Intel's (INTC) stock took a major hit Friday, falling more than 10% after the chipmaker reported fourth quarter results that topped estimates, but missed badly on analysts' first quarter projections.

In the first quarter, Intel said it anticipates earnings per share of $0.13 on revenue of between $12.2 billion and $13.2 billion. Analysts were looking for $0.34 per share on revenue of $14.2 billion.

In particular, Intel CFO David Zinsner said the company expects to see revenue for its Data Center and AI business to decline by double-digit percent sequentially in Q1. CEO Pat Gelsinger, however, said that the declines are fairly seasonal quarter-to-quarter. That's after the business missed analysts' expectations in Q4, with revenue coming in $4 billion, short of $4.1 billion in revenue projected by Wall Street.

Both Zinsner and Gelsinger, however, said they expect to see growth in the segment beyond the first quarter. But analysts are looking for the company to report gains sooner than later.

The Data Center and AI business is Intel's opportunity to take on rivals including Nvidia (NVDA) and AMD (AMD). In December, CEO Pat Gelsinger showed off the company's upcoming Gaudi3 accelerator designed to power generative AI software and services.

Overall Intel reported Q4 adjusted earnings per share of $0.54 on revenue of $15.4 billion. That was better than the $0.44 on $15.2 billion analysts anticipated.

The chipmaker's Client Computing Group posted fourth quarter sales of $8.8 billion, better than the $8.4 billion forecast by analysts and up 33% from a year ago. Intel's Client Computing Group is still its main breadwinner, and with PC sales expected to return to growth throughout 2024, the segment will be especially important for the company in the year ahead.

Intel is hoping its Core Ultra line of chips will further goose PC sales in the year ahead as consumers and enterprise customers who purchased new laptops and desktops at the onset of the pandemic in 2020 begin to look for replacement devices.

The Core Ultra includes a neural processing unit, or NPU, which allows PCs to run certain AI applications locally rather than requiring users to rely on cloud-based applications. The idea is that NPUs will let users access AI apps without having to be online. That would ensure users’ data stays on their devices rather than requiring them to send it to an AI firm’s servers.

But it’s still unclear exactly how useful onboard AI will actually be for consumers. And even Intel admits that it’s not quite sure what AI applications will look like outside of a few early examples, such as local versions of ChatGPT-like apps and AI image-editing software.