導值，為負 1,400 萬美元，這主要是未實現的公允價值調整的結果，這是根據預計將在本季度末結束的資產負債表交易的經濟學信息得出的。

Taken together, net revenue for Q1 came in at $103 million, slightly above guidance and representing a 67% contraction year-over-year. The volume of loan transactions across our platform in Q1 was approximately 84,000 loans, down 82% year-over-year and representing over 53,000 new borrowers. Average loan size of $12,000 was up 22% versus the same period last year.

總的來說，第一季度的淨收入為 1.03 億美元，略高於預期，同比下降 67%。第一季度我們平台上的貸款交易量約為 84,000 筆貸款，同比下降 82%，代表超過 53,000 名新借款人。平均貸款規模為 12,000 美元，比去年同期增長 22%。

Our contribution margin, a non-GAAP metric, which we define as revenue from fees minus variable costs for borrower acquisition, verification and servicing as a percentage of revenue from these came in at 58% in Q1, up from 47% last year, and 3 percentage points above our guided expectation for the quarter.

我們的邊際收益是一項非公認會計準則指標，我們將其定義為收費收入減去借款人獲取、驗證和服務的可變成本，佔這些收入的百分比在第一季度為 58%，高於去年的 47%，以及比我們對本季度的指導預期高出 3 個百分點。

We continued expanding our margins in Q1 through higher rates of automation, which attained the peak of 84%, improved marketing efficiency and increased take rates. Operating expenses were $235 million in Q1, down 15% year-over-year, but up 14% sequentially due to onetime restructuring costs incurred as part of our reduction in force, as well as a onetime noncash charge resulting from the cancellation of a single executive performance-based equity award, which accounted for approximately $40 million of the overall expense base in this past quarter.

我們通過更高的自動化率繼續擴大第一季度的利潤率，達到了 84% 的峰值，提高了營銷效率並提高了採納率。第一季度的運營費用為 2.35 億美元，同比下降 15%，但環比增長 14%，這是由於我們裁員產生的一次性重組成本，以及因取消單一業務而產生的一次性非現金費用。基於高管績效的股權獎勵，這在上個季度的總支出基礎中約佔 4000 萬美元。

The majority of the year-over-year reduction was achieved through sales and marketing, which declined by 76% following the trend in volume. We continue to limit hiring to only a handful of key strategic positions. Altogether, Q1 GAAP net loss was $129 million, and adjusted EBITDA was negative $31.1 million, both comfortably above our guided numbers.

同比減少的大部分是通過銷售和營銷實現的，隨著銷量的趨勢下降了 76%。我們繼續將招聘限制在少數關鍵戰略職位。總的來說，第一季度 GAAP 淨虧損為 1.29 億美元，調整後的 EBITDA 為負 3110 萬美元，均高於我們的指導數字。

Adjusted earnings per share was negative $0.47 based on a diluted weighted average share count of 81.9 million. We ended the quarter with loans on our balance sheet at $982 million, down sequentially from $1.01 billion the prior quarter. Of that total, loans made for the purposes of R&D, principally within the auto segment, represented $493 million of that total.

根據稀釋後的 8190 萬股加權平均股數計算，調整後每股收益為負 0.47 美元。本季度末，我們資產負債表上的貸款為 9.82 億美元，低於上一季度的 10.1 億美元。在這一總額中，為研發目的發放的貸款（主要在汽車領域）佔總額的 4.93 億美元。

Our corporate liquidity position remains strong with $452 million of total cash on the balance sheet and approximately $632 million in net loan equity at fair value. In our remarks of last quarter, we expressed optimism that the consumer trends impacting credit appeared positive to us and that the worst is likely behind us, and this is indeed what we have perceived over the last 90 days.

我們的企業流動性狀況依然強勁，資產負債表上的現金總額為 4.52 億美元，按公允價值計算的淨貸款權益約為 6.32 億美元。在我們上個季度的評論中，我們樂觀地表示，影響信貸的消費趨勢對我們來說似乎是積極的，最壞的情況可能已經過去，這確實是我們在過去 90 天裡所感受到的。

Consumers reducing expenditures indicates to us that they are overspending less relative to income. Decreasing numbers of job openings in the economy indicates to us that Americans are returning to work. While both of these dynamics could be interpreted as a prelude to economic slowdown, it is clear to us that both have been beneficial for credit. This continues to be our perspective as we look to Q2.

消費者減少支出向我們表明，相對於收入，他們的超支較少。經濟中職位空缺數量的減少向我們表明美國人正在重返工作崗位。雖然這兩種動態都可以被解釋為經濟放緩的前奏，但我們很清楚它們都對信貸有利。當我們展望第二季度時，這仍然是我們的觀點。

We also expressed confidence last quarter that through a combination of margin expansion, workforce reduction and expenditure control, we could create a path to return to EBITDA breakeven at our current lower scale. And indeed, we now anticipate achieving this in the second quarter.

上個季度我們還表達了信心，即通過擴大利潤率、裁員和控制支出，我們可以創造一條途徑，以目前較低的規模恢復 EBITDA 盈虧平衡。事實上，我們現在預計在第二季度實現這一目標。

While painful, the workforce reduction was carried out with minimal impact to the velocity at which we are developing our newest round of bets in auto lending, small dollar loans and HELOCs, all areas in which we look forward to sharing more updates in the coming quarters.

雖然痛苦，但裁員對我們在汽車貸款、小額美元貸款和 HELOC 領域開展最新一輪投資的速度影響微乎其微，我們期待在未來幾個季度分享更多更新的所有領域.

With these specifics in mind, for Q2 of 2023, we expect total revenues of approximately $135 million, consisting of revenue from fees of $130 million, and net interest income of approximately $5 million, contribution margin of approximately 60%, net income of approximately negative $40 million, adjusted net income of approximately negative $7 million, adjusted EBITDA of approximately 0 and a diluted weighted average share count of approximately 83.1 million shares.

考慮到這些具體情況，我們預計 2023 年第二季度的總收入約為 1.35 億美元，其中包括 1.3 億美元的費用收入和約 500 萬美元的淨利息收入，邊際貢獻率約為 60%，淨收入約為負數4000 萬美元，調整後淨收入約為負 700 萬美元，調整後 EBITDA 約為 0，稀釋後加權平均股數約為 8310 萬股。

We are happy to be signaling a return to sequential growth and cash profitability in the current market environment. Our improving guidance is clearly not deriving from obvious improvements to the macro economy just yet. It is flowing from a combination of tenacious execution, operating discipline, margin expansion and deal making.

我們很高興地表示，在當前的市場環境下，我們將恢復連續增長和現金盈利能力。我們改進的指引顯然不是來自宏觀經濟的明顯改善。它源於頑強的執行力、經營紀律、利潤率擴張和交易促成。

As always, a huge note of gratitude to the various Upstart teams who have been heads down and laser-focused on getting us through the storm, and we'll remain resolute over the past year in the face of such challenging external circumstances. While there may yet be a few more twists and turns along the way, we are optimistic that we have weathered the worst of it, and barring any reversal in consumer trends, that we are back on an ascending flight path.

一如既往，非常感謝各種 Upstart 團隊，他們一直低著頭，專注於讓我們度過難關，面對如此具有挑戰性的外部環境，我們將在過去的一年裡保持堅定。雖然一路上可能還會有一些曲折，但我們樂觀地認為我們已經度過了最糟糕的時期，並且除非消費趨勢發生任何逆轉，否則我們將回到上升的軌道上。

With that, Dave and I are happy to open up the call to any questions. Operator?

有了這個，戴夫和我很樂意打開任何問題的電話。操作員？

(Operator Instructions) Our first question comes from the line of Simon Clinch with Atlantic Equities.

（操作員說明）我們的第一個問題來自 Simon Clinch 與 Atlantic Equities 的對話。

Dave, Sanjay, congrats on a pretty good quarter here. I was wondering just -- could you give us a little bit more detail around the long-term funding commitments perhaps? And how we should think about how that might be applied or flow through for the remainder of the year? Are there any constraints, product categories that's focused on or anything like that we need to know about?

Dave、Sanjay，恭喜你在這裡度過了一個相當不錯的季度。我想知道 - 你能否就長期資金承諾向我們提供更多細節？我們應該如何考慮在今年餘下的時間裡如何應用或流動？是否有任何我們需要了解的限制、關注的產品類別或類似內容？

Yes, Simon. This is Sanjay. Let's see. I mean, I guess I would say upfront, each agreement is a bit bespoke. So it's hard to sort of broadly generalize. But maybe a couple of headline thoughts. First of all, I would say these agreements, they more or less flow from the ability we've demonstrated over the past couple of quarters to be expensive with our margins and to improve our take rates.

是的，西蒙。這是桑傑。讓我們來看看。我的意思是，我想我會提前說，每個協議都是定制的。所以很難廣泛概括。但也許有幾個標題想法。首先，我想說這些協議，它們或多或少源於我們在過去幾個季度中展示的能力，即我們的利潤率很高，並提高了我們的接受率。

And because we have expanded margins, really, the way to think about these agreements are we're able to sort of share some modest preferential economics with long-term committed investors. Those preferential economics could take the form of sort of return premiums or take rate co-investment or modest discounting in this chain. All investors have a slightly different set of preferences and objectives, so it tends to be a bit different by counterparty.

而且因為我們擴大了利潤率，真的，考慮這些協議的方式是我們能夠與長期承諾的投資者分享一些適度的優惠經濟學。這些優惠經濟可以採取某種回報溢價的形式，或者採取利率共同投資或在這個鏈條中適度貼現的形式。所有投資者的偏好和目標都略有不同，因此交易對手往往會有所不同。

They're all currently focused on personal loans. So they're sort of restricted to our core business. Beyond that, they tend to mirror our broader institutional programs. So apart from some preferential economics that essentially flow from our enhanced unit economics, they tend to look very similar to what we would normally do for the capital markets.

他們目前都專注於個人貸款。所以他們有點局限於我們的核心業務。除此之外，它們傾向於反映我們更廣泛的機構計劃。因此，除了一些本質上來自我們增強的單位經濟學的優惠經濟學外，它們看起來與我們通常為資本市場所做的非常相似。

Great. That's really helpful. And maybe a follow-up question to that. Just on your comments around take rate and I guess, the very impressive contribution margins that you generated this quarter and anticipate to generate in the next quarter. How should we think about the sustainability of that as you go through perhaps an economic recovery and loan acceleration, at some point? Should we expect those contribution margins to decline? And then what does that mean in terms of, I suppose, those funding commitments become less important in that regard, but I'd just like to think about how all that ties together.

偉大的。這真的很有幫助。也許還有一個後續問題。就您對 Take Rate 的評論，我想，您本季度產生的非常可觀的邊際收益，並預計將在下個季度產生。當你在某個時候經歷經濟復甦和貸款加速時，我們應該如何考慮它的可持續性？我們是否應該預期這些邊際收益會下降？然後這意味著什麼，我想，這些資金承諾在這方面變得不那麼重要了，但我只想考慮所有這些是如何联系在一起的。

Sure. Yes. I mean, I guess, the punchline is there, I think they're probably sustainable for as long as we want them to be. There's really 2 underlying factors I would point to. One is the extent to which we are sort of investing for the short term versus the long term, and the other is the general elasticity of the loan demand on the borrower side.

當然。是的。我的意思是，我想，妙語就在那裡，我認為只要我們希望它們存在，它們就可能是可持續的。我確實要指出兩個潛在因素。一個是我們短期投資與長期投資的程度，另一個是藉款人方面貸款需求的總體彈性。

So currently, as an example, elasticity is quite high. Credit is in demand on the borrower side. And so that creates a sort of pricing ability on our side. But maybe the more important thing is when we're sort of solving for the near-term P&L, we are optimizing take rates as much as possible against that elasticity, and we are managing our marketing programs to deliver loans that are profitable in the near term. I think as the economy evolves and certainly, as it improves, what you would see is on the one hand, a declining in elasticity, so there'll be more sort of provision of credit on the supply side and so borrowers will have more choice.

所以目前，作為一個例子，彈性是相當高的。借款人需要信貸。因此，這在我們這邊創造了一種定價能力。但也許更重要的是，當我們解決近期的損益問題時，我們正在根據這種彈性盡可能地優化利率，並且我們正在管理我們的營銷計劃以提供短期內有利可圖的貸款學期。我認為隨著經濟的發展，當然，隨著經濟的改善，一方面你會看到彈性下降，因此供應方將提供更多種類的信貸，因此借款人將有更多選擇.

But the other factor would be, in our case, we would then start to rebalance how we trade off the short term for the long term. And so by reducing take rates and by growing marketing campaigns, we can generate more volume, harvest more lifetime value of the customer, harvest more gains on the model learning side, things that won't necessarily show up in this quarter's P&L but may provide value over time.

但另一個因素是，在我們的情況下，我們將開始重新平衡我們如何權衡短期與長期。因此，通過降低採用率和增加營銷活動，我們可以產生更多的銷量，收穫更多客戶的生命週期價值，在模型學習方面收穫更多收益，這些不一定會出現在本季度的損益表中，但可能會提供隨著時間的推移價值。

And so when you put all of that together, you get an improving economy. I think you'd probably see take rates that may not go back to where they were before, but would probably moderate with the economy improving.

所以當你把所有這些放在一起時，你就會得到一個正在改善的經濟。我認為你可能會看到採取的利率可能不會回到以前的水平，但可能會隨著經濟的改善而緩和。

Our next question comes from the line of Ramsey El-Assal with Barclays.

我們的下一個問題來自 Ramsey El-Assal 與巴克萊銀行的合作。

This is John Coffey on for Ramsey. I had a question for you on your Slide 18, really, regarding your conversion rates. So it looks like your conversion rates had declined from about 21% last year to 8%. I was just trying to understand the drivers of this a little bit better. Is most of it just the supply of credit? Or is it sometimes that potential borrowers are hitting that ceiling for interest rates or they're deciding to defer some kind of purchases to later on when they're maybe a little bit more of a stable economic condition?

這是 John Coffey 代替 Ramsey。我在幻燈片 18 上有一個問題要問你，真的，關於你的轉化率。所以看起來你的轉化率從去年的 21% 左右下降到 8%。我只是想更好地理解這方面的驅動因素。大部分只是信貸供應嗎？還是有時潛在的借款人正在觸及利率上限，或者他們決定將某種購買推遲到以後經濟狀況可能更加穩定時？

Yes, John. Yes, it's got -- it doesn't have much at all to do with the supply side. I think maybe the single most explanatory variable would be what we call our Upstart Macro Index, or UMI, which is also in our investor materials. And that basically reflects the fact that apples-to-apples, the same consumer is defaulting at a rate that is maybe sort of 2x to 3x higher than they were in mid-2021.

是的，約翰。是的，它有——它與供應方沒有太大關係。我認為也許最具解釋力的變量就是我們所說的 Upstart 宏觀指數或 UMI，它也在我們的投資者材料中。這基本上反映了這樣一個事實，即同一消費者的違約率可能比 2021 年中期高出 2 到 3 倍。

And because of that consumer repayment pattern change, we are pricing loans very differently, right? We're including much higher default premiums in the loans. And then compounding that, of course, is the higher base interest rates, which are requiring investors to demand higher sort of returns. When you add those 2 up, our like-for-like price for a loan has gone up quite dramatically and in some cases, 1,500 to 2,000 basis points.

由於消費者還款模式的改變，我們對貸款的定價非常不同，對吧？我們在貸款中包括了更高的違約金。當然，還有更高的基準利率，這要求投資者要求更高的回報。當你將這 2 個加起來時，我們的貸款同類價格已經大幅上漲，在某些情況下，上漲了 1,500 到 2,000 個基點。

And because of that change in pricing, 2 things are happening. One is a lot less borrowers are getting approved, right? So a lot of them are being pushed above the 36% APR threshold. And then even for those who are still getting approved, their prices are a lot higher and maybe less predisposed to taking the loan. So those 2 things combined have resulted in the contracting conversion rate.

由於價格的變化，發生了兩件事。一是獲得批准的借款人少了很多，對吧？所以他們中的很多人都被推到了 36% 的年利率門檻之上。然後即使對於那些仍在獲得批准的人來說，他們的價格要高得多，而且可能不太願意接受貸款。所以這兩件事結合起來導致了合同轉換率。

All right. Great. And just for a very short follow-up. I noticed that your loans on your balance sheet actually declined a little bit quarter-over-quarter. Should we think of last quarter, the fourth quarter, as a bit of a high-water mark? Or is that still too early to say?

好的。偉大的。只是為了很短的跟進。我注意到你資產負債表上的貸款實際上環比下降了一點。我們是否應該將上一季度，即第四季度視為一個高水位線？或者現在說還為時過早？

Good question, John. I mean, I think that we, I think, continue to think of the balance sheet along the lines of the parameters we've expressed the market, which is there's a certain number we won't go above, and that's probably roughly where we were last quarter. We also sort of said in our remarks that there is a transaction that did not complete in Q1, but we expect it to complete in Q2, and that will bring our balance sheet down next quarter.

好問題，約翰。我的意思是，我認為我們，我認為，繼續按照我們表達市場的參數來考慮資產負債表，這是我們不會超過的某個數字，這可能是我們的大致位置是上個季度。我們也在評論中表示，有一項交易在第一季度沒有完成，但我們預計它會在第二季度完成，這將使我們的資產負債表在下個季度下降。

Now from there, we may sort of continue to use the balance sheet as a platform tool. But I think that you'll probably see us remain sort of in a volume where we are sort of peaking at the $1 billion range and then ebbing or flowing from there based on whether we're transacting or accumulating.

現在，我們可能會繼續使用資產負債表作為平台工具。但我認為你可能會看到我們仍然處於某種交易量中，我們在 10 億美元的範圍內達到頂峰，然後根據我們是在交易還是在積累，從那裡退潮或流動。

Our next question comes from the line of Peter Christiansen with Citi.

我們的下一個問題來自花旗銀行的 Peter Christiansen。

I just wanted to dig a little bit into the secured funding, the $2 billion. I just want to understand, is this -- are these funds like securing minimums from existing partners? Or is it incremental from new funding partners? If you could just provide a little bit more color on that, and how it is compared to the existing run rate, I guess, of business that you have?

我只是想深入了解一下有擔保的資金，即 20 億美元。我只想了解，這些資金是不是像從現有合作夥伴那裡獲得最低限度？還是來自新的資金合作夥伴的增量？如果你能提供更多的顏色，以及它與現有的運行率相比如何，我猜，你的業務？

Sure. Pete, this is Sanjay. Yes, I think a lot of it is coming from partners that are new to the platform that we've never worked with before. Some of it is coming from existing partners who would either sort of stepped away from funding during the turbulence of the past few quarters and are coming back to the platform or they're re-upping in significant size and committing forward in exchange for the longer-term agreement. So I think in almost every case, you can view it as sort of being additive to what we've been doing recently, and sort of underpins the -- both the improving guide that we have in our Q2 numbers as well as the sort of more general optimism we're signaling qualitatively.

當然。皮特，這是桑傑。是的，我認為其中很多來自我們以前從未合作過的平台的新合作夥伴。其中一些來自現有的合作夥伴，他們要么在過去幾個季度的動盪期間放棄融資，然後又回到平台，要么重新擴大規模並承諾向前發展，以換取更長的時間- 期限協議。因此，我認為在幾乎所有情況下，您都可以將其視為對我們最近所做工作的補充，並在某種程度上支持我們在第二季度數字中的改進指南以及那種我們正在定性地發出更普遍的樂觀情緒。

That's helpful. And then I guess, in the past, Upstart has made a number of model changes to the AI platform. And you've seen subsequent results change as well. I guess in the current status -- I guess assuming the status quo environment, like how do you think these model changes will influence results in the coming quarters?

這很有幫助。然後我猜想，Upstart 在過去對 AI 平台進行了一些模型更改。您也看到了後續結果的變化。我想在當前狀態下——我想假設現狀環境，比如你認為這些模型變化將如何影響未來幾個季度的結果？

Pete, this is Dave. I mean, generally speaking, on the margin, the upgrades to the AI tend to improve automation or improve accuracy of the models, and those tend to be, in aggregate, positive to growth. So as I was kind of saying towards the end of my remarks, we believe we can grow without the economy improving, but it will certainly be doing it against what is currently a pretty challenging economy with respect to funding markets, base credit rates, interest rates, et cetera. So -- but we do believe, I mean, the reason we've been able to grow and over time is predominantly because the technology and the models get better, and those tend to lead to growth. And even in a difficult environment as we have today, that is still true.

皮特，這是戴夫。我的意思是，一般來說，人工智能的升級往往會提高自動化程度或提高模型的準確性，而這些總體上往往對增長有利。因此，正如我在發言結束時所說的那樣，我們相信我們可以在經濟不改善的情況下實現增長，但它肯定會在目前在融資市場、基准信貸利率、利率等方面極具挑戰性的經濟中實現增長率等等。所以 - 但我們確實相信，我的意思是，我們能夠隨著時間的推移而增長的原因主要是因為技術和模型變得更好，而這些往往會導致增長。即使在我們今天所處的困難環境中，這仍然是正確的。

Our next question comes from the line of Giuliano Bologna with Compass Point.

我們的下一個問題來自 Compass Point 的 Giuliano Bologna 系列。

I was curious in looking at kind of the mix of bank versus non-bank and the institutional investors on the platform. I'd be curious if there's a sense of like where that mix may have shifted this quarter, where that might go going forward?

我很想看看平台上銀行與非銀行以及機構投資者的混合情況。我很好奇是否有這樣一種感覺，即本季度這種組合可能發生了哪些變化，未來可能會發生什麼變化？

Giuliano, you're asking about the sort of mix of banks versus institutional investors?

Giuliano，你問的是銀行與機構投資者的混合類型？

That's right. Yes.

這是正確的。是的。

Yes. I mean, I guess, look, at a headline level, I would say that, obviously, the events that we went through in March that particularly impacted the banking sector definitely had an impact. And so the dynamic there as we continue to bring new sort of lending partner, bank, credit union partners aboard, but existing ones are obviously becoming a little bit more conservative.

是的。我的意思是，我想，看，在標題層面，我會說，顯然，我們在 3 月份經歷的對銀行業產生特別影響的事件肯定產生了影響。因此，當我們繼續引入新型貸款合作夥伴、銀行、信用合作社合作夥伴時，動態就會出現，但現有合作夥伴顯然變得更加保守。

I think that impacted the capital markets less so. So you may have seen a bit of a shift towards the institutional dollars. How that plays out for us is a bit hard to tell. I mean the banking sector is obviously right now in a bit of turbulence, and that may increase or it may moderate. And I think depending on what that banking -- what that sector experiences over the coming quarters, if you look at price, it reflected it. That reflected in their ability to or their appetite to lend and to take more balance sheet risk.

我認為這對資本市場的影響較小。所以你可能已經看到了一些向機構美元的轉變。這對我們的影響如何有點很難說。我的意思是，銀行業現在顯然有點動盪，而且動盪可能會加劇，也可能會緩和。而且我認為取決於銀行業 - 該行業在未來幾個季度的經歷，如果你看價格，它會反映出來。這反映在他們放貸和承擔更多資產負債表風險的能力或意願上。

That makes sense. And then kind of looking through some of the past disclosures. You kind of have 2 banks that seem to be the primary enablers of the loan sales or kind of the marketplace side of the platform. But if you add up the volumes of those 2 banks, it seems to imply that they're actually much more than the loans that are being sold by the institutional investors and retained by Upstart. It could imply that 1 or 2 of those banks accounts for 50-plus or even close to 60% of all of your kind of bank funding side of the platform. So I'm curious what level of concentration there is on the bank side of the funding platform, kind of have you seen any of those partners pull back in their buying activity recently?

這就說得通了。然後瀏覽一些過去的披露。您有 2 家銀行，它們似乎是貸款銷售或平台市場方面的主要推動者。但是，如果將這兩家銀行的交易量加起來，這似乎意味著它們實際上遠遠超過機構投資者出售並由 Upstart 保留的貸款。這可能意味著其中 1 家或 2 家銀行佔該平台所有此類銀行融資方的 50% 以上，甚至接近 60%。所以我很好奇融資平台銀行方面的集中程度如何，你有沒有看到這些合作夥伴中的任何一個最近撤回了他們的購買活動？

Yes. There are really 3 banks in our platform that serve as conduits where they originate loans; in some cases, they hold some of them; and in other cases, they end up selling them to institutional partners. So there are 3 of those banks on our platform. And it's quite possible to move volumes between them. It's quite purposeful that we can do that.

是的。我們的平台上實際上有 3 家銀行充當發放貸款的渠道；在某些情況下，他們持有其中的一些；在其他情況下，他們最終將它們出售給機構合作夥伴。所以我們的平台上有 3 家銀行。並且很有可能在它們之間移動卷。我們這樣做是有目的的。

Generally speaking, the fees and the profits on the loans that end up in the institutional side are higher than they are on the bank -- on the what tends to be the primary loans that banks are originating and holding on their balance sheet. So that sort of creates for the structure that you're seeing there. But it's not really, in a true sense of revenue concentration, these are, again, many, many institutional buyers are behind those banks, but we do have multiple what we call marketplace lending partners on the platform.

一般來說，最終出現在機構方面的貸款的費用和利潤高於銀行——這些往往是銀行發起並持有在其資產負債表上的主要貸款。所以那種為你在那裡看到的結構創造的。但這並不是真正意義上的收入集中度，這些銀行背後有很多很多機構買家，但我們在平台上確實有多個我們稱之為市場借貸的合作夥伴。

Got it. And I'm curious, when you think about funding concentration with any partners, are there any partners that are kind of 10% or more of your overall funding or 15% or more of your overall funding? Or is it more diversified than that?

知道了。我很好奇，當你考慮與任何合作夥伴的資金集中度時，是否有任何合作夥伴佔你總資金的 10% 或更多，或者佔你總資金的 15% 或更多？或者它比那更多樣化？

Giuliano, yes, I think what you'll see in our revenue concentration disclosures when the financials come out is that those concentrations are going down? And I think in terms of source capital dependency, I don't think anything is far greater than the ballpark that you just mentioned.

朱利亞諾，是的，我想當財務數據公佈時，你會在我們的收入集中度披露中看到這些集中度正在下降？而且我認為就來源資本依賴而言，我認為沒有什麼比你剛才提到的範圍大得多。

The next question comes from the line of James Faucette with Morgan Stanley.

下一個問題來自 James Faucette 與 Morgan Stanley 的對話。

I wanted to go back on the secured funding. I think you made the comment that there were some preferential terms that you were able to offer to them. I guess a couple of questions associated with that. Going forward, what would be the right level of mix or targeted mix for that group? And it sounds like you're planning to add more to that, firstly. And secondly, how much should we expect that you may have to harmonize at least some of those terms into other sources of capital?

我想重新獲得擔保資金。我想你曾評論說你可以向他們提供一些優惠條件。我想有幾個與此相關的問題。展望未來，該群體的正確組合水平或目標組合是什麼？聽起來您打算首先添加更多內容。其次，我們應該在多大程度上期望您可能必須將至少其中一些條款與其他資本來源協調一致？

James, this is Dave. Good question. So I would say, first of all, one of the comments I had made in my remarks earlier was that we would like to have, in any particular month, capital committed such that we can be cash flow positive as a business. So sort of a baseline of capital that we feel very good about being solid will be there, and that would be a pretty dramatic reduction in cyclicality. So that's, I think, what we would try to do.

詹姆斯，這是戴夫。好問題。所以我想說，首先，我在之前的發言中發表的其中一項評論是，我們希望在任何特定月份投入資本，以便我們可以作為一家企業實現正現金流。因此，我們對穩固感到非常滿意的某種資本基線將會存在，這將大大減少週期性。所以，我認為，這就是我們要嘗試做的事情。

I generally think we are benefiting from the fact that we have very strong margins in that we can share a bit of them with the preferred partner, and a preferred partner is one that is making a longer-term commitment to us. So I think that's a structure that we would expect to have for the long haul.

我通常認為我們受益於我們擁有非常高的利潤率這一事實，因為我們可以與首選合作夥伴分享其中的一部分，而首選合作夥伴是對我們做出長期承諾的合作夥伴。所以我認為這是我們希望長期擁有的結構。

And I think it makes sense that if they're going to commit to you, you're going to make some special commitment to them. And there can all -- of course, also continue to be plenty of sort of at-will participants in the marketplace month in, month out. And we don't expect that to go away. We think it's a good thing. But I do think we want a significant fraction of the funding to be more long-term committed and getting a little bit of a preference for doing that.

我認為，如果他們要對你做出承諾，你就會對他們做出一些特殊的承諾，這是有道理的。並且可以 - 當然，也可以繼續在市場上月復一月地有大量的隨意參與者。我們不希望這種情況消失。我們認為這是一件好事。但我確實認為我們希望將很大一部分資金用於更長期的承諾，並在這樣做時獲得一點偏好。

And lastly, as we've said, I think we feel comfortable, given our margin structure, that we can afford to do that. The other thing I'll add is this is all, as Sanjay said earlier, entirely related to personal loans today. I think we will push for this kind of structure on other products as well. I kind of believe that secured products are ones that are much primer. Like the HELOC product, we'll have probably, I would just say, less of a need for it because they would just tend to be more normal familiar products and ones that lenders lean toward even in more difficult times.

最後，正如我們所說，鑑於我們的保證金結構，我認為我們有能力做到這一點。我要補充的另一件事是，正如 Sanjay 之前所說，這完全與今天的個人貸款有關。我認為我們也會在其他產品上推動這種結構。我有點相信安全產品是入門級產品。就像 HELOC 產品一樣，我想說的是，我們對它的需求可能會減少，因為它們往往是更普通的熟悉產品，即使在更困難的時期，貸款人也會傾向於使用這些產品。

So that's how we're thinking about it. I don't know if -- I think there will always be kind of -- we'll hopefully harmonize these types of agreements into a consistent structure. Sanjay said they're starting off a little bespoke by -- probably not hard to imagine that, that would happen. But over time, I think we would really like to almost programmatize it so it's a little more structured. But I do believe there'll be sort of longer-term partners and there'll be at-will partners coexisting in our platform.

這就是我們考慮的方式。我不知道是否——我認為總會有——我們希望將這些類型的協議協調成一個一致的結構。 Sanjay 說他們開始有點定制——可能不難想像，那會發生。但隨著時間的推移，我認為我們真的很想將其程序化，使其更加結構化。但我確實相信會有一些長期合作夥伴，並且會有隨意合作夥伴在我們的平台中共存。

Yes. And then on -- I appreciate all that, Dave. And then how are you feeling about the state of your cost base now on a run rate basis? I know that here in the in the June quarter, you're expecting to roughly be adjusted EBITDA breakeven now that those programs have been fully implemented.

是的。然後——我很感激這一切，戴夫。然後，您如何看待現在按運行率計算的成本基礎狀況？我知道，在 6 月份的這個季度，由於這些計劃已經全面實施，您預計會粗略調整 EBITDA 盈虧平衡點。

But should we take that to mean that you feel like you're at the right cost base and that the kind of the first quarter would have been roughly the bottom from a revenue generation perspective and that you can grow at least for the remainder of this calendar year on a sequential basis? Or are there incremental actions that may need to be taken? Just trying to think through kind of how you're seeing the business and its evolution from this point.

但是我們是否應該認為這意味著您覺得自己處於正確的成本基礎上，並且從創收的角度來看，第一季度的成本大致處於底部，並且您至少可以在本季度的剩餘時間內實現增長按順序計算日曆年？或者是否有可能需要採取的漸進行動？只是想通過某種方式思考您如何看待業務及其從這一點開始的演變。

James, this is Sanjay. Yes, I think we're in very good shape. I think from a personnel perspective, I think we did what we needed to do in Q1, and we're positive. I think our workforce numbers are in very good shape now. And barring any dramatic reversal in the business and the economy, I think we'll sort of grow from here.

詹姆斯，這是桑傑。是的，我認為我們的狀態非常好。我認為從人員的角度來看，我認為我們在第一季度做了我們需要做的事情，我們是積極的。我認為我們的員工人數現在狀況非常好。除非商業和經濟出現任何戲劇性逆轉，否則我認為我們會從這裡開始增長。

Look, there's still more work to be done on the OpEx side. As Dave mentioned, there's sort of programs to sort of reduce. We're large consumers of compute. Our machine learning models require a lot of sort of training resources, and we want to get more efficient at that, our engineering footprint. The resources we consume can get more efficient as well. So I think that this will be an ongoing set of initiatives.

看，在 OpEx 方面還有更多工作要做。正如 Dave 提到的，有一些程序可以減少。我們是計算的大消費者。我們的機器學習模型需要大量的培訓資源，我們希望在這方面提高效率，我們的工程足跡。我們消耗的資源也可以變得更有效率。所以我認為這將是一系列持續的舉措。

But I think by and large, your sort of -- your model is right, which is I think we're at a good place. We can tighten up a little bit more, but we're sort of indicating that we think there's a nice path for growth from here on out. And as we demonstrated before, our cost base is going to scale very efficiently with the top line, and it will sort of provide nice margin materialization as we grow.

但我認為總的來說，你的那種 - 你的模型是正確的，我認為我們處在一個好地方。我們可以收緊一點，但我們有點表明我們認為從現在開始有一條很好的增長道路。正如我們之前所展示的那樣，我們的成本基礎將隨著收入的增加而非常有效地擴展，並且隨著我們的成長，它將在某種程度上提供不錯的利潤實現。

Our next question comes from the line of Reggie Smith with JPMorgan.

我們的下一個問題來自摩根大通的 Reggie Smith。

Congrats on the quarter and on securing the $2 billion in funding. My question, I know in the past, you've been asked about becoming a bank and not wanting to do that, which I can totally appreciate. My question for you, and I guess it kind of relates back to the funding as well. Would you consider like a warehouse-type funding mechanism? And does any of your $2 billion, is any of it structured in that kind of way where you pay a fixed rate and you hold more loans on balance sheet? Or is all that $2 billion that you talked about -- is all that $2 billion that you talked about arm's length off-balance sheet financing?

祝賀本季度並獲得 20 億美元的資金。我的問題是，我知道過去曾有人問過你關於成為一家銀行但不想這樣做的問題，我完全可以理解。我想問你的問題，我想這也與資金有關。你會考慮像倉庫式的融資機制嗎？你的 20 億美元中的任何一個，是否以支付固定利率並在資產負債表上持有更多貸款的方式構建？還是您所說的所有 20 億美元——您所說的所有 20 億美元都是公平的資產負債表外融資？

Yes. Reggie, yes, that's generally what we would term forward flow commitment, so loans that are purchased monthly by a third party, an outside third party. So that's what that is. It's not on our balance sheet. There's no warehousing on our side. They can be using -- they could be leveraging on their side if they chose to, but that's separate from anything going on with us. So let me see what was the other part of the question?

是的。 Reggie，是的，這通常就是我們所說的遠期流量承諾，即由第三方（外部第三方）每月購買的貸款。就是這樣。它不在我們的資產負債表上。我們這邊沒有倉庫。他們可以使用 - 如果他們願意，他們可以利用自己的一面，但這與我們發生的任何事情都是分開的。那麼讓我看看問題的另一部分是什麼？

So would you consider warehouse? Is that also something that you would have to pursue?

那麼你會考慮倉庫嗎？這也是你必須追求的東西嗎？

Yes. We've had modest warehouse capacity for many years. So when the loans are on our balance sheet, to some extent, we have -- those are financed and it's more efficient use of our equity capital. But that's still -- that's under that umbrella of the loan total. As Sanjay said, we aim to stay at no more than $1 billion on our balance sheet, and those tend to be warehoused or at least some of them are. So we do use that for capital efficiency, but it's not a primary funding mechanism just as our balance sheet ultimately is not a primary funding mechanism.

是的。多年來，我們的倉庫容量一直不大。因此，當貸款出現在我們的資產負債表上時，在某種程度上，我們已經為這些貸款提供了資金，並且可以更有效地利用我們的股本。但這仍然是——那是在貸款總額的保護傘下。正如 Sanjay 所說，我們的目標是將資產負債表上的資金保持在不超過 10 億美元的水平，而且這些資金往往是入庫的，或者至少其中一些是入庫的。所以我們確實使用它來提高資本效率，但它不是主要的融資機制，就像我們的資產負債表最終不是主要的融資機制一樣。

Got it. No...

知道了。不...

The loans on our balance sheet, when we say we have $1 billion of assets on the balance sheet, less than half of that is the actual loan equity of ours, the rest is financed. But we still consider that to be our balance sheet. It's really -- what we're interested in with the long-term arrangements is third parties as the risk engine, and they can finance this as it suits them.

我們資產負債表上的貸款，當我們說我們的資產負債表上有 10 億美元的資產時，其中不到一半是我們的實際貸款權益，其餘部分是融資。但我們仍然認為這是我們的資產負債表。這真的 - 我們對長期安排感興趣的是第三方作為風險引擎，他們可以根據自己的需要為其提供資金。

Yes. No, I appreciate that. The -- I guess the crux of the question was, would you consider being more, I guess, balance sheet intensive. But it sounds like you're comfortable with $1 billion, that's kind of where you want to remain, at least for the foreseeable future.

是的。不，我很感激。我想問題的癥結在於，我想你會考慮增加資產負債表密集度嗎？但聽起來你對 10 億美元感到滿意，這是你想留在那裡的地方，至少在可預見的未來。

My second question, guidance, very strong numbers. In particular, I guess, the interest -- net interest and kind of fair value. I guess that implies a pretty sharp sequential improvement there. I was curious what was driving that. And if you could talk a little bit about the performance you're seeing of your loans that are held on balance sheet, loss rates and things of that nature.

我的第二個問題，指導，非常強大的數字。我想，尤其是利息——淨利息和某種公允價值。我想這意味著那裡有相當明顯的連續改進。我很好奇是什麼驅動了它。如果你能談談你看到的資產負債表上持有的貸款的表現、損失率和類似性質的事情。

Yes. Sure, Reggie. This is Sanjay. So as far as the guidance, I would say that Q1 was the anomaly in a sense, which is normally, you'd expect your balance sheet to have some modest positive income. We tend to hold our loans at fair value, so we mark-to-market. We don't use CECL accounting. And so when interest rates are going up, which has happened a lot in the last year, it's taken a toll on the valuations.

是的。當然，雷吉。這是桑傑。所以就指導而言，我會說第一季度在某種意義上是異常的，通常情況下，你會期望你的資產負債表有一些適度的正收入。我們傾向於以公允價值持有我們的貸款，因此我們按市值計價。我們不使用 CECL 會計。因此，當利率上升時（去年經常發生這種情況），會對估值造成影響。

Q1 in particular, didn't necessarily have adverse sort of interest rate movements, but what it did have is we book unrealized fair value marks that reflected a balance sheet transaction. Now it was not a balance sheet transaction that completed in Q1. So that's why it's unrealized.

特別是第一季度，不一定有不利的利率變動，但它確實有我們記錄反映資產負債表交易的未實現公允價值標記。現在它不是在第一季度完成的資產負債表交易。所以這就是為什麼它沒有實現。

But because it was a balance sheet transaction that we were anticipating, we were expecting and we expect it to close early in Q2, we sort of reflected the economics of that transaction in Q1, and that will be in our disclosures. So I think you could think of it as of Q2 going forward, we're not really anticipating any large significant transactions beyond the one we've already accounted for.

但因為這是我們預期的資產負債表交易，我們期待並且我們預計它會在第二季度早些時候結束，我們在某種程度上反映了第一季度該交易的經濟性，這將在我們的披露中。所以我認為你可以把它看作是第二季度的未來，我們並沒有真正期待任何超出我們已經核算的交易之外的重大交易。

And therefore, I think what you're seeing now in our guidance for interest income just sort of reflects ongoing normality. It's sort of like income rate -- sorry, interest rate stability, no large transactions and some modest income from our remaining balance sheet. So that's how I would think about the guidance.

因此，我認為你現在在我們的利息收入指南中看到的只是某種程度上反映了持續的常態。這有點像收入率——抱歉，利率穩定，沒有大筆交易，我們剩餘資產負債表的收入不多。這就是我對指南的看法。

As for performance, I think the best way to think about it is in our investor materials, we generally display what we're intending to do as a blended average across our platform with respect to gross return delivery and how we think each vintage is trending. The simple summary is vintages that are on our balance sheet from back in 2021 or 2022 are likely going to under-deliver as they are across the broader platform.

至於表現，我認為最好的思考方式是在我們的投資者材料中，我們通常會展示我們打算在整個平台上做的關於總回報交付的混合平均值以及我們認為每個年份的趨勢.簡單的總結是，我們資產負債表上的 2021 年或 2022 年的年份可能會交付不足，因為它們跨越了更廣泛的平台。

I think anything as of certainly Q4 forward, so anything from the last 6 to 9 months, are well on track to over-deliver in our opinion. And so the performance of our balance sheet will just reflect the underlying vintages that you see on the broader platform.

我認為，從第四季度開始，過去 6 到 9 個月的任何事情，在我們看來都有望超額交付。因此，我們資產負債表的表現將僅反映您在更廣泛平台上看到的基礎年份。

Got it. And if I could sneak one more in. On the take rate, you talked about, I guess, stronger economics. Curious, is there a way to kind of parse the impact of stronger pricing origination fees that are paid by borrowers versus maybe what your partners are paying you? Was there any change there? Or can we talk about better pricing?

知道了。如果我能再偷偷溜進來一次。關於採納率，我猜你談到了更強大的經濟學。很好奇，有沒有一種方法可以分析借款人支付的更高定價發起費與您的合作夥伴支付給您的費用的影響？那裡有什麼變化嗎？或者我們可以談談更好的定價嗎？

I don't think there's anything necessarily more to parse there other than fees primarily borne by borrowers that are getting loans funded through institutions tend to flex up at times like this. So that's really what you're seeing. I mean base yield requirements are going up, loss expectations are going up and also the fees are going up, all of which we certainly don't love, but it's the reality of the economy we're sitting in that we have to play in. And we would really be happy to see all that reverse itself over the next year if we're so fortunate.

我認為除了主要由通過機構獲得貸款的借款人承擔的費用外，沒有什麼需要進一步分析的了，在這種情況下往往會出現這種情況。這就是你所看到的。我的意思是基本收益率要求在上升，損失預期在上升，費用也在上升，我們當然不喜歡所有這些，但這是我們所處的經濟現實，我們必須參與其中. 如果我們幸運的話，我們真的很高興看到所有這些在明年發生逆轉。

No, I got it. That wasn't clear, I'm sorry. I got a couple of calls going on. So I didn't catch that part that it was primarily borrower fees. Okay.

不，我明白了。沒說清楚，抱歉。我接到了幾個電話。所以我沒有註意到這部分主要是藉款人的費用。好的。

Our last question comes from the line of Simon Clinch with Atlantic Equities.

我們的最後一個問題來自 Simon Clinch 與 Atlantic Equities 的對話。

Just curious, going back to your -- the pace of your new model rollouts or updates through the quarter was pretty staggering. And I'll just -- someone who doesn't really know about these things, I was wondering if you could talk us through perhaps the risks or challenges of that kind of pace of rollout, and what to -- I guess, how to think about that in terms of the benefit going forward. Because previously, we haven't needed that many rollouts to create significant benefits for you guys.

只是好奇，回到你的——你的新模型在整個季度推出或更新的速度非常驚人。我只是 - 一個並不真正了解這些事情的人，我想知道你是否可以告訴我們這種推出速度的風險或挑戰，以及 - 我想，如何考慮到未來的好處。因為以前，我們不需要那麼多的推出來為你們創造顯著的好處。

Yes, Simon, it's a good question. It's important, I would just say to state that we have different models in production for each product that we have. So these 4 products, 2 auto products, a personal loan product, small dollar loan product, each of whom have related but different models that push to production. There's also models that are more focused on automation than the pricing. So things that deal with fraud and those kind of things.

是的，西蒙，這是個好問題。重要的是，我只想說我們擁有的每種產品都有不同的生產模型。所以這4個產品，2個汽車產品，1個個人貸款產品，小額貸款產品，每個產品都有相關但不同的模型推向生產。還有一些模型更注重自動化而不是定價。所以處理欺詐和這類事情的事情。

So all across -- I don't know exactly how many AI models, distinctive AI malls we have, but it's quite a few, and those teams are working in parallel. So this isn't the same model being updated or retrained every 3 days. It's less than that. But there's a lot of -- and each one of them generally are making some part of our product line that much better. So that's a pace which a couple of years ago, we were probably maybe doing one a month or so. So it's quite a difference.

所以總的來說——我不知道我們有多少人工智能模型，有多少獨特的人工智能購物中心，但數量不少，而且這些團隊正在並行工作。所以這與每 3 天更新或重新訓練的模型不同。比那還少。但是有很多 - 他們中的每一個通常都在使我們產品線的某些部分變得更好。所以這是幾年前的速度，我們可能一個月左右做一次。所以這是一個很大的不同。

This concludes today's question-and-answer session. I will turn the call back for any additional or closing remarks.

今天的問答環節到此結束。如果有任何補充或結束評論，我將轉回電話。

Just want to thank everybody for listening today. And as we've said, we're happy with what we've achieved in the first quarter. We're actually pretty optimistic about 2023. So thanks for sticking with us.

只想感謝大家今天的收聽。正如我們所說，我們對第一季度取得的成就感到滿意。我們實際上對 2023 年非常樂觀。所以感謝您堅持我們。

This concludes today's call. Thank you for your participation, and you may now disconnect.

今天的電話會議到此結束。感謝您的參與，您現在可以斷開連接。