Semiconductor giant Intel (INTC +4.36%) is in a precarious position. The troubled chip designer and manufacturer has lost considerable market share to AMD in the core PC and server central processing unit (CPU) markets; its foundry business has few external customers despite massive investments to catch up with TSMC; and new CEO Lip-Bu Tan is shaking up the company with layoffs and strategy shifts.
Under pressure from the Trump administration, Intel agreed to a deal that gave the U.S. government a sizable stake in the company in exchange for funds that were already awarded, although mostly not yet delivered, via the CHIPS Act and other initiatives. While this arrangement could benefit Intel by incentivizing chip designers to consider Intel for manufacturing, it carries significant risks as well.
The next year for Intel will be critical as the company attempts to regain lost market share and turn its foundry business into a sustainable endeavor. Here's what the company needs to do to get its stock back on track.
Image source: Getty Images.
Fixing the PC business
Intel is still the leader in the PC CPU business, but AMD has greatly eroded that lead over the past decade. In the desktop CPU market, Intel's unit share has tumbled from more than 90% in 2016 to less than 70% today. A similar story has played out in the laptop CPU market, although the share losses have been less dramatic.
Intel struggled with manufacturing delays in the past, forcing it to use aging process nodes, while AMD hopped to superior technology from TSMC. Intel squeezed out as much performance as it could from its Alder Lake and Raptor Lake chips, but those chips were also plagued with instability issues.
Arrow Lake, which ended up using an advanced manufacturing process from TSMC rather than an in-house process, was a mixed bag. On the laptop side, Arrow Lake chips produced impressive performance and garnered positive reviews. On the desktop side, poor and inconsistent gaming performance overshadowed gains in productivity performance and efficiency.
Intel has two upcoming products that should put the company in a better position. Panther Lake, which will be focused on laptops, is set to launch by the end of the year with limited SKUs and additional SKUs coming in 2026. Nova Lake will follow in 2026 and be aimed at the desktop market. Both product lines will use the Intel 18A process, which should deliver meaningful improvements in performance and efficiency while eliminating the issues that plagued Arrow Lake.
Intel will need both Panther Lake and Nova Lake to succeed in order to start clawing back market share from AMD, and that means the Intel 18A process needs to deliver. The company has reportedly been having problems with yields which, if not resolved, could hurt the profitability of both products. Intel will need to get yields to acceptable levels -- and fast.
NASDAQ: INTC
Key Data Points
Making a comeback in servers
Intel's server CPU business has also been hurt by its manufacturing delays. The company fell way behind AMD in terms of core counts, performance, and efficiency, enabling its competitor to push its market share from close to zero a decade ago to almost 30% today.
Intel has caught up to a degree with Granite Rapids, which uses its Intel 3 process, although CFO Dave Zinsner noted during a recent conference that the company is still not competitive in some areas of the market. Diamond Rapids is expected sometime in 2026, using the same Intel 18A process that Intel's upcoming PC lineup will utilize. Again, profitability will depend on yields, as well as how well the chip performs compared to AMD's products.
Winning foundry customers
Intel has won some external business for its semiconductor foundry, but it hasn't yet scored a major customer for Intel 18A. Intel 18A will be a long-lived node for Intel, and Zinsner noted that there are still opportunities to win external customers for the node despite early struggles.
Intel 14A comes next, with products rolling out in 2028 and 2029. Intel 14A will be the first in the industry to use High-NA EUV tools, which make the process more capital intensive to scale and more expensive to produce wafers compared to Intel 18A.
While Intel will make plenty of its own products using Intel 14A, the company absolutely needs to win significant external customers for the process. Eventually, Intel won't have enough internal volume to justify the staggering expense of keeping up with TSMC in manufacturing. That day is likely nearing, especially given the company's market share declines, so external customers are the key to ensuring Intel's manufacturing operations are sustainable.
Intel will need to secure customer wins for Intel 14A in the lead-up to its launch. For the foundry business to work, the company must book meaningful external volume for Intel 14A throughout 2026 and 2027. If it doesn't, Intel may have little choice but to split itself up.
