Intel CFO sees strong AI server chip demand

Dave Zinsner, Intel’s chief financial officer, said at a Morgan Stanley conference in San Francisco that demand for server processors remains strong, and he expects it to continue into next year.

He said supply shortages at Intel’s factories and across the semiconductor industry limit the company’s ability to meet growing orders.

Zinsner said the server market grew more than 20% in units last year, and he expects meaningful growth again this year.

He noted that runaway demand for AI infrastructure is driving orders for processors, including Intel’s, that perform tertiary roles, spreading the supply crunch to a more diverse set of suppliers.

Many Intel factories are running above 100% capacity, and the shortfall should continually ease but persist through the end of the year, he added.

Intel shares rallied as much as 6.9% to US$46.08 in New York on Zinsner’s remarks, adding to a 17% gain this year through March 3’s close.

🔗 Source: Bloomberg

🧠 Food for thought

Implications, context, and why it matters.

Intel’s capacity problem is its rivals’ market-share opportunity

Intel’s supply crunch is not alone. Rival AMD has seen an “unexpected” jump in server central processing unit (CPU) orders tied to AI workloads. CEO Lisa Su said supply is tight 1.
Short supplies across the sector helped AMD. It reached a record 28.8% of the server CPU market in Q4 while Intel struggled to ship enough chips 2.
Intel has shifted more factory output to server CPUs to meet demand. That left fewer client PC CPUs, mainly low-end models. AMD gained share in desktop and mobile processors 2.

The AI boom is colliding with chip manufacturing economics and power-grid constraints

AI buildouts are pushing up orders for supporting processors. One industry analysis says cost scaling has ended. Transistors per dollar have cost more since the early 2010s 3.
Infrastructure budgets keep climbing. Major technology companies are expected to commit more than $1 trillion in 2025–2026 4.
Electricity supply adds another bottleneck. Morgan Stanley Research projects U.S. data center demand at 74 gigawatts (GW) by 2028. Available power access could fall short by about 49 GW 4.