Intel has reported double-digit growth in client computing revenue, growing demand for AI PCs, and promised of strong gains in the second half of 2024 - but also reported a first quarter loss that sent the chip biz's stock sliding in after-hours trading Thursday.

On the x86 giant's Q1 earnings call, CFO David Zinsner's second quarter revenue forecast of between $12.5 billion and $13.5 billion fell short of the $13.6 billion financial analysts anticipated.

While the market was less than impressed by Intel's Q2 predictions, its Q1 revenues landed just ahead of expectations, growing nine percent year over year to $12.7 billion. Much of this was driven by a 31 percent YoY ramp in client computing sales in the first three months of this year, which accounted for $7.5 billion of Intel's overall revenues. But that wasn't enough to stem an overall $437 million loss for the chip giant.

It seems Intel has the Core Ultra processors that power its AI PC push to thank for at least some of the surge in client computing sales. On Thursday's earnings call, Intel CEO Pat Gelsinger praised the processors and said that he expects sales to double in the second quarter.

In fact, if it weren't for bottlenecks in wafer level assembly, Gelsinger implied that its shipments would be even higher. Instead, Zinsner blamed these challenges as likely to contribute to flat client revenues in the second quarter.

Intel nonetheless expects to exceed its original forecast of 40 million AI PCs sold in 2024.

Intel forecasts a better second half

While Intel's Client Computing Group enjoyed a strong rally during the quarter, the Datacenter and AI side of the house saw more modest five percent year over year gains during the quarter, with revenue landing at $3 billion.

The recent debut of third-gen Gaudi AI accelerators won’t help revenue to improve until at least Q3.

Having said that, Intel has more than one iron in the datacenter fire. In addition to Gaudi3, the chip biz has a pair of Xeon 6 processors due out later this year, beginning with a 288-core monster based on the Sierra Forest platform.

Gelsinger is counting on these launches to help win back lost market share and compete more aggressively with rival Nvidia.

"We now expect over $500 million in accelerator revenue in the second half of 2024 with increasing momentum into 2025 based on Gaudi3's vastly superior TCO as well as our own expanding supply," Gelsinger said.

$500 million is a very modest sum compared to Nvidia's GPU revenue, but with the supply of GPUs expected to remain constrained through the remainder of 2024, Intel could end up winning share simply by having chips to ship.

A long road ahead for Intel Foundry

Overall, Intel's Product division, which includes its client, datacenter, and networking product groups, grew 17 percent year over year in Q1, however, under its new reporting scheme, its Foundry unit remains challenged. The Foundry group saw revenues decline 10 percent year over year in Q1 to $4.4 billion.

Zinsner said he expects startup costs associated with the company's dash to bring five process nodes to market in four years to peak in 2024. He warned investors that Foundry's road to growth would be a long one, as the group looks to expand beyond servicing its own product teams to building chips for others.

"Beyond 2024, as volume begins to shift toward leadership manufacturing nodes with a competitive cost structure, scale improves - including the return of compute tiles to internal process nodes - and our efficiency actions begin to flow through the P&L, we expect to see rapid profitability improvement," he predicted. ®

https://www.theregister.com//2024/04/26/intel_q1_2024/