Workday's Subscription Growth Has Slipped to 14% While the Machines It Sells Eat Its Own Seats

---

Every great software franchise eventually meets the law of large numbers, and the meeting is rarely dramatic. There is no single bad quarter, no fraud, no collapse — only a growth rate that drifts down a point or two a year until one day the company that was priced as a compounder is revealed to be a mature utility wearing a growth multiple. Workday is somewhere on that path, and the precise location matters more than usual, because the company is simultaneously selling the market a story about acceleration. The story is artificial intelligence: agentic systems that automate the back-office tasks Workday has always hosted, a new revenue surface, a reason to believe the deceleration reverses. The forensic question is whether that story is an accelerant or a solvent — whether the AI Workday is racing to sell expands the business or quietly dissolves the seat-based model the business is built on. The numbers in the first quarter of fiscal 2027 do not resolve the question, but they frame it sharply, and the frame is not flattering.

Start with the figure the bulls and the bears both have to live with. Subscription revenue — the recurring core, roughly 93% of the total — grew 14.3% year over year in the April quarter, to $2.354 billion. Total revenue grew 13.5% to $2.542 billion. And the guide for the full fiscal year puts subscription revenue at $9.93 to $9.95 billion, which is growth of 12 to 13%. For a company whose investor identity was built on twenty-plus-percent subscription growth sustained across many years, this is the most important number on the page, and it is not a rounding error away from the old trajectory. It is a different trajectory. Everything else in this piece is an attempt to understand what that new trajectory is worth, and whether the AI narrative bolted on top of it can do the work the growth rate no longer does on its own.

The metronome slowed, and the guide says it slows again

The single most consequential fact about Workday in 2026 is that its subscription growth has a one in front of the four. For most of its public life the company compounded subscriptions in the low-to-mid twenties; the deceleration to 14.3% in the first quarter of fiscal 2027 is not a one-quarter wobble but the visible endpoint of a multi-year glide path, and management's own guidance confirms the direction by pointing lower still. Full-year subscription revenue of $9.93–9.95 billion implies 12–13% growth, and the second-quarter subscription guide of roughly $2.46 billion implies about 13%. When a company guides its own forward growth rate below its just-reported rate, it is telling you, in the politest possible language, that the deceleration is not finished.

This is the denominator problem in its purest form, and it deserves to be named plainly rather than waved away. A company adding revenue onto a base approaching ten billion dollars cannot post the percentage growth it managed off a two-billion-dollar base, no matter how good the product or how loyal the customer. The absolute dollars Workday adds each year may still be large; the percentage they represent shrinks mechanically as the base swells. None of this is a scandal. It is arithmetic, and it happens to every successful subscription business eventually. The scandal, if there is one, lives not in the deceleration itself but in what the market is asked to pay for a business decelerating through the mid-teens toward something lower — and in the narrative deployed to distract from the arithmetic.

The backlog tells the same story with a slight lag. The twelve-month subscription revenue backlog — the contracted revenue due to be recognized within a year, and the cleanest forward indicator software offers — was $8.81 billion, up 15.5%. That is a hair above the reported subscription growth rate, which is mildly encouraging; it suggests the next twelve months are not about to fall off a cliff. But the total subscription backlog, stretching out across all contract durations, grew only 10.9%, to $27.29 billion. The longer-dated book is growing more slowly than the near-dated book, which is the signature of a business whose growth is being pulled forward by shorter commitments rather than extended by longer ones. The metronome is not broken. It is simply ticking at a mid-teens tempo and, on the company's own numbers, slowing toward the low teens.

A third of revenue is profit — until you count the stock

Here is where the quality-of-earnings work begins, and it is the part of the Workday story that the adjusted headlines are engineered to obscure. In the first quarter of fiscal 2027, Workday reported non-GAAP operating income of $809 million — a luxurious 31.8% non-GAAP operating margin. That is the number in the press release headline, the number management lifted its full-year guidance around (to roughly 30.5% adjusted operating margin), the number that supports the narrative of a high-margin compounder. On a non-GAAP basis, nearly a third of every dollar of revenue drops to operating profit.

Now look at what GAAP says about the same quarter. GAAP operating income was $338 million — a 13.3% operating margin. GAAP net income was $222 million, or $0.87 per share. Set the two earnings figures side by side: non-GAAP EPS of $2.66 versus GAAP EPS of $0.87. More than two-thirds of Workday's reported profitability disappears when you stop excluding the things the company would prefer you exclude, and the largest single thing being excluded is stock-based compensation. In fiscal 2026, Workday's stock-based compensation ran at roughly 16% of revenue — close to one dollar in six of everything it sold, paid not in cash but in shares carved out of existing owners. That is not a non-cash nicety to be waved through. It is real compensation, it is a real cost of running the business, and the dilution it creates is a real transfer of value away from shareholders and toward employees.

The defense of excluding stock comp is that it is non-cash, and that is true as far as it goes. But the cash isn't conjured from nothing; it is conserved precisely by paying people in equity instead, and the equity is replenished by issuing more shares or by spending actual cash on buybacks to mop up the dilution — buybacks that are then themselves excluded from the adjusted picture as if they were optional. The honest way to read a software company that books a 31.8% adjusted operating margin and a 13.3% GAAP operating margin is that roughly half of its advertised profitability is a function of how it chooses to pay its workforce. Workday is genuinely profitable on a GAAP basis — 13.3% operating margin and positive net income are not nothing, and many software peers cannot say as much. But the gulf between the 31.8% the company leads with and the 13.3% the accounting confirms is the single most important thing an investor underwriting Workday should hold in mind, because the valuation is built on the larger number while the economics are governed by the smaller one.

The seat model versus the machine it is selling

Now the structural tension, which is sharper at Workday than almost anywhere in software. Workday sells back-office systems of record — human capital management and financial management — and it is paid, at its foundation, on a per-employee, per-seat, per-module basis. The value of a Workday contract scales with the size of the customer's workforce: more employees to administer, more payroll to run, more headcount logged into the system. The entire commercial logic assumes a large population of humans doing back-office work that Workday hosts and bills for.

Agentic AI — the technology Workday is now building its identity around — is designed to do that back-office work without the humans. If an AI agent processes the expense report, reconciles the ledger entry, handles the routine HR request, and answers the benefits question, the task still runs through Workday's platform, but the seat that used to do it may no longer exist. This is the innovator's dilemma stated in its most literal form: Workday must build the automation that threatens its own per-seat economics, because if it does not, a competitor or an AI-native upstart will, and yet every seat that automation eliminates is, under the legacy pricing model, a seat Workday no longer bills. The company's answer — and it is the only credible one — is to monetize AI on a consumption or per-agent basis, charging for the work the agents perform rather than the humans they replace, so that automation becomes a new meter rather than a subtraction from the old one. But that is a pricing transition the market is assuming will land cleanly, and the history of business models asked to cannibalize themselves gracefully is not reassuring. The party whose product makes its own billing unit obsolete is not in an enviable spot, however good the product.

The company is firing the proof

The most striking evidence that this tension is not abstract is that Workday is demonstrating it on its own payroll. Across 2025 and into 2026 the company executed significant headcount reductions explicitly framed around AI-driven efficiency — a roughly 8.5% cut of about 1,750 employees, followed by a further reduction of around 2% of the global workforce in February 2026, on the order of 375 positions. The concentration of those cuts is the tell: they fell heavily inside Global Customer Operations — customer success, adoption services, support, implementation enablement — the human-intensive functions that AI is supposed to automate first.

Read charitably, this is a company eating its own cooking, using its AI to run leaner and proving the value proposition it sells to customers. Read forensically, it is a live demonstration of exactly the dynamic that threatens the revenue line. If Workday can replace its own back-office and customer-operations humans with agents, so can every customer that pays Workday by the seat. The company is, in effect, advertising to its own buyers that the headcount they license Workday to administer can be reduced — which is wonderful for Workday's margins and potentially corrosive for Workday's seat count. There is a reason the layoffs and the AI story arrive together: they are the same story told from two ends. The margin expansion the bulls celebrate and the seat erosion the bears fear are produced by the identical mechanism, and a company cannot indefinitely claim the first while denying the second.

A founder returns, which is rarely a sign of strength

In February 2026, the same month as the latest round of cuts, Workday replaced its chief executive. Carl Eschenbach stepped down as CEO and from the board, and co-founder Aneel Bhusri — who had been co-CEO, then CEO, then co-CEO again, then executive chair across the company's history — returned to the chief-executive chair. Founder restorations are sold as a return to vision and DNA, and sometimes they are exactly that. But the base rate is sobering: companies generally summon the founder back when the professional management hired to scale the business has run into something the board does not believe that management can navigate, and the something is usually a growth problem, a strategy problem, or both.

The timing here is not subtle. A founder returns to lead "the next chapter" precisely as subscription growth decelerates into the mid-teens, as the company executes AI-justified layoffs, and as the market begins to question whether the seat-based model survives the agentic era. The most charitable reading — that Bhusri is the right person to steer the AI pivot — may well be correct; he built the company and knows it intimately. But a leadership change executed alongside cost cuts and a decelerating top line is not the posture of a business confident that the existing plan is working. It is the posture of a board reaching for a reset. Investors paying a re-acceleration price should weigh the possibility that the people closest to the numbers reached for the founder because the existing trajectory was not the one the price assumes.

Priced for a re-acceleration that the guidance denies

The valuation is where the threads tie together, and it cuts both ways more than a simple short thesis would like. On a forward earnings basis, Workday is not expensive in the way it once was: a forward P/E in the low double digits — roughly 11 to 12 by mid-June 2026 — looks almost cheap for a profitable software franchise, and the bull leans hard on that number. But forward P/E here is calculated on non-GAAP earnings, the $2.66-style figure that excludes the 16%-of-revenue stock comp. Measured on GAAP earnings — the $0.87 quarter — the trailing P/E balloons toward the high forties, and the picture inverts: a company growing subscriptions 12–13% trading at a premium GAAP multiple is not cheap at all. The gap between "cheap on forward non-GAAP" and "expensive on trailing GAAP" is the same gap as the margin chasm, viewed from the valuation side, and which number is real depends entirely on whether you believe stock-based compensation is a cost.

The asymmetry is what matters. To justify even the modest forward multiple on the optimistic earnings basis, Workday has to do one of two things: re-accelerate subscription growth back toward the high teens, or convert the AI story into enough genuine, GAAP-real margin expansion to offset a top line settling in the low teens. The company's own guidance — 12–13% subscription growth for the full year, decelerating from the just-reported 14.3% — actively denies the first. That leaves the second: the entire bull case rests on AI both reversing the growth deceleration and doing so without eroding the seat base it depends on, while expanding margins on a GAAP and not merely an adjusted basis. That is a great deal to ask of a single technology, and the price embeds the favorable answer to every part of it. When a stock is priced for a re-acceleration that management's own guidance points away from, the burden of proof sits squarely on the optimist, and the downside if the proof does not arrive is a re-rating to the multiple of a low-teens grower.

The competition is aiming at the same back office

Workday's moat is its position as the system of record for the enterprise's people and money — sticky, mission-critical, brutally expensive to rip out. That stickiness is real and it is the best thing the company has. But the back office is precisely where the largest software companies and a wave of AI-native challengers are now aiming. SAP and Oracle have spent years and fortunes contesting the financials and HCM ground Workday occupies; Microsoft's reach into the enterprise gives it a platform from which to encroach; and a generation of AI-first startups is attempting to rebuild back-office workflows with agents at the center rather than bolted onto a legacy seat model. The agentic era that Workday is counted on to win is, for the first time in years, an era that genuinely reopens the question of who owns the back-office layer.

The threat is not that Workday loses its installed base quickly; switching a Fortune 500 company off its HCM and financials system is a multi-year ordeal that almost no one undertakes lightly. The threat is slower and subtler: that AI lowers the cost of building and migrating back-office tools over time, that broader-platform incumbents with their own AI to bundle nibble at the edges, and that the pricing power Workday has enjoyed as the entrenched system of record erodes at the margin as credible alternatives multiply and as the per-seat unit itself comes under pressure from automation. A 14% growth rate and a swelling backlog do not show this erosion, because entrenchment is a lagging indicator and competitive pressure surfaces slowly in renewal pricing before it ever shows in a growth print. But the valuation is a claim on a decade of pricing power, not a quarter of it, and the back office has not been this contested in a long time.

What the bulls genuinely get right

In fairness, the bull case is substantial, and the skeptic who ignores it is not being honest. Workday is a genuinely excellent, deeply entrenched franchise: its HCM and financials platforms are systems of record for an enormous roster of the world's largest enterprises, and the switching costs are about as high as software offers. That entrenchment gives the revenue exceptional durability — gross retention is strong, customers expand over time, and a back-office system simply does not get ripped out the way a point solution does. The 14% subscription growth, while down from its peak, is still healthy in absolute terms for a company approaching ten billion in revenue, and the twelve-month backlog growing 15.5% suggests the near-term pipeline is intact rather than collapsing. Crucially, Workday is genuinely GAAP-profitable — a 13.3% GAAP operating margin and positive net income put it ahead of many software peers that are still losing money outright, and the non-GAAP free cash flow is real, substantial, and growing. The margin expansion is happening: management raised its full-year adjusted operating margin guidance, and the AI-driven efficiency, whatever it implies for seats, is producing genuine cost leverage.

And the AI opportunity is not merely a defensive story. Workday sits on exactly the back-office data and process surface — payroll, financials, workforce records — where agentic automation should be most valuable, and if any back-office vendor can convert AI into net-new consumption revenue rather than seat cannibalization, Workday's position is among the strongest. The valuation, on a forward non-GAAP basis, has already compressed substantially from the company's growth-stock heyday, meaning much of the prior excess has been worked off; the stock is no longer priced for the twenties-growth fantasy. A founder back at the helm may be exactly the catalyst the AI pivot needs. For an investor who believes Workday monetizes agentic AI as a new meter while the core compounds and the GAAP margin keeps climbing, today's multiple is defensible and perhaps even modest. The honest synthesis is that this is a high-quality, durable business whose growth has matured and whose central narrative — AI as accelerant — must overcome the same technology's tendency to be a solvent. The bull is right about the quality, the stickiness, the cash, and the falling multiple. The skeptic notes that the growth is guided lower, that two-thirds of the headline profit is an add-back, that the seat model and the agent are at war, and that the price still embeds the optimistic resolution of all of it.

The reconciliation that does the heavy lifting

It is worth lingering on the reconciliation between non-GAAP and GAAP, because that single bridge carries more of the bull thesis than any other line in the filings. The journey from $338 million of GAAP operating income to $809 million of non-GAAP operating income is mostly a journey through stock-based compensation, and the size of that bridge — nearly half a billion dollars of add-backs in a single quarter — is the size of the assumption an investor makes when they accept the adjusted framing. If you believe stock comp is a genuine non-cash convenience, the 31.8% margin is the truth and the stock is cheap. If you believe it is real compensation funded by real dilution, the 13.3% margin is the truth and the stock is dear. There is no neutral position; the reconciliation forces a choice.

What makes the choice consequential is that stock comp at 16% of revenue is not a stable, shrinking line that will quietly fade as the company matures. It is a structural feature of how high-end software talent is paid, and it tends to persist or grow even as companies age, because the labor market for engineers and salespeople does not care about a company's GAAP optics. A business that has decelerated to mid-teens growth has less room to outgrow its dilution than one compounding in the twenties — the equity issued each year is being spread across a top line that is expanding more slowly, which means the per-share cost of that compensation is rising in relative terms. The market is being asked to underwrite a margin story whose most flattering version depends on permanently excluding one of the company's largest and most persistent costs, at exactly the point in the company's life when that cost is hardest to grow away from. The reconciliation is not a footnote. It is the thesis.

The kicker

Workday is not a fraud and it is not a falling knife; it is something more interesting and more common — a great business arriving, on schedule, at the awkward middle of its life. The growth that defined it has slipped to the mid-teens and is guided lower; the profitability it advertises is a third of revenue on the adjusted page and a thirteenth on the GAAP one; the AI it is racing to sell is built to eliminate the very seats it bills for, and the company is proving the point by firing its own people; and the founder has been summoned back to write the next chapter precisely because the last one stopped accelerating. None of these facts is hidden. All of them are in the filings and the press releases, stated in the calm language of a company that would prefer you read the non-GAAP line and the AI slide and skip the reconciliation. The price still asks you to believe in a re-acceleration that the guidance denies, funded by a technology that is as plausibly a solvent as an accelerant. The bet is not that Workday breaks. The bet is on which number is real.

The whole of the Workday question lives in the distance between two figures from a single quarter — the $809 million of adjusted operating profit the company leads with and the $338 million of GAAP operating profit the accounting confirms — and in the strange fact that the artificial intelligence sold to close that gap is the same automation that quietly empties the seats the company is paid to fill, so the investor underwriting today's price is wagering, simultaneously, that the smaller profit is the illusion and that the machine eating the back office is a friend, when the more sober reading of the same documents is that the larger profit is the illusion and the machine is doing exactly what it was built to do.