UiPath
UiPath sells software robots that do repetitive computer work inside big companies, and is adding AI agents that decide what to do next.
The scores
Growth3/4 ↓
ARR up 12.4% and revenue 13.4%, but part was bought and the core grows about 4%; the Q3 guide implies 7.6%.
Business quality3/4 →
Rising retention and fast-growing large accounts show a real switching cost, but screen-operating AI models are a determined entrant.
Financial position4/4 →
$1.405B of cash and no debt; free cash flow funds buybacks and acquisitions without outside money.
Execution record3/4 ↑
Fiscal 2027 guidance raised twice and the first profitable year delivered, but the finance chief just changed and the steady record is short.
Per share capture3/4 ↑
Buybacks cut the count about 3% a year, but offsetting share pay took about two-thirds of trailing free cash flow.
Potential2/4 →
The robot market is a few billion dollars, largely held; AI upside needs adoption, disclosure and beating Microsoft and the model labs.
Risk profile3/4 →
No customer is 10% of revenue; the founder's 84% of the votes is one serious, survivable dependency.
Valuation3/4 ↑
About 24 times plan cash after share pay, net of cash, pays for the plan working, not for the AI products becoming large.
Evidence quality3/4 →
Financials are filed and the threat is sourced to the attackers, but the AI figures are management's and the prize count is four years old.
If it works 1.0× · if it fails you keep about 75%.
The Verdict
We rate UiPath neutral. It sells software that does repetitive office work on computers for large companies. The business is solid: customers keep adding to it, and it has plenty of cash and no debt. But its main business grows slowly, and AI models that can work a screen themselves are a real threat. The price is fair: about right if today's plan holds, with no extra credit for the new AI products. As a bet it is defensive: even if those products do well, the company is worth only a little more than today; if they stall, the old business and the cash still cover most of the price.
The long caseAs companies put AI to work, someone must connect it to old systems and check its work; UiPath must become that layer before Microsoft or model makers do.
The evidence under this - 6 claims
- Retention was 107% at January 31, 2026 in the full-year release, so 109% is a recovery rather than a high (S09).
- Attrition is concentrated among the smallest customers, by the company's own account (S04).
- The AI figures are rounded, so the remainder could be $25M to $40M.
- If WorkFusion's ARR was counted outside AI products, the core's own growth is lower still.
- Share pay may not stay at the first half's pace: new grants of 3.1M shares went to four officers on September 3, 2026 (S20).
- The vendor's company value is about $6.44B because it counts only part of the marketable securities; the filed balance sheet is used here.
On the filed balance sheet the company value is about $5.43B, which would put the doubling need near $1.8B of AI revenue, still about seven times today's book.
- Anthropic's own documentation describes the tool for developers to build with, not as a governed enterprise product that runs thousands of processes.
- A model that reasons each step is slower and less repeatable than a recorded robot, the chief executive's argument on the Q2 call (C21).
- Neither figure is filed: both were said aloud, as 'nearly' and 'more than', so the six-month gain could be anywhere from about $50M to more.
- WorkFusion, an AI agent business for bank compliance bought in February 2026, sits inside the second figure and not the first (C05).
- The total includes document reading, which UiPath sold for years before agents, so the agentic share is not disclosed.
The company
Big companies run on old systems that do not talk to each other, so staff copy data between screens all day. UiPath's software robots do that copying instead: a developer records the steps once and the robot repeats them, logging every click.
Its buyers are mostly banks, insurers, governments and manufacturers; 387 of them spend more than $1M a year. It now also sells AI agents that read documents and decide the next step.
The evidence under this - 4 claims
As older systems gain modern interfaces or are replaced, the need to drive their screens shrinks, which is the slow erosion inside the core's 4% growth.
- Retention was 107% at January 31, 2026 in the full-year release, so 109% is a recovery rather than a high (S09).
- Attrition is concentrated among the smallest customers, by the company's own account (S04).
A deal including AI says nothing about how much of its value AI is.
How it makes money
Customers pay a yearly licence or subscription, mostly for how many robots and how much of the platform they use, and most renew and add more. About $1.38B of signed contracts are still to be delivered.
Delivering software costs little, so most of each dollar is left over. The catch has been paying staff in shares: over the last year that cost took about two-thirds of the spare cash, though it is now falling fast.
The evidence under this - 4 claims
- Share-based pay is falling fast, $98.3M in the first half against $154.4M a year earlier, so the trailing figure overstates the run rate (C13).
- Cash taxes on vested shares, about $56M over the same four quarters, are paid in financing and sit outside both figures.
Where the money comes from, stream by stream
| Stream | Share of sales (Q2 fiscal 2027) | How it earns it |
|---|---|---|
| Subscription services | 64.9% | Cloud software and the yearly maintenance and support on licences, billed over the contract and recognised evenly; the line that has grown. |
| Licences | 30.2% | Term licences for software customers run themselves, booked when each term starts, so this line is lumpy and has been flat for three years. |
| Professional services and other | 5.0% | Paid help building and running automations, sold at little or no profit to get customers live. |
Who runs it
Daniel Dines founded UiPath in 2005 and runs it again after two years sharing the job and a brief step aside; a new finance chief, promoted from inside, took over in September. Dines owns about a fifth of the company and, through special shares, more than four-fifths of the votes, so outside shareholders cannot overrule him. He takes almost no pay. The other officers' bonuses ride on growing yearly subscriptions and on cash, not on profit after the cost of paying staff in shares.
The evidence under this - 4 claims
9.6M of the Class A shares belong to his former spouse and are counted because he keeps their votes, so his own economic stake is nearer 17.8%.
Adjusted free cash flow does not charge share pay, the cost that took about two-thirds of trailing cash (C09), and no measure pays on GAAP profit, per-share figures or relative share return.
Why it wins, and who could beat it
Why customers come back
A large customer may have hundreds or thousands of robots built in UiPath's tools, each tuned to its own screens and rules. Moving them to a rival means rebuilding and re-testing every one, so customers mostly add rather than leave: existing customers spend $1.09 for every dollar they spent a year ago.
That wall is getting lower. AI coding tools make rebuilding cheaper, and UiPath itself is using them to move a big Canadian bank's robots over from a rival. What helps UiPath win can help a rival win back.
The evidence under this - 4 claims
- A Canadian bank is moving its whole automation estate onto UiPath from a rival using coding agents, which shows the rebuild cost is falling in both directions (C35).
- Licences are mostly annual, so a customer can leave at each renewal (S01).
- Retention was 107% at January 31, 2026 in the full-year release, so 109% is a recovery rather than a high (S09).
- Attrition is concentrated among the smallest customers, by the company's own account (S04).
The customer and the rival it is leaving are unnamed, so the win cannot be checked.
As older systems gain modern interfaces or are replaced, the need to drive their screens shrinks, which is the slow erosion inside the core's 4% growth.
The threat
AI models can now work a computer screen themselves: Anthropic sells a tool that lets its model look at the screen and use the mouse and keyboard, the robot's job. Microsoft bundles its own robot software, and UiPath's annual report admits AI lets companies build what they used to buy. UiPath answers that a model is slower, can be wrong and still needs something to run and check it.
The evidence under this - 4 claims
- Anthropic's own documentation describes the tool for developers to build with, not as a governed enterprise product that runs thousands of processes.
- A model that reasons each step is slower and less repeatable than a recorded robot, the chief executive's argument on the Q2 call (C21).
- The analyst's question exists because the market believes the opposite, and the shares fell 16.6% the next day (C22).
- UiPath's own 10-K concedes AI agents let customers build what they used to buy (C16).
The rivals: MSFT, NOW, APPN, PEGA, SSNC, and 3 more that are private, in-house or listed abroad.
Who else does this, rival by rival
| Ticker | Rival | Where it competes | How far along | Priced against it |
|---|---|---|---|---|
| MSFT | Microsoft Power AutomatePriced against it: Shown, not counted | Core automation platform | On sale | Shown, not counted |
| Private | Anthropic's computer use tool and OpenAI's computer-using agentsPriced against it: Private | Core automation platform | On sale | Private |
| In-house | Customers' own developers using coding agentsPriced against it: In-house | Core automation platform | On sale | In-house |
| NOW | ServiceNow AI agentsPriced against it: One segment only | AI products (agentic, document reading and Maestro) | On sale | One segment only |
| APPN · PEGA | Appian and Pegasystems process platformsPriced against it: Yes | AI products (agentic, document reading and Maestro) | On sale | Yes |
| Private | Automation AnywherePriced against it: Private | Core automation platform | On sale | Private |
| SSNC | SS&C Technologies HoldingsPriced against it: Shown, not counted | - | On sale | Shown, not counted |
How big it could get
How big this could get
The robot market is not large. The last outside count, from the research firm Gartner, was $2.9B for 2022, and UiPath alone now has about $1.94B a year of recurring revenue, so it already holds much of it.
The bigger hope is AI. The Census Bureau finds 37% of US companies with 250 or more staff now use AI, but nobody counts what they spend on agents. Management's $142B estimate counts work AI might do, not software bought today.
The evidence under this - 4 claims
- The figure is four years old and Gartner now counts a wider category it calls business orchestration and automation technologies, with no free revenue figure.
- UiPath's own ARR at the time was well over a third of it, so the count may be narrower than what UiPath sells.
- Part of the growth was bought: WorkFusion joined on February 5, 2026 and Peak in 2025, and neither's revenue is broken out (C05).
- Licence revenue is booked up front, so quarterly revenue swings with renewal timing; ARR is the steadier figure.
Gartner's last outside count of the robot software market was $2.9B (C29); the $142B counts work that AI might do, not software bought today.
The results
The numbers
Sales grew 15% over the twelve months to Jul 2026, about the same pace as the 13% in the year to Jan 2026.
Operating margin is 8.9%, up from −13% in the year to Jan 2024.
Running costs take 74¢ of each sales dollar, down from 137¢ in the year to Jan 2022.
Each dollar put into the business earns 8.5% a year after tax, up from 2.9% in the year to Jan 2026.
It turned 21% of sales into free cash, $363M over the last twelve months.
Yearly sales growth has ranged from +9.3% to +47% over the last 5 years.
The share count rose 15% in 4.5 years, and pay in shares costs 14% of sales.
All the numbers
Filed figures from the company's 10-K and 10-Q, read 2026-10-04. Fiscal years end in January.
| Jan 2022 | Jan 2023 | Jan 2024 | Jan 2025 | Jan 2026 | 12 mo to Jul 2026 | |
|---|---|---|---|---|---|---|
| Size and growth | ||||||
| Sales | $892M | $1.06B | $1.31B | $1.43B | $1.61B | $1.72B |
| Sales growth | +47% | +19% | +24% | +9.3% | +13% | +15% |
| Profit | ||||||
| Gross margin | 81% | 83% | 85% | 83% | 83% | 83% |
| Operating margin | −56% | −33% | −13% | −11% | 3.5% | 8.9% |
| Net margin | −59% | −31% | −6.9% | −5.2% | 18% | 21% |
| What growth costs, per sales dollar | ||||||
| Research and development | 31% | 27% | 25% | 27% | 24% | 21% |
| Sales and marketing | 78% | 66% | 55% | 52% | 42% | 40% |
| General and admin | 28% | 23% | 18% | 16% | 13% | 12% |
| Pay in shares | 58% | 35% | 28% | 25% | 18% | 14% |
| Capital spending | 1.0% | 2.2% | 0.6% | 1.0% | 1.2% | 0.6% |
| Cash | ||||||
| Cash from operations | −$55.0M | −$9.98M | $299M | $321M | $371M | $373M |
| Free cash flow | −$63.8M | −$33.8M | $292M | $306M | $352M | $363M |
| Balance sheet | ||||||
| Cash and short-term investments | $1.87B | $1.76B | $1.88B | $1.63B | $1.47B | $1.28B |
| Returns and shares | ||||||
| Return on capital | - | - | - | - | 2.9% | 8.5% |
| Shares (diluted) | 455M | 548M | 564M | 560M | 545M | 523M |
What growth costs
Growth has been making UiPath more profitable. Sales have risen by more than half since 2023 while spending on selling stayed flat, and paying staff in shares has shrunk fast. The catch is that the easy gains are mostly taken: with the core growing slowly, more profit now has to come from holding costs down rather than from selling more.
The evidence under this - 2 claims
- Stock pay sits inside every cost line and fell faster than any of them, so cash costs per sales dollar fell less than reported costs did.
- Gross margin barely moved, 83.0% to 82.6%, and the 10-Q expects hosting costs to rise as more customers take the cloud version (S03).
With the core growing about 4% (C04), further gains must come from holding costs rather than from growth, which is a smaller lever.
The last report (3 Sep)
Quarter ended July 2026, filed figures.
In Q2 fiscal 2027, the three months to July, revenue rose 13.4% to $410.3M and recurring revenue 12.4% to $1.94B. UiPath made $31.6M of operating profit under the official accounts, against a loss a year earlier, and existing customers spent more.
The problem was the guide. Management expects about 7.6% growth in Q3, and on the call it said AI products had passed $250M a year without filing the figure. Part of that came from WorkFusion, a company it bought in February.
What it did to the thesisIt confirmed the business is profitable and customers are spending more, but the guide for the next quarter showed growth slowing, and the AI products' size is still only management's word, so the question of how big they are stayed open.
The quarter in detail
Sales growth from the previous quarter, filed figures.
Earnings per share $0.15 against $0.15 expected; $0.07 as filed, the gap being what the company adds back.
The evidence under this - 5 claims
- Part of the growth was bought: WorkFusion joined on February 5, 2026 and Peak in 2025, and neither's revenue is broken out (C05).
- Licence revenue is booked up front, so quarterly revenue swings with renewal timing; ARR is the steadier figure.
- Neither figure is filed: both were said aloud, as 'nearly' and 'more than', so the six-month gain could be anywhere from about $50M to more.
- WorkFusion, an AI agent business for bank compliance bought in February 2026, sits inside the second figure and not the first (C05).
- The total includes document reading, which UiPath sold for years before agents, so the agentic share is not disclosed.
- UiPath has beaten its revenue guide in each of the last four quarters, so the guide has been a floor (S20).
- Full-year guidance was raised in May and again in September (C45).
- Retention was 107% at January 31, 2026 in the full-year release, so 109% is a recovery rather than a high (S09).
- Attrition is concentrated among the smallest customers, by the company's own account (S04).
What management did with the money
It spent $268.5M buying back 22.8M of its own shares in the first half, at about $11.28 each, and $160M in cash on WorkFusion. Paying staff in shares fell to $98.3M from $154.4M a year earlier, so the share count is now shrinking.
Insiders in 2026Q3: 8 purchases and 15 sales, from their Form 4 filings.
The evidence under this - 3 claims
The price, and where we disagree
In short. Fair. Once its cash is set aside, the price pays for this year's plan working and steady growth after it, and pays nothing extra for the AI products succeeding.
The shares fell 16.6% the day after the results, on the slower guide and an analyst's question about a new AI model that does office workflows, and kept falling: 28% in a month.
The market's view is now gloomy. Only 3 of 21 analysts say buy, and a quarter of the shares that trade are sold short, bets that the price will fall. When that many people expect the worst, the next results can matter more than usual.
The evidence under this - 5 claims
- UiPath has beaten its revenue guide in each of the last four quarters, so the guide has been a floor (S20).
- Full-year guidance was raised in May and again in September (C45).
- The analyst's question exists because the market believes the opposite, and the shares fell 16.6% the next day (C22).
- UiPath's own 10-K concedes AI agents let customers build what they used to buy (C16).
How we read the multiple
Over the last year UiPath made about $362.9M of spare cash after running costs and investment. But paying staff in shares is a real cost:
- Take off $234.6M of share pay and about $128.3M is left.
- The $6.84B price is about 53 times that.
- But $1.4B of it is cash in the bank, leaving about $5.43B for the business.
- This year's plan, about $425M of spare cash less share pay near $200M, puts that near 24 times: fair for steady growth near 10%.
Against its own history. Cheaper than its own past on spare cash, because for years it barely made any; the business has changed more than the price has.
The evidence under this - 5 claims
- Share pay may not stay at the first half's pace: new grants of 3.1M shares went to four officers on September 3, 2026 (S20).
- The vendor's company value is about $6.44B because it counts only part of the marketable securities; the filed balance sheet is used here.
- Share-based pay is falling fast, $98.3M in the first half against $154.4M a year earlier, so the trailing figure overstates the run rate (C13).
- Cash taxes on vested shares, about $56M over the same four quarters, are paid in financing and sit outside both figures.
- Q3's guide implies growth of about 7.6% against 13.4% in Q2, which the market read as a slowdown (C08, S18).
- Pegasystems had flagged AI-driven delays in customer purchases, the same risk to this guide (S18).
Priced against its rivals
The competitors its own annual report names, each confirmed against the rival's filing. Measured on price for each $1 of spare cash, cheapest first:
| Ticker | Company | price for each $1 of spare cash |
|---|---|---|
| PEGA | Pegasystems | 11.1× |
| PATH | UiPath | 19.2× |
| APPN | Appian | 35.4× |
| SSNC | SS&C Technologies Holdings not in the median | 10.5× |
| MSFT | Microsoft not in the median | 57.4× |
The middle of the other 2 is 23.2. SS&C Technologies Holdings, Inc., Microsoft Corporation compete with it in one product line inside a much bigger company, so their price says little about this business and they are not counted in the middle. UiPath Inc. is 2nd cheapest of 3.
Rivals' figures come from a data provider, so read the spread of the group rather than any single pair.
For each dollar of spare cash, buyers pay more than for one rival and less than the other, so the price is the group's, not its own premium.
Private or foreign rivals that cannot be priced: Automation Anywhere, Inc., OpenAI, Anthropic.
The bet: what it pays if it works
The bet is on the AI products alone: agents, document reading and the software that coordinates them. The robot business is counted separately, at what its closest rival is worth for each dollar of sales.
Even if the AI products keep growing fast for two more years, the whole company ends up worth only a little more than today, so it is an expensive bet. But if they stall, the robot business and the cash still cover most of the price, which makes the downside small.
| AI products (agentic, document reading and Maestro) revenue in two to three years | AI product books of today's size needed | Valued like its peers | Like the cheapest |
|---|---|---|---|
| $450.00M | about 2 at 100% kept | 0.9× | 0.9× |
| $650.00M (the big win) | about 3 at 100% kept | 1.0× | 1.0× |
| $900.00M | about 4 at 100% kept | 1.2× | 1.1× |
To pay 2× (a Fair bet) it would need about $2.87B of AI products (agentic, document reading and Maestro) revenue (about 11 AI product books of today's size); to pay 5× (a Cheap bet), about $9.66B of AI products (agentic, document reading and Maestro) revenue (about 39 AI product books of today's size), at the same multiple.
Each row is AI products (agentic, document reading and Maestro) alone, plus the rest of the company (Core automation platform) counted at about $5.12B.
The steps it has to take
- Done AI product ARR passes $250M, by management's account
- Not started UiPath discloses AI product ARR again, split from document reading or WorkFusion due 2026-12-02
- Not started AI product ARR reaches about $360M due 2027-09
- Not started AI product ARR reaches about $650M due 2029-01
The detail behind the label
The evidence under this - 6 claims
- Neither figure is filed: both were said aloud, as 'nearly' and 'more than', so the six-month gain could be anywhere from about $50M to more.
- WorkFusion, an AI agent business for bank compliance bought in February 2026, sits inside the second figure and not the first (C05).
- The total includes document reading, which UiPath sold for years before agents, so the agentic share is not disclosed.
On the filed balance sheet the company value is about $5.43B, which would put the doubling need near $1.8B of AI revenue, still about seven times today's book.
- The AI figures are rounded, so the remainder could be $25M to $40M.
- If WorkFusion's ARR was counted outside AI products, the core's own growth is lower still.
Where we disagree with the market
Old robot software that AI models will hollow out: only 3 of 21 analysts rate it a buy.
The robots are sticky and cash per share is rising, so the price is fair, though not cheap enough to call the market wrong.
The market reads one quarter's guide. The filed numbers say existing customers are spending more, large customers grew 21% in a year, and share pay is falling fast.
The case against us is the threat itself: if AI models do the robots' job well enough, customers will stop renewing, and spending per customer, not the guide, will show it first.
What analysts say, and the evidence
The evidence under this - 6 claims
- Retention was 107% at January 31, 2026 in the full-year release, so 109% is a recovery rather than a high (S09).
- Attrition is concentrated among the smallest customers, by the company's own account (S04).
- Anthropic's own documentation describes the tool for developers to build with, not as a governed enterprise product that runs thousands of processes.
- A model that reasons each step is slower and less repeatable than a recorded robot, the chief executive's argument on the Q2 call (C21).
- UiPath has beaten its revenue guide in each of the last four quarters, so the guide has been a floor (S20).
- Full-year guidance was raised in May and again in September (C45).
What could go wrong, and how we would know
The risks
- AI models do the job. Models that work a screen themselves, from Anthropic and others, could replace the robots, and Microsoft bundles its own robot software with products customers already buy.
- The core is slowing. The old robot business grows about 4% a year, and the guide for Q3 fiscal 2027 implies total growth of about 7.6%.
- AI figures are unchecked. The $250M of AI products is management's spoken figure, not filed, and part of it was bought with WorkFusion.
- One man holds the votes. The founder holds about 84% of the votes, so other shareholders cannot force a sale or a change of course.
- Share pay could rise again. Paying staff in shares took $290.7M in fiscal 2026; new grants to four executives in September could push it back up.
The evidence under this - 8 claims
- Anthropic's own documentation describes the tool for developers to build with, not as a governed enterprise product that runs thousands of processes.
- A model that reasons each step is slower and less repeatable than a recorded robot, the chief executive's argument on the Q2 call (C21).
- The AI figures are rounded, so the remainder could be $25M to $40M.
- If WorkFusion's ARR was counted outside AI products, the core's own growth is lower still.
- UiPath has beaten its revenue guide in each of the last four quarters, so the guide has been a floor (S20).
- Full-year guidance was raised in May and again in September (C45).
- Neither figure is filed: both were said aloud, as 'nearly' and 'more than', so the six-month gain could be anywhere from about $50M to more.
- WorkFusion, an AI agent business for bank compliance bought in February 2026, sits inside the second figure and not the first (C05).
- The total includes document reading, which UiPath sold for years before agents, so the agentic share is not disclosed.
What we still do not know
- How much of the AI products is new agents rather than older document reading, and how much came with WorkFusion.
- How many customers have already moved screen work to an AI model's agent instead of a robot.
- How much spending customers keep each year before they add more; UiPath files only the figure after additions.
The evidence under this - 4 claims
- Neither figure is filed: both were said aloud, as 'nearly' and 'more than', so the six-month gain could be anywhere from about $50M to more.
- WorkFusion, an AI agent business for bank compliance bought in February 2026, sits inside the second figure and not the first (C05).
- The total includes document reading, which UiPath sold for years before agents, so the agentic share is not disclosed.
- Management says it will disclose AI ARR periodically, so the figure may become reportable (S05).
- 18 of the top 20 Q2 deals included AI, so it is in most large contracts (C39).
- Anthropic's own documentation describes the tool for developers to build with, not as a governed enterprise product that runs thousands of processes.
- A model that reasons each step is slower and less repeatable than a recorded robot, the chief executive's argument on the Q2 call (C21).
- Retention was 107% at January 31, 2026 in the full-year release, so 109% is a recovery rather than a high (S09).
- Attrition is concentrated among the smallest customers, by the company's own account (S04).
What would change our view
- Neutral→AttractiveThe price falls on an unchanged business
It needs all of:
- ○The market value falls to 40 times the cash left after paying staff in shares, from about 53 today.about $9.85 today, 31% below the last close
- ○Existing customers keep spending at least $1.08 for every dollar they spent a year earlier.checked at the next results, due by 2 Dec
- Neutral→AttractiveThe AI products prove large
It needs all of:
- ○UiPath reports AI products of $400M a year or more.checked at the next results, due by 2 Dec
- ○Existing customers spend $1.12 or more for every dollar they spent a year earlier.checked at the next results, due by 2 Dec
- Neutral→UnattractiveThe core starts to shrink
It needs any of:
- ○Existing customers spend $1.05 or less for every dollar they spent a year earlier.checked at the next results, due by 2 Dec
- ○Recurring revenue grows less than 8% in a year.checked at the next results, due by 2 Dec
We would also be wrong if
- The AI products grow less than 30% in a year.
- Paying staff in shares rises back above $250M a year.
What is coming
- 2 December 2026 Q3 fiscal 2027 results: Whether revenue lands in the $440M to $445M guide and ARR reaches the $1.992B test with retention at 109% or better; a miss says the core is slowing further.
Every dated event ahead
- 2 December 2026 Q3 fiscal 2027 results: Whether revenue lands in the $440M to $445M guide and ARR reaches the $1.992B test with retention at 109% or better; a miss says the core is slowing further.
- March 2027 Q4 fiscal 2027 results and fiscal 2028 guidance: Whether ARR reaches the $2.065B guide, whether AI ARR gets a filed figure, and whether the fiscal 2028 guide implies core growth above the 4% estimate.
- May 2027 2027 proxy statement: Whether pay moves beyond ARR, revenue and adjusted free cash flow to a measure that charges share pay, and whether the chief executive's pledge grows or shrinks.
Since we wrote this
- 29 OctBet moveBet re-struck: Safe, little upside to Priced for successRe-priced at the 28 October close.
- 28 OctBet moveBet re-struck: Priced for success to Safe, little upsideRe-priced at the 27 October close.
- 27 OctBet moveBet re-struck: Safe, little upside to Priced for successRe-priced at the 26 October close.
- 5 OctThesis reviewThe bet now names what it ridesThe owner asked that every bet name the macro trend behind it, after BWXT's bet was missing its chief executive's case for nuclear over gas.
- 4 OctThesis writtenNeutral, medium convictionThe robot core grows slowly but customers keep adding to it, the AI products are real but small, and after cash the price is fair rather than cheap.
- 4 OctThesis reviewOwner's checklist: revenue streams, cost of growth, management, catalystsThe owner's checklist found the method silent on who runs the company, what growth costs, where the money comes from and what is dated ahead; this pass fills those four blocks from the proxy and filings.
Sources 24
Sources - 24
Every claim above cites one of these, or states the reasoning that derived it. The quote is frozen at the date it was read: a link can rot, and an agent fetch returns a reading of a page rather than the page. Backs counts the whole claim ledger, so it can name a claim this page does not surface.
- S01primarycompany filing2026-03-25Item 1 platform description (Maestro, Agent Builder, robots, IDP, Test), the Competition section's seven classes including AI model research organisations and coding agents, the 'build software internally' passage, no customer at 10% of revenue, Class B control at about 84% of votes, stock compensation by year, 3,981 employees, full-year cash flowbacksC01C13C14C15C16C47C49C54C55
- S02primarycompany filing2026-09-03Q2 and first-half fiscal 2027 income statement, balance sheet, cash flow and non-GAAP reconciliations; ARR $1.938B; net new ARR $37M; retention 109%; Q3 and full-year fiscal 2027 guidancebacksC02C07C10C13C45C46
- S03primarycompany filing2026-09-08ARR table with customers above $1M and $100K and retention; WorkFusion and Peak acquisition notes; remaining performance obligations of $1,377.6M; buybacks of 22.8M shares; 521.2M shares outstanding on the coverbacksC02C05C06C11C28C47C48C49C54
- S04primarytranscript2026-09-03Q2 fiscal 2027 earnings call transcript (FMP)data/cache/transcripts/PATH-2027Q2.jsonthe Canadian migration using coding agents; about 10,350 customers with attrition among the smallest; 18 of the top 20 deals included AI; the Barclays question on a frontier lab's workflow model and the chief executive's answer; fiscal 2027 adjusted free cash flow of about $425M; no debt
one of Canada's largest financial services companies, working with Ashling Partners to migrate its entire automation footprint to UiPath and plans to use coding agents to power that migration
backsC01C07C10C21C35C39 - S05primarytranscript2026-05-28Q1 fiscal 2027 earnings call transcript (FMP)data/cache/transcripts/PATH-2027Q1.jsonthe free cash flow guide; 20M shares bought in Q1 at $11.47; a competitive displacement at an electronics manufacturer; management saying AI product ARR will be disclosed periodically
we continue to expect fiscal year 2027 non-GAAP adjusted free cash flow of approximately $425 million
read, but no claim rests on it - S06primarytranscript2026-03-11Q4 fiscal 2026 earnings call transcript (FMP)data/cache/transcripts/PATH-2026Q4.jsonAI product ARR of nearly $200M at January 31, 2026; 90% of $1M customers and about 60% of $100K customers using AI products; AI buyers spending nearly 3 times as much
AI product ARR, which includes agentic, IDP and Maestro, reached nearly $200 million this quarter
backsC03C27 - S07primarytranscript2026-09-22AI ARR above $250M; retention of 114% for $100K customers and 120% for those using both robots and AI; the $142B addressable market; dilution held below zero; Maestro built on the open-source Temporal engine; top 25 customers' average ARR up from $2M to $10M since 2020
At the end of Q2, our ARR from our AI ARR was more than $250 million.
backsC03C26C36C37 - S08primarycompany filing2026-09-03Ashim Gupta keeps only the chief operating officer role; Hitesh Ramani promoted to chief financial officer; a new directorbacksC40C53
- S09primarycompany filing2026-03-11first full year of GAAP operating profit ($57M); fiscal 2027 guidance first set at $1.754B to $1.759B of revenue; a new $500M buyback; the WorkFusion acquisition announcedbacksC05C45C46
- S10primarycompany filing2026-02-20Appian sells process automation with robotic process automation, document processing and AI inside one platform, and names RPA providers as a competitor classbacksC20
- S11primarycompany filing2026-02-18Pega combines attended and unattended robotic automation with workflow and case management for the same large banks, insurers and governmentsbacksC20
- S12primarycompany filing2026-07-30Power Automate is sold inside Dynamics 365 with no revenue of its ownbacksC17
- S13primarycompany filing2026-01-29organizations can choose to rapidly deploy thousands of out-of-the-box ServiceNow AI agentsbacksC19
- S14primaryother2026-10-04a general AI model sold to developers that operates a desktop through screenshots, mouse and keyboard, the same interface UiPath's robots use
Claude can interact with computer environments through the computer use tool, which provides screenshot capabilities and mouse/keyboard control for autonomous desktop interaction.
backsC18 - S15secondaryindustry data2022-08-01Gartner's last public RPA software revenue figure: $2.9B for 2022, growing 19.5%, and a projected 17.5% in 2023
Global robotic process automation (RPA) software revenue is projected to reach $2.9 billion in 2022, an increase of 19.5% from 2021
backsC29 - S16secondaryindustry data2019-06-24Gartner's finding that buyers use robots chiefly to connect older systems that lack modern interfaces
The ability to integrate legacy systems is the key driver for RPA projects
backsC30 - S17primarygovernment data2026-05-28AI use at 19.8% of US businesses on May 3, 2026, within 17% to 20% since December 2025; 37% of firms with 250 or more employees; Finance and Insurance 33.9%
37% of firms with at least 250 employees reported using AI in their business operations
backsC31 - S18secondarynews2026-09-04the market's reading of the Q3 guide as about 7.6% growth, and that Pegasystems had flagged AI-driven purchase delays
though PEGA earlier flagged the same AI-driven customer purchase delays now hitting UiPath
read, but no claim rests on it - S19secondarynews2026-09-23the long-term operating margin target above 30%; new-logo selling prices doubled; hundreds of customers running Maestro in production
It also reiterated long-term operating-margin targets above 30% and expects continued GAAP profitability.
backsC26C36 - S20secondaryvendor data2026-10-04PATH research pack (FMP, FINRA, SEC XBRL), frozencontent/PATH/research-pack.jsonprice and return series, the price-move split, analyst ratings and target drift, short interest, insider statistics, headcount, and the filed-competitor multiplesbacksC22C23C24C43
- S21primarycompany filing2026-10-0477 open-market sales in the year to October 2026, every one flagged under a Rule 10b5-1 plan, and no purchasesbacksC25
- S22primarycompany filing2026-02-27Blue Prism, bought in 2022, is SS&C's robotic process automation product, with no separate revenueread, but no claim rests on it
- S23primarycompany filing2026-10-04UiPath XBRL company facts (SEC), frozen as content/PATH/financials.jsoncontent/PATH/financials.jsonfive fiscal years and the four quarters to July 2026 of revenue, R&D, sales and marketing, admin, stock pay, capital spending, operating income and the diluted share count, as the filer tagged thembacksC49C54
- S24primarycompany filing2026-05-12beneficial ownership at April 28, 2026 including the pledge and the former spouse's shares, the chief executive's tenure and nominal pay, the fiscal 2026 bonus metrics and payout, the PSU revenue goals and the iARR gatebacksC51C52C53