Guidewire (GWRE): Shares Fall 20% as FY2027 Growth Guide Lands at 16%

Guidewire Software reported fourth-quarter and full-year fiscal 2026 results after the US close on Thursday 3 September 2026, beat on both revenue and earnings, and then fell about 20% on Friday, closing at $162.42 against a prior close of $202.86. The quarter itself was fine. What broke the stock was a first-quarter revenue guide of $372m to $378m against a consensus of roughly $387m, and a full-year fiscal 2027 revenue range of $1,707m to $1,727m.

That full-year range is the figure to hold on to. Against fiscal 2026 total revenue of $1,475.4m, the midpoint of $1,717m implies growth of roughly 16%, down from the 23% Guidewire just delivered. A seven-point deceleration in guided revenue growth is what a 20% de-rating looks like. But the more interesting number sits underneath it, and almost nobody printed it.

What actually happened

Take the quarter first, because it is quickly dealt with. Fourth-quarter total revenue was $411.1m, up 15% year on year and ahead of a consensus around $402.5m. Non-GAAP net income was $0.99 per share on 83.6m diluted weighted-average shares, against roughly $0.94 expected — and it is worth being explicit that this is the non-GAAP figure, since Guidewire's GAAP and non-GAAP earnings differ substantially on share-based compensation.

The full year was better still. Total revenue of $1,475.4m was up 23% on fiscal 2025. Annual recurring revenue as at 31 July 2026 was $1,242m, measured on exchange rates as at 31 July 2025, representing 19% constant-currency growth. Fully ramped ARR grew 22%. The company also repurchased 4,085,350 shares during the year at an average price of $148.41, for an aggregate $606.3m.

Now the guidance. For the first quarter of fiscal 2027 Guidewire guided total revenue of $372m to $378m, subscription and support revenue of $279m to $283m, and ending ARR of $1,253m to $1,259m. For the full year it guided total revenue of $1,707m to $1,727m and ending ARR of $1,450m to $1,460m, which management framed as roughly 18% constant-currency ARR growth at the midpoint.

And then the sentence that did the damage. The fiscal 2027 outlook assumes ARR attrition normalises relative to fiscal 2026 — because fiscal 2026's attrition was a record low. Gross ARR attrition came in below 1.5% across all ARR and below 1% for core systems customers, who represent the vast majority of the base. Management stated that this record-low attrition contributed roughly one percentage point to fiscal 2026 ARR growth.

Why the market reacted the way it did

On the surface, the reaction is straightforward: a 3% miss on the near-term revenue guide, and a headline growth rate stepping down from 23% to 16%. Software trading on a high multiple of revenue is priced for the growth rate, and the growth rate moved.

But the sell-off was not really about the 3%. It was about that attrition sentence, and about what it implies for the reliability of everything the market thought it knew.

Guidewire is valued as a durability story. Property and casualty insurers run core policy, billing and claims systems for decades; they change them roughly never; and the whole investment case rests on the proposition that once a carrier is on the platform, the recurring revenue is close to permanent. A company that discloses gross ARR attrition below 1% on core systems is confirming exactly that. So when management then says the next year assumes attrition normalises higher, the market hears something uncomfortable: the durability everyone was paying for is a variable that management estimates, not a constant they can rely on. Worse, it is unfalsifiable in the near term. You cannot verify a churn assumption until the churn either happens or does not.

That is a legitimate thing to worry about. It is also, on the numbers, a strange thing to sell 20% on.

The Openbook read

Here is the arithmetic that reframes the whole event. Fiscal 2026 ARR growth was 19% on a constant-currency basis. Management said roughly one percentage point of that came from record-low attrition that will not repeat. Normalise it, and the underlying fiscal 2026 ARR growth rate was about 18%. Fiscal 2027 is guided to roughly 18% constant-currency ARR growth at the midpoint.

On a like-for-like basis, Guidewire has guided its recurring revenue growth to be flat, not lower. The apparent deceleration from 19% to 18% is entirely the removal of a one-off benefit management themselves identified and disclosed. The underlying engine is running at the same speed it ran last year. That is not the profile of a business that just got worse. It is the profile of a business whose disclosure got more conservative.

So why does guided revenue growth still fall from 23% to 16%? Because revenue and ARR have been converging, and in fiscal 2026 revenue was running ahead. Revenue grew 23% while ARR grew 19% — revenue catching up to a recurring base built in prior years, as subscription contracts ramped into full recognition. Fiscal 2027 guides revenue growth of about 16% against ARR growth of about 18%, which is the first year revenue growth crosses below ARR growth. Arithmetically, that crossover was always coming; a revenue line cannot outrun its own recurring base indefinitely. What the market read as demand deterioration looks a great deal more like the end of a catch-up.

With that established, the five factors fall out as follows.

Momentum is unambiguously damaged and there is no arguing with it. A 20% single-session fall on a stock that had been trading above $200 is a genuine trend break, and momentum scores that read price will read it correctly. This factor should not be rationalised away.

Growth is where the mechanical reading and the considered reading diverge most sharply. A factor model consuming guided revenue growth marks Guidewire down seven points. A factor model consuming underlying, attrition-adjusted ARR growth marks it down by roughly nothing. Our view is that the second is the better description of the business and the first is the better description of the near-term reported numbers, and both belong in the score — but the gap between them is the opportunity or the trap, depending on whether the attrition assumption proves conservative.

Profitability improved and continues to. Non-GAAP earnings of $0.99 for the quarter beat expectations, and the multi-year gross margin expansion from migrating customers onto Guidewire Cloud is the structural story that has been running underneath the growth debate all along. Nothing in Thursday's release disturbed it.

Solvency is not a pressure point. The company generated enough cash to fund $606.3m of repurchases in a single fiscal year without strain.

Reward/Risk has genuinely improved, and this is the crux. At $162.42 on roughly 83.6m diluted shares, the market capitalisation is about $13.6bn, or roughly 7.9 times the midpoint of guided fiscal 2027 revenue — arithmetic from published figures rather than a reported multiple, and it ignores any net cash position. The session before, the same guided revenue was being capitalised at close to 10 times. Nothing about the recurring revenue engine changed between those two prices. What changed is that the company told the market its churn had been unusually good and would probably not stay that way.

There is one more detail worth sitting with. Guidewire spent $606.3m buying its own stock during fiscal 2026 at an average of $148.41. Friday's close of $162.42 leaves that programme in the money by about 9% — and leaves the company facing a share price not far above the level at which it was already a willing buyer. A management team that believes its own attrition guidance is conservative has an obvious way to demonstrate it.

The read-across

The first read-across is the one to resist. This is the fourth "beat the quarter, fall on the guide" print in barely a week, after Marvell, Dell and Ciena, and the temptation is to file it as another data point in an AI-infrastructure capex wobble. It is not. Guidewire sells core administration software to property and casualty insurers. Its guidance issue is an assumption about customer churn in a market with essentially no connection to AI data-centre spending. Reading it as a fifth AI capex signal would be a genuine analytical error.

The correct read-across is narrower and more useful: any subscription business whose recent growth was flattered by abnormally low churn faces the same disclosure problem. Record-low attrition is not a permanent state, and a company that has enjoyed one has borrowed growth from the future without necessarily saying so. The screen worth running — and one our screener supports — is for recurring-revenue businesses whose net retention has been running above its own multi-year average, since those are the companies most exposed to a Guidewire-style reset in the next guidance cycle.

Within insurance technology specifically, the direct comparables are Sapiens, Duck Creek — now in private hands — and, further along the value chain, the claims and underwriting analytics names. Verisk and CCC Intelligent Solutions sell into the same carriers but on different contract structures, and neither is exposed to Guidewire's specific core-systems replacement cycle. The test worth applying is simple: if Friday's move were really about insurer IT budgets, it should show up across that group rather than in one name. A single-stock move on a single-company disclosure is the more economical explanation, and it is the one the guidance language supports.

What to watch next

• First-quarter fiscal 2027 results, due in early December. Guidewire reported the equivalent quarter on 3 December 2025, so expect a similar slot. The single number that matters is ending ARR against the guided $1,253m to $1,259m — and, more than the number, whether management repeats or revises the attrition normalisation assumption.
• Any disclosure of actual gross ARR attrition during fiscal 2027. If it stays near the sub-1.5% level rather than normalising higher, the fiscal 2027 ARR guide is conservative by roughly a full percentage point of growth, and the case that the market mispriced Friday strengthens materially.
• Guidance revisions through the year. A company that guides conservatively in September and raises in December is telling a very different story from one that holds the range all year.
• Buyback activity. Whether the repurchase pace picks up at prices below the $148.41 fiscal 2026 average is the clearest available signal of what management actually thinks the shares are worth.

Factor scores and full financial history for Guidewire are available on Openbook, and will refresh with the December print. The question this article leaves the reader with is the one the numbers pose directly: is this a business that got worse, or a disclosure that got more careful? The attrition arithmetic points fairly firmly at the second.