Semiconductor giant Intel (INTC +1.93%) was expecting demand in its core markets to pick back up in the second quarter, but that demand failed to materialize. In the company's conference call with analysts on Thursday, chief financial officer David Zinsner said, "... the market has not recovered as expected, and we're obviously not satisfied with our results."
Intel missed analyst expectations across the board, reported an unexpectedly large loss, and guided for a bigger revenue decline in the third quarter. The company is taking drastic actions to lower costs and free up cash as it continues to plot its turnaround.
Intel will reduce its head count by more than 15%, slash operating expenses by billions of dollars over the next few years, reduce its capital spending to better reflect market demand, and suspend its dividend until cash flow improves. The market reacted to this barrage of bad news as one would expect, sending shares down nearly 30% by Friday afternoon.
There's a lot to unpack in Intel's second-quarter report, but here are a few key things that investors need to know.
NASDAQ: INTC
Key Data Points
1. Customers have too much inventory
The PC market has begun to bounce back after a post-pandemic plunge, but the recovery has been sluggish. Gartner reported that global PC shipments increased by just 1.9% year over year in the second quarter and that PC inventories were still a bit elevated.
That lines up with Intel's results. While Intel's client computing group saw a 9% jump in revenue during the second quarter, the company noted that customer inventories are still not quite normal. In the third quarter, it expects its client segment to be flat to down as that inventory is digested.
2. AI chips are crowding out CPUs
Intel faces multiple problems in its data center business. First, the company has fallen behind rival AMD in performance and efficiency. It is looking to change the story with two new server chip families this year.
The Sierra Forest central processing unit (CPU), with a large number of efficiency cores, is aimed at cloud workloads, while the more traditional Granite Rapids is meant for workloads that need raw power. Sierra Forest has already launched, and Granite Rapids will launch in the third quarter. Both are built on the Intel 3 process, and the next-gen versions of each will move to the upcoming Intel 18A process.
The second problem for Intel is that data center customers are focused on building out their AI footprints. Much of the spending in the data center is going toward AI accelerators, where Intel has a small presence, rather than CPUs.
AI servers still need CPUs, and Intel is having success positioning its Xeon server chips as the "head node" used to manage AI server clusters. However, that's not enough to drive meaningful growth.
The company does expect the situation to improve in the second half of the year as the traditional server market rebounds, but only modestly.
3. The foundry strategy is on track
Despite the layoffs and cost cuts, Intel remains on track to complete its rollout of five nodes in four years, key to its plan to grow into the world's second-largest foundry by 2030.
The foundry business has started to generate some external revenue from advanced packaging services, although this number is currently small. Once the Intel 18A process is ready in early 2025 and ramps up in volume in 2026, external foundry revenue will grow in earnest. During the earnings call, CEO Pat Gelsinger reiterated that 2027 is still a good target for when the foundry business is expected to break even.
Getting its advanced processes up and running is crucial for Intel's product segment as well. Manufacturing of Lunar Lake, the company's upcoming laptop CPUs, is outsourced to Taiwan Semiconductor Manufacturing, leading to higher costs. Much of Intel's existing CPU lineup uses its older nodes, which have "uncompetitive cost structure and power performance," according to Zinsner.
Intel remains confident that its target of producing $15 billion in annual external foundry revenue by 2030 is still achievable despite the cost cuts.
A painful period for Intel
Turnarounds aren't easy, especially for a sprawling company like Intel. The foundry strategy is still a few years away from really paying off, and in the meantime, demand for the company's core products remains muted.
The good news is that Intel has a solid product line through the end of 2025. On the PC side, Lunar Lake for laptops and Arrow Lake for desktops arrive this year, and the Intel 18A-built Panther Lake comes in 2025.
On the server side, after Sierra Forest and Granite Rapids this year comes Clearwater Forest next year and an unnamed successor to Granite Rapids. As the company moves chips onto its Intel 18A process, competitiveness should increase, and the cost structure should improve.
Intel stock is not for the faint of heart. The company's overall strategy still looks promising, but the path forward is riddled with potholes, land mines, and brick walls. Expect the stock to be volatile as Intel scrambles to turn itself around over the next few years.
