Reading the numbers behind the headlines.

Two things deserve a closer look before extrapolating. First, the geographic breakdown: China shipments dropped to 19% of system revenue from a peak above 45% two years ago, and the gap has been filled almost entirely by a single Korean memory customer pulling forward High-NA capacity. Customer concentration in the EUV book is now more pronounced than at any point since 2019. That is a vulnerability, not just a tailwind.

Second, gross margin guidance for Q2 of 51–52% is being read as conservative, but the mix mechanics explain most of it: the High-NA systems being recognised carry materially lower initial margins until the install base scales. The real margin question is not whether 2026 prints in line — it will — but whether the 56–58% long-term band still holds once High-NA is a third of revenue rather than a rounding error.

Service revenue of €1.6bn, up 11% y/y, is the quietly important number. The installed base is now generating recurring economics that look much more like a software company than the sell side gives credit for. That is the part of the story that, in my view, justifies the multiple, far more than any one quarter of bookings.

The peak-NII debate in European banking is over — this print settles it. Eurozone deposit repricing is no longer a tailwind, and the structural hedge contribution that flatters several UK names does not exist in the same form here. Fee income at Global Banking partly offset the drag (Equity & Prime Services had a strong quarter, +14% y/y), but fee businesses are not what the equity story has been priced on for the last eighteen months.

Cost of risk came in at 28bps, well below the through-the-cycle guidance of 40bps. Provisions on Stage 1 and 2 exposures were released net — never a number to celebrate, in my experience — and the headline EPS beat depends almost entirely on this line. Strip cost of risk back to guidance and the print is a 4% miss.

"Provisions are an opinion. Net interest income is a fact. Read the press release in that order."

CET1 of 13.1% gives plenty of room for the announced €1.05bn buyback to complete without stressing the capital plan, and management's reiteration of the >12% RoTE ambition for 2027 is at least directionally credible. But the path to that number now requires either a steeper curve than the market is pricing or a meaningful pickup in Corporate & Institutional Banking. Neither is in the bag.

For four years the bear case on the European majors has been that renewables capex destroys returns and the integrated transition is a slow-motion impairment cycle. This quarter is the first one I would point to as a counter-data-point that doesn't require squinting. Integrated Power generated cash, the LNG portfolio captured a wider arbitrage than the prompt suggested, and Refining & Chemicals held up despite a soft European margin environment.

Net debt to capital employed of 11.3%, organic free cash flow of $5.1bn, and a continued $2bn-per-quarter buyback alongside the $0.85 dividend means the shareholder return policy now consumes almost exactly 100% of distributable cash at $75 Brent. That is a tighter cushion than at Shell, and it is the line I would watch into Q3 if the macro slips.

One quiet flag: working capital absorbed $1.7bn in the quarter, partly seasonal but partly reflecting an inventory build in the trading book. Trading P&L was strong enough that no one is going to complain, but it is the sort of number that, four quarters in a row, stops looking seasonal.

The transition narrative was an extraordinary one to trade: predictable revenue mix shift, multiple expansion as the market gave the company SaaS comparables, and a credible restructuring story. All of that is now in the price. From here, SAP has to re-accelerate organic cloud growth from the high-20s into the 30s, or the multiple compresses.

I am not sure the Q1 print supports the higher growth thesis. The current cloud backlog growth rate decelerated to 28% from 31% the prior quarter on a sequential basis — modest, but a deceleration nonetheless — and the Business AI revenue contribution remains disclosed only qualitatively, which after four quarters of build-up is starting to feel like a tell.

Free cash flow of €3.6bn (vs €2.9bn prior year) is genuinely impressive and underwrites the €5bn buyback commitment without any balance sheet strain. As a cash story, this is a fine print. As a growth story, the next two quarters need to do more.