Tech Newsround – ASML & Nvidia

ASML Q2 25 – China bites

• ASML has reported good Q2 2025 results, but removed its forecast of growth in 2026, as I suspect that China will be fast-falling as a percentage of sales as a result of its inability to buy most of ASML’s products.
• Bookings, the most watched number, came in at €5.5bn, ahead of estimates of €4.8bn, a good indication of medium-term strength, but bad news followed.
• Q2 2025 revenues / EPS were also good at €7.7bn / €5.91 just ahead of estimates of €7.55bn / €5.24.
• However, Q3 25 guidance was also a bit soft with revenues of €7.4bn – €7.9bn (€7.7bn) adrift of forecasts of €8.3bn, which was put down to macroeconomic and geopolitical conditions.
• This looks to me to be more signs of the Chinese stockpiling of equipment ahead of export restrictions coming to an end, which is a theme that ASML appears unwilling to address in its commentary.
• With its main customers still expanding capex driven by demand for AI, ASML should have good visibility on its earnings, meaning that something else is impacting its business.
• I don’t think this issue has yet been fully baked into the numbers, meaning that there is more of this to come.
• ASML remains expensive for a company with this level of uncertainty.

Nvidia vs. Intel vs. Qualcomm – Advantage Qualcomm

• The long-awaited Nvidia CPU for laptops looks like it has been delayed again, highlighting that it is harder than it looks to get this to work, meaning that Qualcomm has a bigger edge in this space than previously thought.
• I have been waiting for some time for both Nvidia and MediaTek, either jointly or separately, to launch a product, but Computex 2025 came and went with barely a murmur (see here).
• Nvidia is definitely working on a product, but it has been having problems with the launch currently expected in early 2026, which would put it on course to headline at Computex 2026.
• However, it looks like there have been further problems with overheating under load and signal quality, which may require the chip to be redesigned.
• This could push the launch back even further, and now it is possible that it could be late 2026 before Nvidia is even ready to launch it, meaning that it might be CES 2027 before we see a real push.
• This is good news for Intel but especially for Qualcomm, as it looks like it will have the Arm-based laptop CPU market all to itself for longer than expected.
• It also points to higher barriers to entry, and by the time that Nvidia is ready with its first generation product, Qualcomm should already be on version 3.
• This gives Qualcomm the opportunity to cement its lead and stay ahead of its competition, meaning higher revenues and profits in this segment.
• I continue to think that Qualcomm’s significant potential in the PC market remains largely ignored by the market, which is why I think that there is still good growth ahead.
• This, combined with the underappreciated position in automotive and whatever it intends to launch for the AI data centre (for which it already has a big customer, (Humain) is why I remain a happy shareholder.