UpstartUPST
$2.7B market capUses AI models to approve personal loans that banks and investors then fund.
Upstart in 36 checks
Upstart at a glance: how it scores on value, growth, quality, health, shareholder returns and trend. The fuller the shape, the stronger the company. Behind each axis are six pass-or-fail checks from its filings, spelled out in the chapters below. Point at an axis to see them.
The business itself is the question here - 5 of 36 checks passed.
Value
What you pay today for what the business produces, measured against UPST's own history and its peers, never a universal rule.
Expensive against its own history and its sector - you're paying up for what you get.
What cash return does the business throw off per dollar of market value?
-10.6%FCF yield today
◌ 2026 = at the latest close (2026-09-04), over the trailing twelve months; earlier years are annual averages
→At -10.6%, you get less cash per dollar of market value than the 5-year median of -2.4% - the market is charging more for the same cash.
What does “Value” actually mean?
Value asks what you pay for each dollar of earnings, sales or cash the business produces. A high multiple isn't automatically bad - fast growers earn theirs - but a price far above the company's own history means the market expects a lot to go right. We compare each multiple to the company's own past and to its sector, never to a universal rule.
How we scored it · 0 of 6 checks passed
| – | Cheaper than its own history (earnings) | no multiple history |
| ✕ | Earnings yield beats a long bond (4%) | 2.0% vs 4.0% |
| ✕ | Better cash yield than its own history | -10.6% vs -2.4% |
| ✕ | Free cash flow yield above 3% | FCF yield -10.6% |
| – | Cheap on enterprise value | EBITDA unavailable |
| ✕ | Price isn't outrunning growth | no positive three-year earnings growth behind the price |
Growth
What the company has actually reported - is it selling more, and is more of it becoming profit?
Growing, but with caveats - revenue +46.3% over the last year.
Revenue: is the business selling more than it used to?
◌ 2026 = trailing twelve months to the latest filed quarter, not a full fiscal year
→Revenue reached $1.0B in 2025, compounding +5% a year since 2022 though the pace has cooled. The trailing twelve months are already running at $1.2B, ahead of the last full year.
Net income: how much of that revenue becomes profit?
◌ 2026 = trailing twelve months to the latest filed quarter, not a full fiscal year
→Net income was $54M in 2025, against $-129M the year before. Trailing twelve-month profit stands at $60M.
How fast is it growing, year by year?
+64%revenue growth, FY 2025
Shown separately because they would flatten the axis: 2021 earnings +2164% - rebounds off a collapsed prior year.
→Revenue grew +64% in 2025. Each point is one year's change against the year before.
What does “Growth” actually mean?
Growth here is what the company has actually reported to the SEC - not a forecast. We look at the last year, the three-year pace, whether growth is speeding up or slowing down, and whether it's been consistent rather than one lucky year. Hatched bars are fourth quarters we derived from annual filings (companies file a full-year 10-K rather than a Q4 report).
How we scored it · 4 of 6 checks passed
| ✓ | Outgrew its sector last year | 46.3% vs 34.1% (sector 70th pct, n=73) |
| ✕ | Sustained growth beats its sector (3 years) | 4.8% vs 17.6% (sector 70th pct, n=55) |
| ✓ | Profits grew last year | swung to a profit of $60M from $-6M |
| ✓ | Profit growth beats its peers | profitable now after losses three years ago |
| ✓ | Growth is speeding up, not slowing | 1y 46.3% vs 3y 4.8% |
| ✕ | Grew per share, not just in total | -11.4% vs 0.0% |
Quality
Whether the growth makes real money - margins, returns on capital, and whether profits turn into cash.
Profitability is thin or negative - the growth isn't turning into money yet.
Margins: of every $1 of sales, how much survives each cost layer?
◌ 2026 = trailing twelve months to the latest filed quarter (2026-06-30), not a full fiscal year
→Operating margin widened 17 points to 4% since 2022. After everything, 5 cents of each sales dollar reaches net profit. UPST doesn't tag a gross-profit line in its filings, so the chart starts at operating margin.
Earnings quality: do the reported profits turn into real cash?
◌ 2026 = trailing twelve months to the latest filed quarter (2026-06-30), not a full fiscal year
→Only -471% of reported profit becomes operating cash. Accounting profit is running ahead of cash collection, which is worth watching in the receivables and inventory lines.
What does it earn on the money it uses?
◌ 2026 = trailing twelve months to the latest filed quarter (2026-06-30), not a full fiscal year
→ROE of 7% on shareholders' capital (ROCE isn't meaningful for this business model).
How much of every sales dollar ends up as free cash?
◌ 2026 = trailing twelve months to the latest filed quarter, not a full fiscal year
→Free cash flow is negative, so no share of revenue is currently converting to spare cash. Every sales dollar is being reinvested or consumed.
What does staying competitive cost, per dollar of sales?
◌ 2026 = trailing twelve months to the latest filed quarter (2026-06-30), not a full fiscal year
→The biggest claim on each sales dollar is research and development, at 25% of revenue (stock compensation 13%, capital spending 0%).
What does “Quality” actually mean?
Quality asks whether the growth makes real money. Margins show how much of each sale survives costs; return on equity shows what shareholders earn on their capital; and the profit-to-cash comparison catches companies whose accounting profits never turn into actual cash.
How we scored it · 1 of 6 checks passed
| ✕ | Better net margins than peers | 4.8% vs 18.2% (sector 70th pct, n=77) |
| ✕ | Runs leaner than peers (operating margin) | 4.0% vs 25.6% (sector 70th pct, n=45) |
| ✓ | Actually profitable | TTM net income $60M |
| ✕ | Earns well on shareholders' money | 7.6% vs 16.1% (sector 70th pct, n=79) |
| ✕ | Earns a real return on its assets | 1.9% vs 5.0% |
| ✕ | Profits are cash, not accounting | -4.71 vs 0.80 |
Health
The balance sheet stress test: could UPST survive a bad year?
The balance sheet carries real risk - read the checks before anything else.
Could it handle its debt if things went wrong?
→Debt isn't clearly tagged in UPST's filings, so treat the balance sheet with extra care rather than assuming zero.
Is the company's own capital growing or shrinking?
→The company's own capital grew from $635M in 2023 to $797M (+26%). The business is building book value rather than consuming it.
What does “Health” actually mean?
Health asks one question: can the business survive a bad year? We check whether near-term bills are covered, whether debt is modest and shrinking, whether earnings comfortably pay the interest, and - for loss-makers - how many years of cash are left at the current burn rate.
How we scored it · 0 of 6 checks passed
| ✕ | Leverage is under control | 0.75 vs 0.60 |
| ✕ | Less levered than its peers | 0.75 vs 0.61 (sector 30th pct, n=106) |
| ✕ | Debt trending the right way | liabilities are 74.9% of assets vs 55.7% five years ago |
| ✕ | Earnings cover the interest | 1.29 vs 5.00 |
| ✕ | Converts sales to cash better than its sector | -22.8% vs 29.3% (sector 70th pct, n=77) |
| ✕ | Self-funding | 1.58 years of cash at current burn |
Shareholder returns
How much cash actually flows back to owners - dividends, buybacks, whether the share count truly falls, and whether what is handed back is affordable.
UPST returns nothing to owners yet, and the share count keeps rising - every dollar stays in the business.
How much goes back to shareholders - and how much leaks out as stock compensation?
→Stock compensation ($132M) flows out with nothing returned - the dilution is winning.
How fast is your ownership being diluted - or concentrated?
→20.2% more shares last year - your stake was diluted by that much.
UPST has issued or retired shares - did shareholders end up better off?
Both lines start at 100 in 2019, so the gap between them is what each share gained or lost. Share counts are split-adjusted.
→UPST issued +650% more shares from 2019 to 2025 and revenue per share still fell -13%. On this measure the new shares have not paid for themselves.
Why compare buybacks with stock compensation?
A company can trumpet billions in buybacks while quietly issuing nearly as much stock to employees. What matters to you is the net effect: is the share count actually falling? If not, the 'return' is mostly recycling.
How we scored it · 0 of 6 checks passed
| ✕ | Share count isn't climbing | shares up 29.9% over 3 years |
| ✕ | Buybacks outpace the stock issued to staff | no buybacks against $146M of stock compensation |
| ✕ | Hands cash back to owners | no dividends and no buybacks in the last twelve months |
| ✕ | Meaningful yield to owners (dividends and buybacks) | 0.00 returned, 0.0% of market value |
| ✕ | Buybacks are sustained, not one-off | pays no dividend and reports no buybacks |
| ✕ | Buybacks growing | pays no dividend and reports no buybacks |
Trend analysis
What the market is doing about all of the above. This is price behaviour, not a fact about the business - read it as the market's current opinion, scored on six checks like every other chapter. The market is the S&P 500, measured by the SPY ETF with dividends included, over 3 months (63 trading sessions) and 12 months (252).
The market is voting against it right now - a falling trend on most measures.
How is UPST's trend actually behaving right now?
Chart by TradingView
→UPST is in a downtrend. The price is below the band where recent trading settled and that band is still falling, so nothing in the picture has turned yet. The last two weeks have rolled over, though the price is still above where it stood a month ago, so the fall is recent rather than long-running.
How the trend above is worked out
The trend read comes from the daily Ichimoku picture, a standard trend indicator, translated out of its jargon. It builds a band from the midpoints of the last 9, 26 and 52 sessions' highs and lows - in effect, the range where recent trading has settled - and draws that band 26 sessions into the future. Price above the band is an uptrend, below it a downtrend, inside it no trend. Because the band is drawn forward, the support for the next few weeks is already fixed and a change of its direction can be seen coming. We also compare the last two weeks against the last month, and today's price against where it stood a month ago. None of this says anything about the business; it describes the price only, and it is not advice. Where a check says 'the S&P 500', the comparison is with the SPY ETF including dividends, on the same adjusted basis as the stock's own price: short means 3 months, 63 trading sessions; long means 12 months, 252.
How we scored it · 0 of 6 checks passed
| ✕ | Trading above its cloud | 0.00 vs 0.50 |
| ✕ | Long-term trend structure is healthy | 30.28 vs 34.05 |
| ✕ | Rising over 3 months | -5.7% vs 0.0% |
| ✕ | Beating the S&P 500 over 3 months | -5.7% vs 4.7% |
| ✕ | Beating the S&P 500 over 12 months | -57.4% vs 20.0% |
| ✕ | Not in a deep hole | -59.3% from its 52-week high |
Insider activity
What the people running the company do with their own shares - reported to the SEC within two days, classified so pay-plumbing doesn't masquerade as conviction.
Are the people running it buying or selling with their own money?
→Insiders bought $6M against $3M of sales - net buying with their own money is the single most bullish signal insiders can send.
Who did what, exactly?
| Date | Insider | Role | Type | Shares | Value |
|---|---|---|---|---|---|
| 2026-08-25 | Natalia Mirgorodskaya | See Remarks | SELL | 913 | $27,417 |
| 2026-08-20 | Sanjay Datta | President, Capital& Enterprise | SELL | 8,945 | $251,298 |
| 2026-08-20 | Scott Darling | Chief Legal Officer | SELL | 7,696 | $216,216 |
| 2026-08-20 | Natalia Mirgorodskaya | See Remarks | SELL | 586 | $16,461 |
| 2026-08-19 | Sanjay Datta | President, Capital& Enterprise | SELL | 10,000 | $314,633 |
| 2026-08-19 | Andrea Blankmeyer | Chief Financial Officer | SELL | 3,000 | $87,990 |
| 2026-08-17 | Sanjay Datta | President, Capital& Enterprise | SELL | 2,033 | $59,683 |
| 2026-08-17 | Scott Darling | Chief Legal Officer | SELL | 814 | $23,898 |
| 2026-08-17 | Andrea Blankmeyer | Chief Financial Officer | SELL | 7,175 | $210,646 |
| 2026-06-09 | Sanjay Datta | President, Capital& Enterprise | SELL | 15,000 | $456,116 |
Why do the transaction types matter so much?
Most insider filings are not trades: stock grants, option exercises and tax withholding are how executives get paid, and gifts are estate planning. The signal lives in open-market transactions - a buy means an insider chose to spend their own cash on the stock. Sells are murkier: many are pre-scheduled 10b5-1 plans set months in advance. That's why the chart counts only open-market activity, and the table labels every row.