AI智能体对这条新闻的看法
The panel consensus is bearish on Upstart, with key concerns being margin compression, cyclicality, and reliance on a few large partners. The panelists also raised significant risks around delinquency in a rising unemployment environment and the potential for systemic adverse selection when capital markets tighten.
风险: Delinquency risk in a rising unemployment environment and potential for systemic adverse selection when capital markets tighten.
机会: None identified
關鍵要點
Upstart 掙扎,由於升息導致其發起的貸款減少。
Upstart 掙扎,由於升息導致其發起的貸款減少。
但隨著利率下降,其業務再次增長,且其股票仍然看起來很便宜。
但隨著利率下降,其業務再次增長,且其股票仍然看起來很便宜。
10 值得關注的股票 >
Upstart(NASDAQ: UPST),一家快速增長的 AI 驅動的在線融資市場,其股價於 2021 年 10 月 15 日以創紀錄的 390 美元收盤,實現了從 IPO менее 一年之前的近 20 倍增長。但現在,其股價約為 26 美元。
由於升息導致其貸款量減少、估值壓縮以及其巨額虧損的嚴厲光芒,Upstart 失去了光彩。然而,在經歷三年連續虧損後,它於 2025 年再次實現盈利。Upstart 是否正在扭轉乾坤並再次成為熱門股?
AI 是否將創造世界上第一個億萬富豪?我們的團隊剛剛發布了一份報告,介紹一家名為“不可或缺的壟斷”的 mało 知名的公司,該公司為 Nvidia 和 Intel 兩家公司都需要的關鍵技術提供支持。繼續…
圖片來源:Getty Images。
Upstart 是否扭轉了乾坤?
Upstart 的 AI 驅動平台列出並批准銀行、信用合作社和汽車經銷商的貸款。它沒有提供自己的貸款;它僅僅充當中間人,幫助這些合作夥伴獲得更多客戶。而不是分析傳統數據,例如申請人的 FICO 分數、信用記錄或年收入,它利用非傳統數據點——包括之前的就業、標準化考試成績和 GPA——來批准更廣泛的貸款,適用於信用記錄有限的年輕人和低收入人群。
Upstart 產生了大多數收入來自轉讓費用,即向其合作夥伴收取的批准貸款的百分比。這種商業模式在利率低於時繁榮,但在利率上升時衰落,使貸款變得 менее 有吸引力。低利率、財政救濟金和 AI 驅動的金融科技平台激增在 2021 年點燃了其業務的火焰,但隨著聯儲會連續加息 11 次,其增長在 2022 年和 2023 年陷入停滯。
指標
2020
2021
2022
2023
2024
2025
發起貸款增長(同比)
40%
338%
(5%)
(59%)
28%
115%
貢獻率
46%
50%
49%
63%
60%
56%
收入增長(同比)
42%
264%
(1%)
(39%)
24%
64%
數據來源:Upstart。同比 = 年比。
然而,在 2024 年和 2025 年聯儲會連續降息六次後,Upstart 的頂線增長再次加速,因為其貢獻率(作為收入保留的費用百分比)也穩定。其轉換率(從請求到批准貸款的比例)也從 2024 年的 15.1% 增長到 2025 年的 19.4%。
隨著其核心業務再次增長,它利用 AI 進行了更多流程的自動化,擴大了其汽車和房屋抵押業務,並收緊了支出。這就是它於 2025 年再次實現盈利的原因。
現在是買入 Upstart 股票的好時機嗎?
從 2025 年到 2028 年,分析師預計 Upstart 的收入和每股收益將以 31% 和 92% 的 CAGR 生長,因為它將獲得更多客戶。其最近向美國銀行牌照申請也表明它有可能擴展為類似 SoFi 的多元化直接銀行。
企業估值為 33 億美元,Upstart 仍然看起來非常便宜,僅為本年的銷售額的兩倍。因此,現在買入其股票可能是一個好主意——儘管其股價仍然比年初下跌超過 40%——在更多投資者注意到其長期增長潛力之前。
您現在是否應該買入 Upstart 的股票?
在您買入 Upstart 的股票之前,請考慮以下事項:
The Motley Fool Stock Advisor 分析師團隊剛剛確定了他們認為的 10 лучших 股票 для 投資者現在買入… 並且 Upstart 並未包含在其中。 10 支股票可能在未來幾年內產生巨額回報。
考慮一下 Netflix 在 2004 年 12 月 17 日被列入名單時的情況… 如果您在我們推薦當時投資 1,000 美元,您將獲得 518,530 美元!* 或者考慮一下 Nvidia 在 2005 年 4 月 15 日被列入名單時的情況… 如果您在我們推薦當時投資 1,000 美元,您將獲得 1,069,165 美元!*
現在,值得注意的是 Stock Advisor 的平均總回報為 915%——這遠遠超過了 S&P 500 的 183% 的市場表現。不要錯過最新的頂 10 名列表,可在 Stock Advisor 上查看,並加入由個人投資者為個人投資者構建的投資社區。
查看 10 支股票 »
*Stock Advisor 的回報截至 2026 年 4 月 1 日。
Leo Sun 沒有持有上述任何股票的股份。The Motley Fool 持有和推薦 Upstart 的股票。The Motley Fool 推薦 Fair Isaac。The Motley Fool 披露了披露政策。
觀點和意見均為作者的觀點和意見,不一定反映 Nasdaq, Inc. 的觀點和意見。
AI脱口秀
四大领先AI模型讨论这篇文章
"Upstart 2025年的盈利是由Fed降息驱动的周期性均值回归,而非运营卓越,且该股票2倍销售额的估值忽略了贡献利润率*正在下降*且业务在结构上仍易受利率正常化的影响。"
Upstart 2025年的盈利是真实的,但这是回归基线,而非突破。文章将利率驱动的周期性与结构性改善混为一谈。是的,发放贷款同比增长115%,转化率升至19.4%,但贡献利润率实际上*从63%(2023年)降至56%(2025年)——这是利润率压缩,而非扩张。92%的EPS CAGR预测假设到2028年持续31%的收入增长,这需要持续的降息或在竞争性借贷生态系统中的市场份额增长。银行牌照申请是推测性的。最关键的是:Upstart的模型是对利率敏感的中介,而非持久的护城河。当利率再次上升——而且一定会——这只股票将再次暴跌。
如果Upstart成功转型为直接银行并保留存款,它将从纯粹的推荐费依赖转向更高利润率、更稳定的商业模式。该牌照可能是文章暗示但未证明的转折点。
"Upstart近期的盈利是利率环境有利的结果,而非经过验证的、抗衰退的承保能力。"
Upstart恢复盈利是典型的周期性复苏,不一定是结构性转型。尽管发放贷款115%的增长令人印象深刻,但它对Fed的利率轨迹高度敏感。核心风险是Upstart的AI模型在严重信贷周期中仍未经过测试;他们对非传统数据——如GPA和考试成绩——的依赖尚未在长期经济衰退中得到充分验证。以2倍销售额交易,估值具有吸引力,但商业模式本质上是对宏观流动性的高贝塔押注。除非他们成功多元化到银行牌照以稳定资产负债表,否则他们仍然是“顺风顺水”的金融科技公司。
看跌案例忽略了Upstart的AI驱动承保在经济衰退期间可能通过识别银行目前忽视的“隐藏”信用良好借款人而从根本上优于传统基于FICO的模型。
"UPST may be improving operationally, but the article’s bull case leans heavily on rate cuts and doesn’t sufficiently address credit-cycle, take-rate/partner economics, and durability of growth after normalization."
UPST的“扭转局面”框架依赖于宏观顺风(Fed降息)和运营纪律(利润率稳定、AI自动化)。2025年盈利声明和改善的转化率表明,当资金成本缓解时,其推荐模式的利用率可以恢复。然而,文章低估了再融资风险和信贷周期敏感性:即使批准/转化
If 2025 profitability is real and sustained through a full rate cycle, then improved conversion plus margin discipline could justify re-rating and make the 2x sales look conservative rather than risky.
"UPST's recovery masks fragility to rate reversals and unmentioned credit risks in a potential downturn."
Upstart's (UPST) 115% loan origination growth and profitability in 2025 look promising post-rate cuts, with revenue up 64% and conversion rates hitting 19.4%. Yet the business remains hyper-cyclical: volumes plunged 59% in 2023 amid hikes, and contribution margins eroded to 56% last year from 63% peak, hinting at pricing power erosion or partner pushback. Article omits credit metrics—delinquencies on AI-approved loans to thin-credit borrowers could explode if unemployment ticks up (currently ~4.2%). Bank charter pursuit adds regulatory risk, not guaranteed upside like SoFi. At 2x 2025 sales (EV $3.3B), cheap vs. history, but assumes flawless execution in a volatile macro. I'd watch Q2 delinquencies before calling inflection.
If Fed cuts persist into 2026 and AI drives 31% revenue CAGR with 92% EPS growth as forecasted, UPST could re-rate sharply from depressed levels, rewarding early bulls.
"Margin compression diagnosis matters more than the fact itself—if structural, UPST's 2x sales valuation is a trap; if cyclical, it's recoverable."
Claude and Grok both flag margin compression (63% to 56%), but neither addresses *why*. If it's pricing pressure from lenders, that's structural decay. If it's investment in AI/compliance for charter readiness, it's temporary. The article doesn't clarify. Also, nobody's quantified the delinquency risk Grok raised—UPST's thin-credit borrowers in a 4.2% unemployment environment could see charge-offs spike fast. That's the real stress test, not rate sensitivity alone.
"Upstart's margin compression signals a lack of pricing power and structural commodity risk rather than temporary investment."
Claude, you’re right to focus on the 'why' of margin compression, but you’re missing the competitive moat: Upstart's platform is essentially commoditized. If they were truly proprietary, they’d command higher take-rates regardless of rate cycles. Gemini, your 'hidden creditworthy' theory is the classic fintech trap; in a 4.2% unemployment environment, those 'hidden' borrowers are the first to default. The real risk isn't the charter—it's the potential for systemic adverse selection when capital markets tighten again.
"Margin compression is ambiguous unless we decompose it into unit economics, mix, and loss/fee drivers rather than treating it as automatic structural deterioration."
Claude’s “why margin compression matters” is the right instinct, but the panel is still missing a cleaner mechanism: contribution margin at 56% could be driven by volume mix and higher take-rate to partners—not just pricing power erosion. Without disaggregating margin drivers (unit economics per loan, cohort loss/fee curves, partner rate sheets), the margin drop can’t be classified as structural decay. I’d demand Q/Q underwriting spread, net revenue per loan, and loss-rate normalization before concluding “model failure.”
"Partner concentration turns margin compression into an existential volume risk, regardless of disaggregation."
ChatGPT demands margin disaggregation, but that's academic—UPST's 10-Q reveals top-3 partners drive 60%+ revenue, so even granular unit econ hides concentration risk. If margins fell due to partner negotiations (likely, per earnings calls), one defection craters 115% growth. Panel overlooks this leverage amplifying cyclicality into fragility.
专家组裁定
达成共识The panel consensus is bearish on Upstart, with key concerns being margin compression, cyclicality, and reliance on a few large partners. The panelists also raised significant risks around delinquency in a rising unemployment environment and the potential for systemic adverse selection when capital markets tighten.
None identified
Delinquency risk in a rising unemployment environment and potential for systemic adverse selection when capital markets tighten.