UiPath's Turnaround Is Real. The Market Hasn't Noticed.

First GAAP profit. $1.2 billion in net cash. AI in 16 of 20 top deals. And a stock price that still trades like a melting ice cube.

The best time to buy a turnaround is when the numbers have already turned and the market still believes the old story. UiPath just crossed that line.

The company reported its first GAAP profitable quarter on May 28. Revenue grew 17% year over year to $418 million. AI modules landed in 16 of the top 20 deals, and AI inclusive expansions were six times larger than deals without. The balance sheet holds $1.2 billion in net cash against just $83 million of debt. A $500 million buyback is actively retiring shares at an average cost of $11.47.

And yet the stock trades at $12.19, down 82% from its 2021 peak, rated Hold by all 16 covering analysts, with a consensus price target of $13.25. That gap between the numbers and the narrative is the opportunity.

The financial inflection

The quarter ending April 30, 2026, UiPath's Q1 FY27, was not just another print. It was the one that made the turnaround thesis numerically undeniable.

GAAP operating income hit $28 million, compared with a $16 million loss a year earlier. Non-GAAP operating margin reached 22%, up 250 basis points year over year. Free cash flow was $130 million in the quarter alone. Annual recurring revenue crossed $1.9 billion, growing 11%, and full year guidance calls for $2.06 billion.

These are not the numbers of a company in decline. They are the numbers of a company that got lean, found its footing, and is now converting revenue into cash at a 31% free cash flow margin. At the current $6.3 billion market cap, the implied free cash flow yield on trailing numbers is roughly 5.7%, and the buyback suggests management agrees the stock is cheap.

But the market does not price stocks on what they just did. It prices them on what it thinks they will do next. And what the market thinks about UiPath is still stuck in 2022.

AI is a tailwind, not a headwind

The bear case on UiPath has always been that robotic process automation is a dead category. Why would anyone buy RPA when AI can do the whole job?

The May earnings call answered that question directly. CEO Daniel Dines described a world where AI builds the automations and UiPath runs them. "AI creates automation," he said. "You run those automations. It is very cheap to run, deterministic, reliable, auditable. And only when these scripts break, you invoke AI again to fix the scripts."

In this framing, UiPath is not competing with AI. It is the runtime environment for AI built automations. The company is model agnostic, plugging into whatever large language model the customer prefers. The orchestration layer, governance, audit trails, credential vaults, and self-healing capabilities are the moat. As one customer put it at the company's DevCon event, "Models are easy. Orchestration is not."

The evidence that this thesis is working in the field is concrete. A healthcare distribution company deployed an end to end agent and automation workflow that generated multimillion dollar annual savings and led to a seven figure expansion deal. A Fortune 500 energy company placed UiPath at the center of a $70 million cost reduction initiative. A global telecom with 2,000 running automations is now building 200 deterministic plus 20 agentic use cases. A global semiconductor company is replacing a legacy RPA vendor with UiPath as its strategic AI automation platform.

The coding agent product, launched at DevCon, is accelerating development cycles in ways that drive stickiness. One consumer electronics company went from a four week build to three hours. A chip manufacturer cut a two month process to a few days. These are not marginal improvements. They are step changes that make the platform harder to rip out.

The OpenAI scare created the entry point

When OpenAI launched Presence in late July, an enterprise platform for deploying AI agents with voice, chat, and policy controls, PATH dropped 15% in a single session. The logic was straightforward: if OpenAI can deploy agents directly, who needs UiPath?

The logic is also wrong, for the same reason the bear case has been wrong since 2023. OpenAI is a model company. UiPath is an orchestration company. The hard part of enterprise AI is not the model. It is making the model work inside a governed environment with compliance, access controls, and audit trails that satisfy a Fortune 500 chief information security officer. UiPath already has those credentials. OpenAI does not.

The Bargo SaaS Disruption Monitor, which runs 32 software companies through daily guardrail tests to detect whether AI exposure predicts fundamental deterioration, puts the fear in context. Its July 27 verdict: the AI eats SaaS fingerprint is "not detectable." The raw correlation between AI exposure and fundamental pressure is 0.25, but it collapses to negative 0.03 after controlling for company age, size, and growth rate. A permutation test puts the result at p equals 0.16, indistinguishable from chance. The fear that is compressing software multiples across the sector has no empirical footprint yet.

If AI is not eating SaaS, then the names that sell AI into the enterprise, like UiPath, are mispriced by the fear itself.

Smart money is positioning for a rerating

The long dated options market tells the story that the sell side will not. LEAPS open interest, contracts expiring six months to two years out, shows 83,900 calls versus 29,715 puts, a ratio of 2.8 to 1. The largest single position is the January 2028 $15 calls with 9,137 contracts. There are 8,385 contracts at the $25 strike and 7,995 at $20. In the December 2028 expiry, 871 days out, someone opened a large new position of nearly 5,000 contracts at the $22 strike in the most recent session.

The options market is not just betting on a bounce. It is betting on a structural rerating over the next 18 months, with conviction concentrated in the $15 to $30 range. At the same time, the gamma structure is supportive. Net dealer gamma is positive at $2.5 million, meaning options dealers are in a long gamma regime that suppresses volatility. The gamma flip sits at $10.45, well below the current spot of $12.19, and the call wall at $13 provides a near term magnet if accumulation continues.

Insider activity is clean. Over the last 180 days, there have been zero open market sales. Every Form 4 filing has been tax withholding or annual director equity grants. The CFO, CPO, CLO, and CAO all held their positions through the Q1 print. The buyback is real: 22 million shares have been retired at an average of $11.47, with capacity remaining under the $500 million authorization.

What could go wrong

The risks are not theoretical. They are specific and measurable.

Small customer churn remains a drag. Dollar based gross retention is 97%, meaning 3% of revenue walks out the door annually, concentrated in smaller clients. Growth depends on enterprise expansion, not logo growth. If enterprise AI adoption spending slows, the growth story stalls.

The September 3 earnings print is binary. The estimated EPS is $0.04, but the real numbers to watch are ARR growth, GAAP profitability continuation, and AI deal metrics. If Q2 reverses any of the Q1 improvements, the turnaround narrative resets. Given the seasonality in UiPath's business, a softer Q2 is possible even if the underlying trend is intact.

OpenAI may not be the only competitive threat for long. Microsoft Power Automate and Copilot have distribution advantages no independent platform can match. ServiceNow is building agentic workflows into its IT service management dominance. If one of these players bundles orchestration with an existing enterprise platform at zero incremental cost, UiPath's value proposition gets harder.

The seat pricing model is under structural pressure across the software industry. The Bargo disruption data shows that seat priced names carry more fundamental stress than consumption priced peers. If enterprises shift from per user licensing to consumption models, UiPath's pricing architecture and gross margins could compress.

Finally, the stock has been a serial disappointer. Down 82% over five years, it carries the kind of reputational damage that takes more than one good quarter to repair. Sentiment will not flip overnight.

The numbers in one table

A company with a pristine balance sheet, accelerating AI adoption, a 17% growth rate, and a 0.42 PEG ratio trading at 13 times forward earnings in a sector where the median forward multiple was once above 30 is not a normal situation. It is a market that has given up on the story and stopped checking the numbers.

The numbers have changed. The market will catch up. The question is whether it happens before or after September 3.

More research at bargo.ai/research.