Workday (WDAY): the founder returns amid an AI panic

2026-08-12 · By Lubin Danilo, founder of Lubin Investment

WDAY: see the full analysis on Lubin Investment

Workday lost nearly half its stock market value in a year, caught up in fears that artificial intelligence could make its payroll and HR software obsolete. Its founder came back to run the company himself to answer that fear. The business still looks solid in my numbers, but the real bet is playing out now, not in the past.

The software nobody talks about, and nobody ever leaves

Workday runs the software that calculates payroll, manages time off, and tracks every hire across a large share of the world's biggest companies. Nobody mentions it at dinner, everyone touches it at the office. This kind of software has a precise name in corporate jargon: a system of record, the place where the data lives that no other application can afford to copy incorrectly, or a paycheck goes out late and a tax filing gets missed.

That is exactly what makes this kind of software so hard to leave. Migrating the payroll, org charts, time-off rules, and career history of twenty thousand employees to a new system is not a project you start on a whim: it is a multi-year undertaking, run with internal teams and specialized integration firms, with a real risk of breaking something badly if a single payroll rule gets copied wrong. A company that has finished installing Workday, or one of its longtime rivals like SAP SuccessFactors, Oracle, or UKG, rarely stays out of affection: it stays because the cost and risk of starting over far outweigh the discomfort of staying put, even as the license price climbs at every contract renewal.

That inertia shows up plainly in the numbers Workday reported for its fiscal year ended January 31, 2026: revenue of $9.552 billion, up 13.1%, and, more tellingly, a subscription revenue backlog of already-signed but not-yet-recognized contracts of $28.101 billion, up 12.2% year over year, according to Workday's official full-year results release. That order book, what software investors call backlog, is not an optimistic forecast written by a marketing team: these are contracts already signed, spanning several years, whose cash will land in the books over coming fiscal years almost regardless of what the stock does in between. With 21,000 employees and revenue per employee of roughly $455,000, Workday sells software, not billable hours: few businesses in my universe offer this much visibility into their own future while staying this light on physical capital.

The founder who came back to save what he built

On February 4, 2026, Workday cut about 400 roles, roughly 2% of its workforce, mostly in customer support and functions that do not directly generate revenue. Five days later, on February 9, the announcement that truly shook the stock landed: Carl Eschenbach, who had led the company for barely thirteen months, stepped down as CEO and left the board, under a departure agreement that grants him several million dollars in severance and accelerated vesting of his shares, while staying on as a strategic advisor to his successor. Taking his place was none other than Aneel Bhusri, Workday's co-founder, who had himself held the CEO role until 2024 before handing it to Eschenbach, according to Workday's official transition announcement.

Bhusri justified his return in unambiguous terms: artificial intelligence, in his own framing, is a bigger transformation than the shift to the cloud that made Workday's fortune in the first place. His read on the problem shows up directly in the headcount numbers: the roles cut in February hit legacy customer-facing functions first, while the company redirects hiring toward engineering and products tied to artificial intelligence. This is not cost-cutting in the usual sense, the kind done to polish a quarterly margin ahead of an earnings call: it is a company-wide reallocation, a bet that Workday's future lies in a direction quite different from what has driven its revenue so far.

This is where a simple reading falls short, in either direction. A founder coming back after two years away can be read two opposite ways, and both are true at once. The positive one: nobody understands the company's inner workings better than the person who built it more than twenty years ago, and his return signals a level of personal commitment a hired executive structurally lacks, especially one who has remained a major shareholder since founding and has no reason to play it safe in the short run. The less comfortable one: thirteen months is an extremely short window to decide an executive has failed, and the board just implicitly admitted it got its previous pick wrong, even though the market had welcomed Eschenbach's arrival as a sign of operational discipline after years of somewhat unruly growth. I always judge a company's quality separately from its price, as I detail in my guide on management, but management stability is part of that first question, and this kind of turmoil at the top is a real thing to watch, not a footnote that fades by next quarter.

What the market actually punished

The market did not wait to render its verdict: the stock plunged more than 8% the very day of the February 9 announcement, hitting a 52-week low of $144.55. That plunge, more than any disappointing quarter, explains most of the value lost since the peak, at a moment when the entire enterprise software sector was already going through a wave of worry over what artificial intelligence means for its future revenue. The stock has since clawed back some ground: as I write this, on August 12, 2026, it trades at $175.29, still well below its highs from before this broad-based panic. This is not a Workday-only accident: the same wave of worry over what room artificial intelligence will leave for subscription-based enterprise software knocked down several other big names in the sector at the same time, a sign the market is pricing in a sector-wide risk almost as much as a story specific to this one company.

At this price, Workday trades at 31.8 times its free cash flow, meaning the cash that genuinely stays in the till once every bill is paid: at that multiple, it would take nearly 32 years of today's cash, compounded, to earn back the price paid for the stock today. I could not pin down precisely where this level sits in the stock's own history at the time of writing, lacking access to the full series, and I would rather say so than invent a percentile that would give a false sense of precision. What my model does calculate reliably is a reasonable buy price of $253.54 for this name, projecting that cash five years out with conservative assumptions: the current price therefore trades at a 44.6% discount to that threshold. That is a rare signal on a name scoring 9 out of 10 in my quality screen, where nine criteria out of ten clear the bar, the only real flagged weakness being modest share dilution and a return on invested capital so high, 811%, that it mostly confirms how little physical capital this business consumes relative to the cash it produces.

That dilution is worth pausing on, because it tells a story that free cash flow, taken alone, completely hides. A large share of the cash Workday generates comes from a specific accounting mechanism: stock-based compensation paid to employees is a real cost to existing shareholders, but it does not leave the company's cash account, so it gets added back into the free cash flow calculation instead of being subtracted from it. Over the last fiscal year, that stock-based compensation amounted to roughly 53% of reported free cash flow, a mechanism I break down in a dedicated article. This is not fraud, nor something unique to Workday, most software companies work this way, but it is precisely why Workday's share count keeps creeping up by about 1.8% a year despite genuine buybacks: part of the cash funding those buybacks is really just offsetting the dilution it just created, not shrinking the net share count.

The real bet: can AI agents replace a seat billed by the month

That leaves the real question, one that thirteen months in the corner office and 400 layoffs only accelerated without answering: can artificial intelligence agents break Workday's own business model? Almost all enterprise software, Workday included, is billed by the seat today: a fixed price per employee, per month, regardless of how much work actually gets processed behind it. If an AI agent can now handle a time-off request, an expense report, or a payroll close from start to finish without a human ever opening the software, does the client company still need to pay for that seat at the same rate as before?

That is exactly the bet Bhusri put on the table with Illuminate, the agent platform Workday has been building since 2024 and significantly expanded in 2025 to cover HR, finance, and industry-specific work, according to Workday's official announcement. As a sign that even Workday does not believe it can sell these agents at the usual flat rate, the company introduced Flex Credits at the same event, a new consumption-based billing unit built specifically for how customers actually use its agents, rather than a simple add-on tacked onto the existing seat price. The idea is not to let a third party sell agents that would siphon Workday's usage from the outside, but to sell those agents itself, built on the enterprise data it already hosts and on the technical architecture shared across all its customers. Josh Bersin, an HR technology analyst who has tracked Workday for years, sums up the shift underway as a move from a simple system of record to what he calls a platform of agents, in his detailed analysis of this reinvention: in his view, already owning a company's most sensitive organizational data, data no outside agent can access without explicit authorization, remains Workday's best defensive asset against this risk, far sturdier than any simple technology lead that could be closed within eighteen months.

The opposite risk is real, though, and it would be dishonest not to name it. Nothing stops a younger competitor, one without twenty years of legacy code to maintain, from building an agent that reads Workday's data through a simple open interface and sells that service directly to Workday's own customers, with their consent, without ever paying Workday a dime in licensing along the way. Bhusri's bet only works if Workday stays faster at building its own agents than these rivals are at working around its system, and only if its customers agree to pay for a layer of intelligence on top of the data layer they already pay for. None of that is guaranteed in advance, and nobody, not even Bhusri, can promise it with any certainty today.

Here is how I come down on this, without rushing the call. The quality of the business remains solid on nearly every criterion that matters to me, with free cash flow still growing at a healthy pace and debt that is nearly negligible against the cash generated each year. The price already bakes in a good chunk of February's panic: my own model flags this name as sitting in buy territory, which does not happen that often on a stock scoring this high in my quality screen. But the real test is neither in February's press release nor in this article: it lands on August 27, 2026, when Workday reports its next quarterly results. That day, I will look at one number before anything else: the growth rate of subscription revenue and of the 12-month backlog. If it holds up, the thesis of a founder returning at the right moment, on the right issue, holds together. If it slows noticeably, February's stock market panic will have been right before I was, regardless of how solid the rest of the balance sheet looks. I would rather wait for that confirmation in the numbers before adding to a position than bet on the strength of a press release alone. If you want to follow this name as it develops, my Workday page runs on the same live numbers quoted here, and I lay out the general logic behind my reasonable buy price and quality screen in my full methodology.

FAQ

What is a system of record in a company?

It's the software where a company's official data lives (payroll, org charts, time off), the one every other tool checks against. That role is exactly why Workday is so hard to replace once it's installed.

Why did Aneel Bhusri come back to run Workday?

After Carl Eschenbach stepped down in February 2026 (thirteen months into the job), co-founder Aneel Bhusri took the CEO role back, arguing that artificial intelligence is a bigger transformation than the cloud and demands a fast change of course.

Why does Workday's share count keep rising despite buybacks?

Because a large share of free cash flow comes from stock-based compensation added back into the calculation (about 53% last year), which dilutes shareholders alongside the buybacks. The buybacks mostly offset that dilution rather than shrinking the net share count.

Is Workday currently in my buy zone?

According to my model, yes: the current price of $175.29 trades at a 44.6% discount to my reasonable buy price of $253.54. But I'm waiting for the August 27, 2026 results to confirm subscription growth holds up before adding to a position.

About the author

Written by Lubin Danilo, founder of Lubin Investment. A self-taught individual investor, I find fundamental analysis fascinating, and it has delivered excellent results. For three years now, my performance has beaten the S&P 500. But analyzing every stock took too much time: sites with incomplete data, calculation methods and criteria never aligned with mine. And spotting the best stocks was just as time-consuming, even with my own well-defined checklist. So I put my software development background to work to build this software, base my investment strategy on its results, and share it with people who share the same passion as me. It judges a company's quality and its price separately, using criteria drawn from the financial literature (Warren Buffett, Michael Mauboussin, Aswath Damodaran).