Intel After Q3: Can ‘IDM 2.0’ Execution, Foundry Ambitions and AI Accelerators Reignite Revenue and Margin Traction?

Intel reported third-quarter revenue of $13.65 billion, beating consensus of $13.14 billion, with adjusted EPS of $0.23. The stock rose roughly 6% after-hours and more than 7% in premarket trading following the report, reflecting relief around a PC recovery and stable core execution. Management guided to fourth-quarter revenue of about $13.3 billion at the midpoint and adjusted EPS of $0.08, broadly in line with street expectations.

The quarter’s headline EPS was muddied by a unique accounting treatment tied to the U.S. government’s $8.9 billion investment negotiated in August. Intel recorded a GAAP per-share loss of $0.37 to account for shares in escrow earmarked for release to the government, while noting it has consulted with the SEC and may revise results after the shutdown. Importantly, Intel disclosed it received $5.7 billion in cash during the quarter related to this transaction, supporting near-term liquidity and capital deployment flexibility.

By segment, the products group generated $12.7 billion in sales, up 3% year over year. Client Computing contributed $8.5 billion, reflecting a firmer PC market backdrop, while Data Center and AI CPUs posted $4.1 billion, down 1% year over year. The company said demand outpaced supply and expects that dynamic to persist into next year—helpful for mix and pricing, but also a reminder of the supply-chain choreography required to fully capture AI-led growth. The Q4 guide excludes the recently completed Altera divestiture, which has implications for both revenue presentation and reinvestment priorities in core and foundry operations.

Intel Foundry remains the strategic centerpiece—and the heaviest lift. The division posted $4.2 billion in revenue in Q3, down 2% year over year, and crucially all of it came from internal wafers. That underscores the work ahead to secure anchor external customers at advanced nodes. The company started production of its most advanced 18A chips in Arizona during the quarter, a milestone necessary to rebuild process credibility and reassure potential partners about roadmap reliability and yield.

Scale is the next hurdle. Management has signaled roughly $100 billion of capital needs to stand up a globally competitive foundry footprint aligned with national security and supply-chain resiliency imperatives. The U.S. government has effectively joined as a strategic partner—backed by cash inflows in Q3 and an equity stake mechanism—in support of domestic leading-edge manufacturing. This public-private model can de-risk some financing and geopolitical exposures, but it cannot substitute for process leadership or customer capture at competitive pricing.

Intel’s operating discipline is also in focus. Headcount stood at 88,400, down sharply from 124,100 year over year, pointing to cost control and a leaner operating model. The company must translate that leaner structure into better foundry cost competitiveness, cycle-time reductions, and customer service improvements. Without external 18A wins and a visible utilization ramp, fixed-cost absorption risks will keep foundry margins under pressure and postpone the flywheel effect that drives returns in leading-edge manufacturing.

The AI buildout cuts both ways for Intel: it is an immediate attach opportunity for its CPUs and a long-term accelerator roadmap challenge. The recently announced $5 billion investment and partnership with Nvidia puts Intel’s server CPUs alongside Nvidia’s AI GPUs in large-scale systems, strengthening Intel’s presence in AI compute nodes even as the accelerator slot is predominantly owned by Nvidia today. If attach rates remain high and platform standards favor balanced CPU-GPU architectures, Intel can benefit from AI capex cycles without owning the accelerator slot outright.

Management’s comment that demand outpaced supply in Q3—and is likely to continue—signals robust AI-driven pull across data centers and select PC segments adopting on-device AI features. The implication: where Intel can allocate supply to higher-mix SKUs, revenue growth and gross margin quality should improve. Sustained delivery on the 18A roadmap will be essential to competing for future AI accelerators and custom silicon, especially as hyperscalers consider heterogeneous compute and memory fabrics that reward packaging, interconnect, and power-efficiency leadership.

The accelerator pipeline will need to show more tangible traction in 2026+ to change consensus views. In the interim, the company’s best path to AI monetization is CPU attach and platform partnerships that put Intel in the critical path of AI systems procurement. The Nvidia tie-up is strategically notable because it can stabilize data center CPU volumes, improve visibility, and potentially support better pricing if platform bill-of-materials grow and customers prioritize qualified CPU-GPU combos for deployment speed.

Taiwan Semiconductor Manufacturing Co. posted another record quarter, with revenue up 30.3% year over year to roughly $33.1 billion and net income at NT$452.3 billion. Advanced nodes at 7nm and below represented 74% of wafer revenue, and high-performance computing, driven by AI and 5G, reached 57% of sales. Management raised full-year 2025 revenue growth guidance to the mid-30% range and lifted its capex floor to $40 billion, reflecting a multi-year AI-driven utilization surge.

This is the foundry bar Intel must clear to capture external share. TSMC’s scale, yield learning, and pricing power are reinforced by a customer roster that includes Nvidia and AMD for advanced AI processors. In that context, Intel Foundry’s lack of a marquee external 18A customer keeps its pricing and margin prospects uncertain. Even with U.S. onshoring incentives and potential tariff dynamics, policy tailwinds can only augment—never replace—competitive process leadership and time-to-market advantages.

For Intel, the practical benchmark is twofold: prove 18A reliability in production at a level that attracts a top-tier mobile or HPC design, and demonstrate enough capacity readiness to de-risk ramp schedules. If those conditions are met, Intel can start to leverage domestic manufacturing as a premium value proposition for security- and latency-sensitive workloads. Without them, utilization gaps can drag on foundry margins even as the product business benefits from AI-related demand.

Delivery versus Q4 guidance will set the tone for 2026 expectations. Investors should track whether the PC recovery remains intact through holiday shipments, how data center CPU demand trends against AI accelerator deployments, and whether mix shifts lift consolidated gross margin. Management’s comment that demand exceeds supply suggests allocation discipline—watch for indications that high-mix products get priority as capacity expands.

Foundry milestones are the swing factor. Concrete external wins on 18A—especially from a top-10 fabless or cloud customer—would be the clearest de-risking event for the IDM 2.0 story. Look for backlog disclosures, tool install updates, and utilization commentary that imply a trajectory toward positive foundry operating leverage. Meanwhile, clarity on the government stake accounting from the SEC will help normalize GAAP comparisons, and details around Altera divestiture proceeds and reinvestment use-cases will inform capital allocation and strategic focus.

Finally, AI product ramps and ecosystem adoption: monitor CPU attach in AI GPU systems, early signals on Intel’s accelerator traction, and the timeline for supply constraints easing. If supply bottlenecks persist but favor higher-value SKUs, near-term margin outcomes could surprise positively; if bottlenecks impede shipments, revenue cadence could remain choppy despite strong order books.

Base case: revenue stabilizes with mid-single-digit growth as PC demand normalizes and data center CPUs re-accelerate on AI attach, while gross margins grind higher on mix and cost discipline. Foundry remains mostly internal through 2025 with incremental external prototypes but limited revenue contribution. The government partnership and cash inflow help fund capex without undue balance sheet strain, but sustained free cash flow improvement hinges on utilization gains.

Upside case: execution on 18A yields one or more meaningful external design wins, providing line-of-sight to double-digit revenue growth exiting 2026 as foundry utilization improves. AI platform partnerships increase CPU dollars per node, while early accelerator traction and advanced packaging revenue contribute to margin quality. Price targets have been trending upward over the past quarter, and if current share prices near the high-$30s are sustained alongside better external foundry visibility, multiple expansion could follow.

Key risks: a slippage in process timing or yields that delays external ramps; the durability of the PC recovery; dependence on partners for AI system wins; and intense competitive pressure from TSMC’s capacity, ecosystem lock-in, and price-performance advantages. Policy tailwinds can mitigate geopolitics and financing risks but will not rescue margins without credible process leadership and customer capture.