Intel’s comeback now depends on 14A
- Intel still makes most of its money from PC and server processors.
- The big swing is Intel Foundry, which aims to make chips for outside customers.
- Management says unfilled demand is in the billions, which points to real product demand but also lost sales.
- The key test is whether a major outside customer commits to Intel 14A in the 2H 2026 to 1H 2027 window.
- The stock needs execution to catch up with the story, since the turnaround still carries heavy factory and balance sheet risk.
The 14A clock is ticking
Intel is a turnaround with one main question: can it prove that its factories matter again? The company still has a large x86 chip business in PCs and servers, but the bigger value case depends on Intel Foundry winning outside customers.
The most important watch item is Intel 14A, the next planned leading-edge chip process. A process is the recipe and factory technology used to make a chip. Intel has said it may pause or stop future leading-edge work if it cannot secure a significant external customer for 14A.
The bull case is that the warning creates urgency, not panic. Management now expects design commitments to start in the second half of 2026 and expand into the first half of 2027. The new Terafab partnership with SpaceX, xAI, and Tesla could become a major outside foundry relationship if it turns into real financial commitments.
The bear case is harsher. If the 14A window passes without a normal anchor customer, Intel may have to slow or stop the very factory roadmap that supports IDM 2.0. That could mean write-downs, weaker access to government incentives, and more reliance on TSMC for Intel’s own products.
Chips today, foundry tomorrow
Intel is an integrated device maker. That means it designs many of its own chips and also runs factories that make them. Its main cash source is selling x86 processors for laptops, desktops, and servers.
The newer model is IDM 2.0. Intel uses outside foundries such as TSMC for some products when that helps performance or supply. At the same time, Intel Foundry tries to sell manufacturing and packaging services to other chip designers, putting Intel in direct competition with TSMC and Samsung.
This model can work well if factory volume is high. Chip plants are expensive, so empty or underused capacity hurts margins fast. That is why outside foundry customers matter so much for 14A.
Demand is not the only problem. Management said Q1 2026 unfulfilled demand was meaningful and started with a B, meaning billions of dollars. That is a better problem than weak demand, but it still caps revenue and risks share loss if customers move to AMD, ARM-based chips, or other suppliers.
Where Intel fights
Client Computing Group
This is the PC chip business, including Core processors for laptops and desktops. Core Ultra is Intel’s AI PC line, with Core Ultra Series 3 built on Intel 18A.
Data Center and AI
This group sells Xeon server CPUs and AI-related products. Intel is leaning into inference and agentic AI, where CPUs can still play a large role next to accelerators.
Intel Foundry
Foundry is the long-term swing factor. It makes wafers, packaging, and related services for Intel’s own products and aims to win outside chip designers.
Custom ASICs
Intel builds purpose-made chips for AI, networking, and cloud customers. Management said this business reached an annualized revenue run rate of more than $1 billion in Q4 2025.
Altera and Mobileye
These are non-core holdings in FPGAs and driver-assistance technology. Intel continues to monetize stakes in these assets to simplify the company and raise cash.
NVIDIA-linked Xeon
Intel is developing a custom Xeon with NVIDIA NVLink technology. If it works, it could make Intel CPUs more useful inside AI systems built around NVIDIA accelerators.
The mix is still product-led
The mix uses fiscal 2025 disclosed segment revenue before intersegment eliminations: CCG $32.2 billion, DCAI $16.9 billion, Intel Foundry $17.8 billion, and All Other $3.6 billion. Foundry includes large internal sales to Intel Products, so its share does not mean outside foundry demand is already that large.
What can break the plan
No 14A anchor customer
High impact · Medium oddsIntel has said it may pause or discontinue future leading-edge process work if it does not secure a significant external customer for Intel 14A. That would strike at the core of IDM 2.0. It could also trigger asset impairments, weaker government support, and more dependence on third-party foundries.
Supply limits become share loss
High impact · Medium oddsManagement says demand is ahead of supply by billions of dollars. That shows customers want the products, but it also means Intel is leaving revenue on the table. If shortages last too long, customers may shift more volume to AMD, ARM suppliers, or in-house chips.
Gross margin stalls
High impact · Medium oddsIntel 18A products are still early in the ramp, which can mean higher costs per wafer. Management also pointed to rising input costs for items like memory and substrates. If margins do not improve as 18A matures, the turnaround becomes much harder to fund.
Terafab does not validate foundry
Medium impact · Medium oddsThe Terafab partnership with SpaceX, xAI, and Tesla is important, but its structure is still unclear. It may not look like a normal foundry customer deal. If it is a special arrangement, it may not prove that broad outside customers will choose Intel Foundry.
Capital plans depend on outside help
High impact · Medium oddsIntel has slowed some factory projects, including Ohio, to conserve cash and match spending to demand. Government incentives are tied to milestones and could be at risk if leading-edge plans change. Heavy capital needs keep financial health a central concern.
Leadership churn hurts execution
Medium impact · Medium oddsIntel has had CEO transitions in 2025, 2024, 2021, and 2019. A turnaround this complex needs steady choices over many years. More top-level change could slow customer talks, capital planning, and product roadmaps.
In one breath
What is Intel 14A and why does it matter?
Intel 14A is Intel’s planned next leading-edge manufacturing process after 18A. It matters because Intel has said future leading-edge work may be paused if it cannot win a significant outside customer for 14A.
Is Intel mostly an AI company now?
No. Intel sells into AI data centers, but its base business is still PC and server CPUs. The AI opportunity is mainly in Xeon CPUs, inference, custom chips, and possible NVIDIA-linked systems.
Why is Intel using TSMC if it owns factories?
IDM 2.0 allows Intel to use outside foundries for some products when that helps performance, timing, or supply. The risk is that if Intel’s own leading-edge roadmap slips, outside foundries could become a crutch instead of a tool.
What should investors watch over the next year?
The top item is a named Intel 14A external customer. After that, watch whether supply catches up to demand, whether gross margin improves after Q2 2026, and whether Terafab becomes a real funded project.