A First €10 Billion Quarter Beats on Revenue and Profit, but a 50% Rally Has Already Paid for It: Downgrading to Hold, Fair Value $231–259

Results vs. Consensus

Q3 2026 Scorecard

Consensus below is the company-compiled panel from October 19, on the same non-IFRS euro basis as the actuals. Media-reported consensus put EPS slightly above the panel median, at €1.83, and total revenue slightly below it, at €10.10B; the beats hold on either basis. No company-compiled backlog consensus exists, so that row uses the media-reported expectation.

Year-over-Year Comparison

Quarter-over-Quarter Comparison

Nine Months 2026

Quality of Beat

• Revenue: The cloud beat is real but flattered by two items. Management said one-off effects, mainly a multi-quarter revenue catch-up from customers returning from financial distress, added about a point to cloud growth, and the three 2026 acquisitions contributed €49M of revenue, close to another point against the prior-year cloud base if all of it is treated as cloud. By our estimate, organic cloud growth was therefore nearer 23% than 25% at constant currency. Total revenue growth of 12% cc is a point above the 2025 rate and is a clean beat; software support declined 6% cc, a smaller decline than the 7% in Q2, and licenses came in well above a small expectation.
• Margins: This is where the headline overstates the quarter. Q3 2025 carried about €0.1B from a change in case law on other tax litigation and about €0.1B of workforce transformation expense, both inside non-IFRS operating profit. Adding them back puts the prior-year base near €2,766M and a 30.5% margin. Against that base, Q3 2026 operating profit grew about 10% (11% excluding €39M of acquisition dilution), and the margin of 29.5% (30.0% excluding acquisitions) was 0.5–1.0pp lower. Cloud gross margin fell for a third straight quarter, research and development rose 13% on 2% headcount growth, and share-based payment expense rose 10% after falling 21% in the first half as the share price recovered.
• EPS: Non-IFRS EPS grew 19%, a point ahead of operating profit, helped by a non-IFRS effective tax rate of 25.8% against 27.9% a year earlier and a 1.2% lower weighted share count, partly offset by weaker net financial income. With a full-year tax guide of about 29% and a nine-month rate of 28.6%, the fourth quarter carries a higher rate, so the tax help does not repeat.
• Cash: Free cash flow grew 36% to €1,718M on 28% operating cash flow growth. The nine-month figure of €7,968M is up 11% after absorbing the €408M Teradata settlement and €783M more in income tax payments than a year earlier. Receivables were a drag of €142M against a €679M inflow in the prior-year nine months, which is the working capital point the CFO raised himself.

Segment Performance

Revenue Lines

Operating Segments and Regions

Current Cloud Backlog: Held at 26%

Current cloud backlog reached €24,144M, up 28% reported and 26% at constant currency, the same constant-currency rate as Q2 and a point below the 27% of Q3 2025. Sequentially it grew 5.3%. The rate beat expectations of roughly 24.7%, and one questioner observed that it looks slightly faster on an organic basis than in Q2, a point management did not dispute.

"While we continue to expect a slight deceleration in constant currencies current cloud backlog growth in 2026, we do foresee continued market share gains." — Dominik Asam, CFO

Assessment: Two consecutive quarters at 26% against a guide for deceleration settle the question that broke the stock in January: demand did not roll over. The fourth quarter now decides the exit rate, and management's own upper bound of about 24% would still be comfortably above the "definitely not a 4% decline" floor set in January. We keep our exit estimate at 24–25%.

Cloud ERP Suite: Faster, With Help

Cloud ERP Suite grew 28% at constant currency to €5,924M, a point faster than Q2, and is now 89% of cloud revenue. Some of the acceleration is the financial-distress catch-up, which management placed inside cloud revenue, and some is acquisition revenue. The CEO's framing was that AI is now the reason customers move.

"In Q3, AI was the main driver behind our customers move to the cloud. There are 3 key reasons. First, AI accelerates the RISE with SAP cloud migration. Second, AI makes the business case for our suite more attractive. And finally, our customers embrace our Business AI platform with RISE with SAP." — Christian Klein, CEO

Assessment: The franchise is taking share at scale, and external data agree: management cited IDC estimating SAP grows about 11 percentage points faster than its market. At 89% of cloud revenue, Cloud ERP Suite is the cloud line, so a sustained step down in its rate would show up immediately in the headline, with no diversification to absorb it.

Extension Suite: The Quiet Deceleration

Extension Suite, the non-ERP cloud lines including SuccessFactors and Concur, grew 8% at constant currency, down from 12% in each of the prior two quarters, and was slightly lower in absolute terms than in Q2. Management did not address the slowdown in prepared remarks; the only related exchange concerned new competitors in travel and expense.

Assessment: At €688M a quarter this line cannot move the group, but it is the part of the cloud portfolio most exposed to best-of-breed and AI-native competition, and it has now slowed to single digits. It weakens the claim that the end-to-end suite pitch is lifting the line-of-business applications as well as ERP.

Software Support: A Smaller Decline This Quarter

Support revenue fell 6% at constant currency to €2,434M, a smaller decline than Q2's 7% and above the €2,385M consensus. The outlook still states that the support decline rate will accelerate in coming years as customers convert to cloud.

Assessment: One quarter of a narrower decline does not change the direction, which management has guided explicitly. Read together with the European commitments agreement and the reported returns from third-party maintenance, it suggests the support base is eroding through conversion rather than through defection, which is the better kind of erosion for this thesis.

Services and the Core Services Segment: Margin Given Up on Purpose

Services revenue fell 5% at constant currency, and the Core Services segment profit fell 40% to €113M, taking the segment margin to 11.6% from 18.9%. Segment cost of cloud rose to €64M from €24M, consistent with consultants delivering cloud and AI adoption work that is not billed as services.

"You get a meager margin on an hour you sell, but if you can convert that into recurring high margin AI data business, and we feel there is more value to be had from that." — Dominik Asam, CFO

Assessment: This is a deliberate transfer: SAP is giving up services revenue and services margin to accelerate adoption of subscription products. The economics are right if adoption converts to cloud revenue, and the cost is visible. It is also now a stated headwind for 2027, which is why we treat it as part of the denominator drag rather than a one-off.

Regions: The Americas Lag Again

Cloud revenue grew 28% at constant currency in both EMEA and APJ and 21% in the Americas. The release again called EMEA and APJ "particularly strong" and the Americas "solid." Asked whether the Americas would catch up, the CEO described the US as the front-runner in AI adoption.

Assessment: The numbers do not support the answer. The Americas has carried the "solid" descriptor for every quarter we have covered and now trails the other regions by seven points of cloud growth. The CEO's explanation that the region is more penetrated in cloud is plausible and would mean a structurally lower growth rate, not a catch-up.

Key KPIs

Key Topics & Management Commentary

Overall Management Tone: Confident on the quarter and deliberately cautious on what follows, a reversal of July, when management was confident on demand and defensive on cost. The prepared remarks spent most of their length on the AI product and almost none defending margin. The CFO then volunteered a list of reasons the 2027 acceleration is harder than it looked, which reads as expectation management ahead of the January guide rather than as a reaction to anything in the quarter.

1. The Margin Beat Is Mostly the Comparison

Non-IFRS operating profit grew 18% to €3,030M and the margin expanded 120bp to 29.5%. The CFO named the reason in his prepared remarks.

"IFRS and non-IFRS operating profit growth benefited from prior year negative impacts of €100 million each from a change in case law affecting SAP's other tax litigation and from a workforce transformation, which were both recorded in the third quarter of 25." — Dominik Asam, CFO

Normalizing Q3 2025 for those two items gives a base near €2,766M, or a 30.5% margin. On that base Q3 2026 grew about 10%, or about 11% excluding the €39M of non-IFRS dilution from Reltio, Dremio and Prior Labs, and the underlying margin was 0.5–1.0pp lower. The CFO added that the investment phase is not over.

"While the operating profit strength this quarter was a result of the factors I mentioned above, it is important to remember that investments to accelerate our AI transformation and to further enhance our products are in full swing." — Dominik Asam, CFO

Assessment: The quarter met the second-half profit commitment we graded from July, and the consensus beat is genuine because the panel already expected 15% growth on the easy base. What it does not show is margin expansion. Like-for-like, SAP gave up half a point to a point of operating margin while funding research and development growth of 13%, a recovering share-based compensation line and acquisition dilution. That is a defensible investment choice, and it is not the margin inflection our original thesis described.

2. Cloud Revenue: 25%, With a Point of Catch-Up

Cloud revenue grew 25% at constant currency, ahead of Q2's 24% and the 23–25% full-year guide. The CFO sized the one-off component himself.

"Cloud revenue grew to approximately €6.7 billion up 25% helped by several 1-off positive effects, which added approximately 1 percentage point of growth in Q3. This was due largely to a positive revenue recognition effect which in turn mainly stem from a catch up of revenues from customers returning from financial distress." — Dominik Asam, CFO

For the fourth quarter he back-solved the guide to about 23% and listed what falls away: easier Q3 2025 comparisons, a half point from SmartRecruiters annualizing, and the slower backlog growth that exited 2025.

"And this is why we now say it is kind of towards the high end of the range, and then you can back solve the implied cloud revenue growth. And you will find it is around 23% there, So there is no magic." — Dominik Asam, CFO

Assessment: The upper end of the €25.8–26.2B range implies about €6.88B of fourth-quarter cloud revenue at constant currency, up 22.6% on €5,610M, which matches the CFO's 23%. The midpoint would have implied 19%, so moving to the upper end is a real raise of about €200M. The financial-distress catch-up is the least repeatable part of the quarter, and the allowance on trade receivables increased by €54M in the nine months against €10M a year earlier, so the same credit cycle that returned some customers is also pushing others the other way.

3. The Backlog Guide Still Says Slow Down, Under a New Macro Assumption

The outlook kept the slight deceleration guide for current cloud backlog but changed its condition. In July the outlook rested on a near-term Middle East de-escalation. The Q3 statement instead says the backlog outlook assumes a stable macroeconomic environment in the fourth quarter. Asked why management would not be more optimistic, the CFO pointed to the size of the quarter and the backdrop.

"But, again, there is half of the bookings of the year to be still executed in Q4, and if I look at the current macro backdrop, it is not so great. I mean, you have seen the oil price going up." — Dominik Asam, CFO

Assessment: The geopolitical condition has become a general macro one, and the only line it still qualifies is backlog. That is a narrower dependency than in July, because the revenue, profit and cash guides are no longer conditional. The CFO's range of roughly 23–24% for the exit rate is a two-to-three point deceleration from 26% inside one quarter, which looks conservative given two quarters at 26%. We treat it as the floor the January report needs to clear, not as the base case.

4. The 2027 Acceleration: Still Planned, Now Qualified

The outlook repeats that total revenue growth will accelerate in 2027. The CFO then spent part of his prepared remarks on why that is harder than it sounds.

"First, as a reminder, the fourth quarter accounts for around 50% of our annual bookings volume. Given the continued political and macroeconomic uncertainty we face today, this could weigh on Q4 bookings, and as a result, our 2027 trajectory." — Dominik Asam, CFO

"As a result, we expect to trend towards the upper end of our cloud and software revenue outlook which will raise the bar against which we will measure our 2027 revenue acceleration. Taken together, we are still planning for total revenue growth to accelerate in 2027, but must be cognizant of all the factors that make this trajectory more challenging." — Dominik Asam, CFO

The three factors are fourth-quarter bookings at risk from macro, a services line likely to keep shrinking next year by design, and a higher 2026 base after this quarter's beat.

Assessment: This is the most important change in the call. In January and April the 2027 acceleration was the reason to look through a flat 2026; it is now hedged by the company that set it. The arithmetic supports the caution. With cloud near 25% and support and services declining, total revenue growth of about 11% in 2026 needs cloud near its current rate in 2027 and a slower decline elsewhere to accelerate. We now model 2027 total revenue growth of 11–12% at constant currency, an acceleration from a 2026 base near 11% that is modest at best.

5. Cloud Gross Margin: A Third Decline, and a New Goal of Stability

Non-IFRS cloud gross margin fell 0.7pp at constant currency to 74.3%, after declines of 0.1pp in Q1 and 0.7pp in Q2. Cost of cloud rose 30% against cloud revenue growth of 26%. The CEO repeated the explanations from July and added hardware costs.

"I mean, when you look onto the gross margin, as I also said in the last earnings call, I mean, yes, this year, we have some 1 timers. We are investing in software and cloud. We, of course, have higher hardware prices." — Christian Klein, CEO

The CFO's framing of the target was more telling than the explanation.

"We are going to embark quite some acquisitions, which is not so much topic on gross margin, but more a topic on operating margin, which have to be absorbed, so we will need all these levers to really make sure that our gross margin, the cloud gross margin is stable and we can sustain that." — Dominik Asam, CFO

Assessment: In our July thesis review we set a third quarter without expansion as the point at which expansion having stopped becomes a decline having begun, and as the test that breaks our margin-and-cash pillar. It has been met: 74.3% is below the 74.6–75.2% band we used in July to argue that only expansion had stopped, and management's own language has moved from expansion to stability. We therefore mark that pillar broken. Cash, its other half, delivered this quarter, and we carry it in the valuation rather than as a pillar. The levers named for 2027, automation of cloud operations and cutting third-party contractors, are credible, and the hardware-price explanation is new and external. In July we had already moved our estimate from expansion to a flat 74.5–75.0%; we now lower it to 74.3–74.7%, and operating margin progress depends on operating expenses.

6. Services Shrink by Design, Likely Into 2027

Services fell 5% at constant currency, and the CFO said the redeployment of consultants into adoption support will likely put more pressure on services revenue next year as well.

"In fact, we believe in knowledge worker industries, business models, which solely rely on selling more hours to customers will be more and more challenged by AI enable lower cost, more scalable and nevertheless higher margin alternatives. So further portfolio adaptation in services is a necessity." — Dominik Asam, CFO

Assessment: SAP is applying to its own services business the AI-disruption argument that the market had applied to SAP. It is strategically consistent, and it costs about €1B of quarterly revenue growing negative into 2027. With support also guided to decline faster, the two lines that are about a third of revenue are likely to subtract more from 2027 growth, not less. This is the denominator drag in our thesis, now likely to persist for another year.

7. AI Is Named as a Cloud Revenue Driver

In the four previous calls we covered, management described AI by deal participation. This time the CEO attributed cloud revenue growth to it.

"The strong growth in cloud revenue was significantly driven by our Business AI, data, and platform business. Continuing to trend from recent quarters 9 out of 10 of our 50 largest deals included AI and the Business Data Cloud." — Christian Klein, CEO

Joule Work, the conversational workbench launched at Sapphire, is now in customers' hands, and SAP described its platform as open to more than 70 frontier and open-weight models without a preferred provider. The CEO also drew a line on data use.

"At SAP, the customer's data remains the customer's data. We never upload customer data or process IP into large language models. Not for training, not for model improvement, not for any purpose." — Christian Klein, CEO

Assessment: The claim that AI and data are driving cloud revenue is the most important qualitative change for the long-term case, and it came without a number. Business Data Cloud, the Business AI platform and the newly acquired Reltio and Dremio sit inside cloud revenue and are not broken out. Model neutrality and the data commitment are a coherent positioning against vendors tied to one model provider. Until SAP sizes the AI and data contribution, we credit the attribution as directional support for Bull #3, not as a measured revenue stream.

8. Internal AI Efficiency: "Up to €3 Billion"

Answering a question on margins next year, the CEO gave a larger efficiency figure than the one management set in January and repeated in April, a run-rate of around €2 billion by the end of 2028.

"We are confident, as I also said, we are confident to really also gain up to €3 billion efficiency gains over the course of the next 2 years. And, yes, we will continuously update you on the progress." — Christian Klein, CEO

He cited internal evidence: Joule Work Desktop live for more than 110,000 employees, productivity gains of 20% in central functions, and roughly 70% more code contributions from developers this year.

Assessment: Up to €3 billion against a 2026 operating profit guide of €11.8–12.2B is a large number, and the call did not reconcile it with the €2 billion run-rate target, state how much reaches the margin rather than funding reinvestment, or address whether it requires a restructuring charge. In January the CEO said there was no restructuring plan but that the position could not be held forever. We treat the figure as support for the 80–90% operating-expense envelope in 2027 and as a reason to expect a charge at some point, not as incremental earnings.

9. The Buyback Comes Back

The release reports cumulative repurchases of 25,205,983 shares at an average of €164.79, about €4.2B, as of September 30, against 16,280,097 shares and about €2.6B at June 30. SAP therefore bought about 8.9 million shares for roughly €1.53B in Q3, at an average near €171, under a second tranche of up to €2.6B that began on July 27 and runs to January 27, 2027.

Assessment: This answers the sharpest criticism in our July note. The tranche started three days after the Q2 print, close to the low, and the shares have risen about 50% since, so the timing is good in hindsight. The funding is less comfortable: nine-month dividends, buybacks and acquisitions of €8.8B exceeded free cash flow of €8.0B, and SAP drew €4.7B of new borrowings. At about €5.8B remaining authorization through 2027, the buyback is roughly 2.4% of the current market value.

10. Free Cash Flow: A Strong Quarter and a Demanding Fourth Quarter

Q3 free cash flow rose 36% to €1,718M, but the approximately €10B full-year guide still requires about €2.03B in the fourth quarter, against €1,034M a year ago. Asked about it, the CFO said migration credits would weigh on cash, though not enough to threaten the guide, and raised receivables himself.

"But to be honest, on Q3, you see a little bit of a for my taste, high accounts receivable. So we have to tighten the screws on our collections a little bit more. Maybe the high interest rates create some incentives here to. For customers to be a little bit slower sometimes if they find a good excuse." — Dominik Asam, CFO

Assessment: The second-half bridge we flagged in July is on track: €1.72B delivered of the roughly €3.75B required. The fourth quarter is the largest collection quarter, so nearly doubling it is not as extreme as the percentage suggests, and Q4 2025 was depressed. The receivables comment, the higher allowance on receivables and slower customer payments under high rates all point the same way, though. This is the line where a miss is most likely, and we keep our full-year estimate at €9.5–10B with the risk toward the lower half.

11. A Fourth Acquisition Inside an Unchanged Guide

After closing Dremio on July 6 and Prior Labs on July 16, SAP agreed to acquire TechWolf, a skills and work-intelligence platform that the CEO tied to HR use of Joule Work. The three closed deals contributed €49M of revenue and reduced non-IFRS operating profit by €39M in Q3. The operating profit guide was unchanged at €11.8–12.2B, which already included the Dremio and Prior Labs dilution from July.

Assessment: Four acquisitions announced since late March is the pattern we flagged in July, and this time it came with the buyback running in parallel rather than instead of it. The dilution is small relative to operating profit. The question that remains is whether January's 2027 guide absorbs TechWolf and any further deals inside the 80–90% expense envelope or adjusts around them, as the 2026 guide did twice.

Guidance & Outlook

Disclosed Currency Translation Impact

Currency was a tailwind in Q3, with reported cloud growth of 26% against 25% at constant currency, roughly the effect we flagged in July though a little smaller. For the fourth quarter the tailwind widens to 2.0 points on cloud. Reported fourth-quarter cloud growth near 25% would therefore be consistent with the constant-currency deceleration to about 23% that management has guided.

Implied fourth quarter: At the upper end of the cloud range, fourth-quarter cloud revenue is about €6.88B at constant currency, up 22.6%. At the upper end of cloud and software, that line is about €9.61B, up 11.5%. The unchanged operating profit range implies fourth-quarter operating profit of €2.92–3.32B at constant currency, between +3% and +17% on €2,829M, with the midpoint at +10%. Free cash flow needs about €2.03B.

Street at: The October 19 panel had FY26 operating profit at €11,780M and EPS at €7.06, with FY27 at €13,645M and €8.37. Nine-month operating profit growth of 17% at constant currency already sits at the top of the 13–17% range, so leaving the range unchanged while raising revenue implies management expects fourth-quarter profit growth near 10% and is holding the difference for investment.

Guidance style: Conservative on revenue and backlog, unchanged on profit. Management has now raised or confirmed the revenue lines at every print this year while adjusting the profit line only for acquisitions, a pattern consistent with a company that prefers to reinvest outperformance rather than report it.

Analyst Q&A Highlights

How Much Caution Is Built Into the Fourth-Quarter Backlog Guide

The opening question quantified the guide: a two-point deceleration from 26% would mean 23% for the fourth quarter, and the questioner asked whether that reflects caution or something in the pipeline. The answer cited macro rather than pipeline and loosened the implied point estimate.

Q: "Think you have said before that slight decel meant about 2%. Which would imply 23% for the fourth quarter versus we ended Q3 at 26% Could you just talk a little bit about why you are not willing to be a bit more optimistic here? Is it just an abundance of caution given the scale of the quarter? Is there anything you are seeing in pipelines or in customer decision making that would give you that greater caution at this point in the year?"
— Adam Wood, Morgan Stanley

A: "Interest rates have gone up, and we have all seen long term analysis that when interest rates reached 5%+, and they have reached that in US now, these interest rates start to develop teeth. So I would say it is probably reasonable to not assume that the macro is super conducive in Q4… So I would say the upper end of that range might be 24% or so. So I do not wanna be necessarily saying that there is a point value of 23%."
— Dominik Asam, CFO

Assessment: The answer named no pipeline weakness, and the CEO separately called the fourth-quarter pipeline strong. The caution is about macro risk to the largest bookings quarter, not about observed demand. Lifting the reference to "24% or so" at the upper end is a small concession that the 23% arithmetic was too low. We read the guide as conservative.

Whether the Backlog Is Running Ahead of Internal Plan

A question noted that backlog growth was stable at 26% and looked slightly faster organically, against an unchanged guide for deceleration, and asked directly whether it is ahead of internal expectations and which products are surprising. The CEO answered with products and customer anecdotes and did not address the plan.

Q: "So 26% constant currency stable versus Q2 and looks to be slightly accelerating on an organic basis. But your guidance has always been for that slight deceleration. And that is unchanged. So is the backlog currently growing faster than your own internal expectations at this point? And if it is, could you help us with what products specifically are surprising you on the upside and why that is happening?"
— Toby Ogg, JPMorgan

A: "Second, the business case. I mean, you know, every business case is now, you know, with AI, as I said. The majority of our deals are closed with AI. Even more important, the business data Cloud, after now 2 years now being in the market… Obviously, we can adopt AI very fast on the customer side, but it takes longer if you have a messy system. And that has nothing to do with our AI or with the technology."
— Christian Klein, CEO

Assessment: The question about internal plan went unanswered, which is normal for a company that does not disclose its plan, but the product answer was specific: Business Data Cloud is the product the CEO singled out, and the clean-data argument ties AI demand to the migration SAP already sells. The admission that AI adoption is slower on messy systems is candid and supports the view that AI pulls RISE migrations forward rather than replacing them.

AI Credits, Consumption Pricing and the Backlog-to-Revenue Bridge

A question asked whether the push toward AI credits is already affecting the income statement and the relationship between booked backlog and revenue. The CFO said credits are amortized over the contract and that consumption pricing will enter gradually.

Q: "And I guess related to AI, I am getting partners beginning to feel the push towards AI credits into the business model Dominik, can you shed some light into whether that is influencing some of the metrics in the P and L yet? And the dynamics between perhaps what you can book a CCB and what ends up in revenue?"
— Charles Brennan, Jefferies

A: "So I would say in the near term, this kind of logic how to bring CCP down to cloud revenues is still quite intact. Yes. It will loosen over time as we are going into more and more hybrid models all the way to full consumption based and as we wanna sell outcomes to customers as opposed to tokens, And for outcomes, it needs consumption and adoption… That is not something you should be kind of concerned about in the bridge to 2027. This will be still quite traditional, I would say."
— Dominik Asam, CFO

Assessment: This is useful for modeling: through 2027, current cloud backlog remains a reliable leading indicator of cloud revenue. Beyond that the CFO has told investors that the indicator will become less reliable as consumption and outcome pricing grow. The thesis has relied on backlog as its anchor metric, so we note that management has set a time limit on that reliance.

The Roughly €2 Billion Fourth-Quarter Cash Requirement

A question asked for confidence in the fourth-quarter free cash flow needed to reach the guide and whether migration credits are starting to be used. The CFO held the guide and pointed to receivables.

Q: "Perhaps for Dominik, I think you need to do about $2 billion in Q4… Are you starting to see from the migration credit being, if you like, utilized by customers?"
— Michael Briest, UBS

A: "Indeed, I mean, we always pointed out to the fact that migration credits at some point in time will affect our free cash flow, not to the degree that I would feel nervous about the full year guidance of €10 billion. Yes. there is a remainder to do, as you mentioned. But I think we have opportunities, and I am relatively confident that this is something we can manage."
— Dominik Asam, CFO

Assessment: "Relatively confident" is hedged language, and he volunteered the receivables problem in the same answer. Migration credits consuming cash is the expected cost of the RISE incentive structure, and management confirmed it has started. The guide is probably achievable, but this is the one line on which the CFO sounded less than certain.

What Holds Margins Up Next Year

A question asked how operating leverage and margin expansion continue into next year with gross margin down. The CEO listed internal AI efficiencies across functions; the CFO reduced the gross margin ambition to stability and anchored on the expense envelope.

Q: "We have gross margins down. You seem pretty confident about the ability to return to growth, but it would be good to have your bit more detail on the rest of the model in terms of operating leverage and ability to keep on expanding margin next year."
— Frederic Boulan, Bank of America

A: "But the overall model is intact. I mean, we are firmly committed to keeping that to containing the growth in operating expenses, total operating expenses, including cost of goods sold, to this 80% to 90% growth of the top line."
— Dominik Asam, CFO

Assessment: The 80–90% ratio is the commitment to grade in 2027, and it implies modest margin expansion of roughly half a point to a point a year at 11–12% revenue growth. Management did not promise more than that, and did not promise gross margin expansion at all. This answer anchors our FY27 framework.

Model Routing as a Gross Margin Lever

A question asked whether routing AI tasks across models could become a source of cost savings for customers or for SAP's margin. The CEO said model switching is increasing, including more open-weight models, and pointed to automation of cloud delivery.

Q: "In the light of the gross margin for the business, we obviously have some headwinds at the moment, but this model routing could become quite a nice source of potential cost savings that you can either pass on to your customers or maybe keep some of that in your margins. Just how should we think about that conceptually going forward?"
— Johannes Schaller, Deutsche Bank

A: "We are working heavily on further automation opportunities so that we are also able to cut some of third-party contractors we deploy in our cloud delivery. So we remain confident that we also see in the upcoming years a healthy development of our cloud gross margins."
— Christian Klein, CEO

Assessment: "Healthy development" is less than a commitment to expansion and fits the CFO's "stable." The concrete levers are internal operations and contractor costs rather than AI model pricing, which suggests token costs are not yet the swing factor in cloud gross margin. Model routing remains an option on future margin; the 2027 cloud gross margin story is an operations-cost story.

Headless Software and Model Neutrality

A question raised the two debates shaping software valuations: whether applications become headless data sources accessed through other vendors' AI interfaces, and whether model-agnostic vendors or exclusive partners win. An audio failure interrupted the call, and the answer that followed addressed agent accuracy rather than either debate.

Q: "You know, firstly, around software potentially going more headless, so accessing data and insights from a plug in versus retaining that AI UX. And then secondly, perhaps, as you mentioned earlier, you know, a divergence between those staying AI model agnostic, giving choice, and those partnering more exclusively."
— Ben Castillo-Bernaus, BNP Paribas

A: "And with that, we see also then, of course, also further optimization of the price outcome ratio. So again, we are working there, and we see very, very good progress quarter over quarter."
— Christian Klein, CEO

Assessment: The headless question goes to whether SAP keeps the user relationship as AI interfaces consolidate, and it went unanswered on this call. The disruption probably explains the brevity rather than evasion, and the prepared remarks made the opposite case through Joule Work and the governed ontology layer. It remains the central open question for the multiple, and TechEd next week, where the CEO promised more platform announcements, is the next chance for an answer.

Travel and Expense Competition at Concur

A question asked about new competitors growing fast in travel and expense, where Concur has long dominated. The CEO separated the expense business from the booking business.

Q: "I am wondering what you think about the travel and expense side of the business where Concur has been so dominant for so long, but now you are seeing some new competitors growing fast?"
— Patrick Walravens, Citizens

A: "With a lot of regulations, customers have Concur in the meantime extremely deeply integrated to our finance solutions. So it is a very, very stable business. Obviously, the very, very minor part is on the booking side, and there are, of course, a lot of new incumbents in the market. But this is not the main part of Concur's business."
— Christian Klein, CEO

Assessment: The answer concedes the booking side to new entrants and defends the expense-to-pay integration with finance, which is the stickier and larger part. It fits the broader pattern: SAP's defense is depth of integration with the system of record. Read alongside the Extension Suite slowing to 8%, it suggests the line-of-business applications face more competitive pressure than ERP.

What They're NOT Saying

1. Why the operating profit guide did not move: Nine-month operating profit growth of 17% at constant currency is at the top of the 13–17% range, and revenue moved to the upper end of its ranges, yet the profit range was left unchanged without explanation beyond investments being "in full swing." The implied fourth-quarter growth near 10% at the midpoint is a choice to reinvest that management did not state as such.
2. How much revenue AI and data generate: The CEO said Business AI, data and platform significantly drove cloud revenue growth. No revenue, attach rate, Business Data Cloud contribution or consumption figure was given, and the release contains none.
3. The credit cycle behind the catch-up: About a point of cloud growth came from customers returning from financial distress, while the allowance on trade receivables increased by €54M in the nine months against €10M a year earlier and the CFO called receivables high. Neither the size of the reversal nor the exposure to customers moving the other way was quantified.
4. How the €3 billion efficiency figure relates to the €2 billion target: The CEO's "up to €3 billion" over two years was not reconciled with the January target, and there was no word on how much reaches the margin or whether a restructuring charge is needed.
5. Whether the backlog is ahead of plan: A direct question went unanswered. With the guide unchanged after two quarters at 26%, the gap between the stated outlook and the run-rate has now persisted for two quarters.
6. The headless-software risk: The question most relevant to the valuation debate received a disrupted, non-responsive answer, and the prepared remarks addressed it only by implication.
7. Why the Extension Suite slowed to 8%: The line slowed four points without comment in the prepared remarks or the release.

Market Reaction

• Pre-print setup: The NYSE-listed shares closed at $232.73 on October 21, up 11.2% over the trailing thirty days and down 4.2% year to date against +13.3% for the S&P 500. They were down 17.0% over the trailing twelve months, and the 52-week closing range was $146.38 to $280.46, the low being the July 23 close before our last note.
• After-hours move: The ADR added about 1.0% after the release, to around $235.
• Next-day session: Shares opened at $237.06, a 1.9% gap, traded between $236.29 and $243.05, and closed at $240.59, up 3.4% or $7.86. That is the highest close since our July note and 14.2% below the 52-week high.
• Volume: 5.2M shares against a 2.0M thirty-day average, about 2.6 times normal.
• Benchmark and peers: The S&P 500 fell 1.2% on the session. SAP moved in line with several US application software peers that rose by similar amounts, while the broad software ETF was flat.

A Beat Met With a Shrug, Then Bought. The Frankfurt listing fell more than 1% after the open before turning higher, and the ADR finished near the upper end of its range on more than twice normal volume. The opening hesitation matches the pre-print argument that an in-line quarter would not be enough after the rally: the print had to clear a raised bar. It did, on every revenue and profit line, and by the close the market had decided the beat and the upper-end cloud guide mattered more than the unchanged profit range and the 2027 caveats.

The Rally Came Before the Print. The larger story is the setup. The shares have risen 50.4% since our July 24 note, against +3.4% for the S&P 500, as the backlog reacceleration, the buyback restart and a broader recovery in software stocks reversed the AI-disruption discount that took the stock to its July low. A 3.4% reaction on a quarter this strong is modest in that context. Most of the repricing had already happened, and the reaction-day move added to a rally rather than starting one.

What the Market Did Not Price. The reaction ignored the comparison-driven margin, the third cloud gross margin decline and the qualified 2027 language. Those are the inputs to the January guide, which will be the next test of the re-rating, and none of them improved this quarter.

Street Perspective

Debate: Is the Backlog Guide Conservative, or Is the Fourth Quarter at Risk?

Bull view: Backlog growth has held at 26% for two quarters against a guide for deceleration, the fourth-quarter pipeline is described as strong, and the guide's implied 23–24% looks built to be beaten. Desks that see the unchanged backlog outlook as conservative relative to demand expect January to show an exit rate in the mid-twenties and a firm 2027 acceleration.

Bear view: The fourth quarter carries half of annual bookings, oil has nearly doubled and US rates are above 5%, and management itself chose this call to list why 2027 is harder. A company with strong pipeline visibility that declines to raise its backlog guide in October is telling investors something about the range of outcomes.

Our take: The bulls are right about the backlog, and the bears are right about 2027. Two quarters at 26% make an exit rate of 24–25% our base case, above management's range. But the backlog is not the whole bridge to 2027: services and support are likely to subtract more next year, and the CFO has raised the 2026 base himself. We expect a January guide with acceleration that is real and small.

Debate: Has AI Turned From a Threat Into a Revenue Driver?

Bull view: Management now attributes cloud revenue growth to Business AI, data and platform; nine in ten of the 50 largest deals include AI and Business Data Cloud; Joule Work is live and used internally at scale; and SAP's governed ontology and refusal to train models on customer data are a defensible position against model-provider-led applications. The AI-disruption discount that drove the stock to its July low has been disproven by the numbers.

Bear view: No AI revenue figure exists. The headless-software question went unanswered, the Extension Suite has slowed to 8%, and the CFO has said consumption pricing will loosen the link between backlog and revenue after 2027. The numbers show that ERP migration is strong, which was never in doubt; they do not show that AI is a separate revenue stream.

Our take: We side with the bulls on direction and with the bears on evidence. The attribution of growth to AI and data is a meaningful change in management's language, and the migration-acceleration argument is consistent with the numbers. The disruption risk was never about the core ERP system of record; it is about the user interface and the line-of-business applications, and the Extension Suite slowdown is the one data point this quarter that bears watching on that score.

Debate: After a 50% Rally, What Is the Stock Worth?

Bull view: The shares remain 14% below their 52-week high, EPS is compounding in the mid-teens through 2028 on consensus, free cash flow is guided to about €10B with a buyback running, and the business just delivered its best quarter of the year. The July low was an AI-panic price, and the re-rating has further to run as the 2027 acceleration is confirmed.

Bear view: At about 30x 2026 earnings, the stock prices a margin expansion that this quarter did not show on a like-for-like basis, a cloud gross margin that has declined for three quarters, and a 2027 acceleration that management has started to hedge. The cheap entry point is gone.

Our take: We side with the bears on valuation, and only partly on the business. Our Outperform rested on a de-rated price and a business executing better than the price implied. Both played out: the business executed on demand, cash and capital return, and the price recovered. The margin pillar did not hold. At 26x our FY27 estimate and a 4.1% free cash flow yield, the stock sits just below the midpoint of our fair value range. We would buy it again on a pullback or on a January guide that makes the 2027 acceleration firm.

Model Update & Valuation Framework

Our FY26 estimate. We estimate fourth-quarter non-IFRS operating profit near €3.18B at actual currency, about 11% growth at constant currency plus the disclosed one-point currency tailwind. After about €0.1B of net non-operating cost and a tax rate near 29.5%, consistent with the full-year guide of about 29%, and about 1,143M average shares, that gives fourth-quarter EPS of about €1.90 and FY26 non-IFRS EPS of €7.05–7.15, in line with the €7.06 panel median.

Our FY27 framework. We assume constant-currency total revenue growth of 11–12% on a FY26 base near €40.3B, cloud growth in the low twenties, support declining faster and services shrinking again. Operating expenses grow at about 85% of revenue growth, the middle of management's commitment, which yields operating profit growth in the mid-teens and a margin near 30%. With the share count declining about 1–2% on the remaining buyback and a tax rate near 29%, FY27 non-IFRS EPS is €8.10–8.40. That is slightly below the €8.37 panel median, because we weight the CFO's 2027 caveats more heavily. The 2027 reporting changes, which will exclude currency effects and tax-related interest from non-IFRS results, will require restating the base; we will rebase when SAP reports 2026 on the new definitions.

Valuation. At the reaction-day close of $240.59 and the 1.122 USD/EUR rate on October 22, the shares are worth about €214 each. On about 1,150M shares, the market value is near €247B. That is 30x our FY26 estimate of €7.10 at the midpoint and 26x our FY27 midpoint of €8.25, with a 4.1% free cash flow yield on the FY26 guide and 4.5% on FY27 panel free cash flow of €11.2B. In July, at $160.00 (about €141), the same measures were 20x FY26 and 6.2%.

Fair value: $231–259. We value SAP at 25–28x our FY27 midpoint of €8.25, or €206–231 a share, which translates to $231–259 at 1.122. That is roughly 29–33x our FY26 estimate, and we anchor it on our own record with this stock rather than on today's price. In January we upgraded at about 23x current-year earnings because that price treated the franchise's growth as in question. In October 2025 we initiated at Hold at about 39x because that price required an acceleration the backlog was not showing. A share-gaining business compounding EPS in the mid-teens belongs between those marks; we stop well short of the upper one because cloud gross margin is no longer expanding, operating margin progress now depends on cost discipline within an 80–90% envelope, and management has qualified the 2027 acceleration. The midpoint of $245 implies about 2% upside, and the €2.50 dividend adds about 1.2%, for a 12-month total return near 3%. A de-rating to 20x on a weaker FY27 of €7.90 would put the shares near $177, about 26% lower. A re-rating to 30x on €8.40 would put them near $283, about 18% higher and at the 52-week high. The scenarios skew to the downside.

Why Hold rather than Underperform. A total return near 3% is below what we would need to expect SAP to beat the S&P 500, and the downside skew rules out staying at Outperform. We stop at Hold because most of the downside case is a de-rating back toward 20x, which in our view would take either a renewed AI-disruption scare of the kind that drove the July low or a January guide that drops the 2027 acceleration. Two quarters of 26% backlog growth make the second unlikely, and the top of our range, about a 9% total return, is within reach on a January guide that formalizes the acceleration. On balance we expect SAP to perform roughly in line with the market, with a wider range of outcomes than usual.

The five-quarter arc, stated plainly: we initiated at Hold in October 2025 at about 39x forward earnings on a 2.9% free cash flow yield, upgraded in January at about 23x and 5.1%, held Outperform in April at 21x and in July at 20x as the price kept falling, and now downgrade at 30x FY26E and 4.1%. From the $200.21 close on which we upgraded in January to this close the shares are up 20.2% against +10.0% for the S&P 500; from July they are up 50.4% against +3.4%. The upgrade was wrong on price for two quarters and right over the full period. The Hold trigger we set in July required a third quarter without cloud gross margin expansion combined with second-half operating profit below the guided range. Only the first has happened: third-quarter operating profit grew 18%, and the fourth quarter needs only about 3% growth to keep the year inside the range. The downgrade is therefore mainly about price. The business delivered on demand, cash and capital allocation, and the one fundamental change, the end of cloud gross margin expansion, lowers the multiple we will pay more than the earnings we expect.

Thesis Scorecard Post-Earnings

Grading Last Quarter's Commitments

Overall: Thesis mixed on the business and weaker on valuation. The core franchise, the backlog, cash and capital allocation all delivered; the margin pillar broke on the test we set in July, and the backlog risk has returned through the multiple rather than the metric. Of eight prior-quarter commitments, three were beaten, delivered or on track, one was right in direction only, three went against the thesis, and one was superseded.

Action: Downgrade to Hold from Outperform; fair value $231–259 (25–28x FY27E EPS midpoint of €8.25), with conviction at 6/10. This is mainly a valuation downgrade; the July Hold trigger did not fire. We upgraded in January because a share-gaining franchise was priced as if its growth were in question, and we held the call through two quarters of a falling price. Since July the shares have risen 50% against +3% for the market, and the quarter, strong as it was, did not lift our estimates by nearly as much: FY26 EPS is unchanged at about €7.10, and FY27 is a new framework of €8.10–8.40 that sits slightly below consensus. At $240.59 the stock trades just below the midpoint of our fair value with about 3% total return to it. We would return to Outperform on a pullback toward $205, about 22x FY27E, with the thesis intact, or on a January guide that formalizes 2027 total revenue acceleration with operating expenses at the low end of the 80–90% envelope. We would move to Underperform if backlog growth falls below 22% at constant currency while margin continues to compress, the trigger we set in July, or, as a new condition, if the January guide abandons the 2027 acceleration. Going forward we watch the January backlog exit and 2027 guide, fourth-quarter free cash flow and receivables, any reconciliation of the efficiency target, and the TechEd platform announcements.