“Again, that starts to improve in the following year as Panther Lake becomes more and more meaningful part of the volume for the client business in 2026.”

In context

“would be very useful. David Zinsner: Yes. So on revenue growth, again, we're not going to, for 2025, I'd say just that we are managing the business in terms of how we're investing in the business to a 3% to 5% growth rate next year, but over time, in our minds. And things turn out to be better, then that's good fall through to us from a profitability perspective. I wasn't suggesting that you take the 39.5%, you move it every quarter. Clearly, every quarter is going to have some unique aspects to it. Only that 39.5% was a quarter in which it was clean and it was a good proxy to start the calculation for the full year of '25. You're right, yes, we do see more headwind in the second half versus the first half given Lunar Lake becoming a more meaningful part of the volume over time for 2025. Again, that starts to improve in the following year as Panther Lake becomes more and more meaningful part of the volume for the client business in 2026. John Pitzer: Vivek, do you have a follow-up question, please? Vivek Arya: Yes. Thank you, John. So maybe one for Pat. Pat, what does the future look like for Intel's data center if there is no competitive AI product. Is just being CPU-centric good enough? At what point does the CPU get commoditized like custom chips or replaced with ARM-based products. What is Intel's AI strategy right now? Patrick Gelsinger: Yes. Thank you, Vivek. Maybe three different perspectives just to highlight. Number one is the CPU plays an increasing role in data center AI compute. Even in cloud-based environments today, head nodes are an area of strength for Intel Xeon already, as I reflected in my comments. As you go into enterprise AI, we expect to place a more prominent role, databases, embedding, refinement”