The One Number Behind UiPath's 24% Selloff — and Whether the Stock Is Actually Cheap

UiPath did the two things that usually push a beaten-down stock higher: it beat Wall Street's revenue estimate for its fiscal second quarter and it raised its full-year outlook. And the stock responded by falling roughly 24% in the days after the September 3 report. When a company beats and gets sold, the market is usually telling you it didn't like the quarter the way you think it did. The useful question isn't "why did it drop." It's which single number drove the drop, because that number is also the one that decides whether the shares are cheap.

That number is net new ARR — the amount of recurring revenue UiPath added in the quarter on top of what it already had. It came in at just $37 million, down from $49 million in the prior quarter. Growth of the existing revenue base is what a subscription software company lives on, and UiPath's is decelerating to a crawl.

The growth meter reads "stalling"

Total annual recurring revenue reached $1.938 billion, up 12% year over year. Total revenue was $410 million, up 13%. Those are decent numbers for most companies but the wrong numbers for a software growth story, because they tell you the engine is no longer accelerating.

The market's real concern was visible in the fine print. Net new ARR of $37 million was 24.5% below the $49 million UiPath added in its fiscal first quarter. That is the lowest step on a staircase that investors were told would keep climbing as the company monetizes AI agents. When the incremental revenue a company adds each quarter shrinks, the whole growth goose is cooked regardless of what the reported revenue line says — and the market read that correctly on September 3. Even the raised guidance contributed to the mood: full-year revenue was nudged up to a midpoint around $1.79 billion, but the fiscal third-quarter midpoint actually came in just below consensus.

None of this is a distressed balance sheet. UiPath ended the quarter with roughly $1.4 billion in cash, no debt, and $31.6 million of GAAP operating income — its fourth consecutive quarter of profitability, with a non-GAAP operating margin near 22%. It even bought back 2.4 million shares. The company is now consistently profitable while its installed base holds: dollar-based net retention was 109%, meaning existing customers spent a bit more, not less.

So this is the odd profile the selloff has left behind: a cash-rich, profitable, recently-buyback-happy software company whose customers aren't leaving — but whose new-business engine is sputtering. Quality improved. Growth is what's broken.

The bear case is factual, not panic

The honest contrarian knows the difference between a selloff that outruns the numbers and one that tracks them. This one mostly tracks them. The deceleration in net new ARR is reported, reproducible evidence, not narrative. AInvest's aggregate analyst signal labels the stock Hold — not a screaming buy from the sell side.

The loudest bear argument is that general-purpose AI agents from the likes of ServiceNow and Salesforce will do to robotic process automation what it does to code-searching — render the incumbent's core product redundant. Management's answer is that AI doesn't replace UiPath so much as it creates demand for the thing UiPath sells: orchestration and governance over agents, robots, and people executing end-to-end processes. Founder and CEO Daniel Dines put it directly, arguing that AI increases the need for "orchestration, governance, and exactness that deterministic automation provides."

That is a real argument and not obviously a losing one. But the distinction the market cares about — and the one that separates "competition exists" from "moat breached" — is showing up in the numbers. Existing customers are renewing and even expanding (109% retention), which says the moat hasn't cracked. The weakness is on the new-customer side: net new ARR rose only modestly year over year, at a time when the whole market is trying to figure out how to price AI agents.

Is it undervalued, or just cheaper?

This is where the "undervalued stock to buy" framing gets tested. UiPath trades at roughly $13.75, down far off its 52-week high near $20, with a market value around $7.2 billion. On trailing sales the stock is about 4x revenue, and against forward sales around 3.3x. That no longer carries a premium for a hyper-growth name — because it isn't one anymore. A 12% to 13% grower at roughly those multiples is not a market-misreading bargain; it's a fairly-priced company whose growth premium has mostly been wrung out.

Calling UiPath undervalued today requires believing the deceleration in net new ARR reverses — that agentic-AI monetization actually begins landing in bigger quarterly increments, and that the new-rulebook transaction-based pricing of AI agents converts into durable recurring revenue the way license-and-subscription revenue did. Nothing in the September 3 report proves that yet. A single quarter in which net new ARR inflects upward wouldn't prove the comeback plan works, but it would start to move the burden of proof.

That's the honest close. The market wasn't obviously wrong to sell UiPath, and a reader should not treat a 24% drop plus a beat as an automatic invitation to catch a falling knife. What the selloff has produced is a profitable, clean-balance-sheet company whose premium is gone — which is interesting, and which becomes a genuine contrarian entry only when the number that caused the drop, net new ARR, starts climbing again. Wait for that meter to turn before assuming the market has it backwards.