Intel’s AI Strategy & Q3 Earnings: NVIDIA Deal, Foundry Growth

Intel’s management spent much of the earnings call detailing how the company is repositioning itself within the AI hardware landscape, with CEO Lip-Bu Tan describing AI as a fundamental driver of Intel’s next growth phase. The most significant strategic development is Intel’s deepening partnership with NVIDIA, which will link Intel CPUs to NVIDIA accelerators using NVLink technology. Tan characterized this collaboration as creating ‘a new class of products spanning multiple generations’ aimed at hyperscale, enterprise, and consumer markets.

This partnership signals a fundamental shift in how Intel defines its role in the AI economy. Rather than competing directly in AI training—where NVIDIA maintains clear leadership—Intel is focusing on AI inference, particularly in what Tan called ‘agentic’ and ‘physical AI’ workloads. CFO David Zinsner emphasized that the collaboration expands Intel’s total addressable market rather than replacing existing business lines, suggesting Intel is positioning itself as a complementary player in the AI infrastructure stack while leveraging its installed x86 base.

Intel’s foundry segment emerged as a clear growth story in Q3, with tangible progress reported on its 18A process node now fully operational at Fab 52 in Arizona. Yields are improving at a predictable pace, while development on the next-generation 14A node is underway with early customer engagement reported across multiple fronts. The foundry narrative gained significant credibility through external validation during the quarter.

Financially, Intel secured $5.7 billion from the U.S. government, $2 billion from SoftBank, and expects an additional $5 billion investment from NVIDIA by the end of Q4. These substantial inflows not only strengthen Intel’s liquidity but demonstrate growing confidence from both private and public partners. Tan emphasized that Intel Foundry is being managed as a ‘long-term trust business,’ prioritizing customer relationships and disciplined investment—a cultural shift from previous cycles of aggressive, preemptive capital spending.

As demand for AI-related semiconductor manufacturing accelerates, Intel’s progress on advanced nodes and packaging technologies—including EMIB and EMIB-T—positions the company to compete for a growing share of outsourced wafer production. The company reiterated that capacity expansion will only occur in response to confirmed external demand, reflecting a more measured approach to capital allocation.

Intel’s financial results showed evidence of stabilization after several challenging years, with the company achieving 40 percent gross margin—four points above guidance—due to stronger product mix and cost control. Operating cash flow of $2.5 billion and positive free cash flow of $900 million marked significant improvements from prior quarters, though the stock remains down 20% from five years ago.

Zinsner highlighted that Intel raised approximately $20 billion in liquidity during the quarter, exiting with $30.9 billion in cash and short-term investments. The company used part of this to reduce debt by $4.3 billion and plans to continue deleveraging in 2026. Capital expenditures remain targeted at about $18 billion for 2025, with focus on foundry readiness and advanced packaging capabilities.

Management acknowledged ongoing structural headwinds, including high costs associated with older manufacturing nodes and early-stage yields on 18A that continue to weigh on margins. These factors are expected to normalize over the next 12 to 18 months as production efficiency improves. Product execution remains central to the recovery, with Intel reaffirming that Panther Lake will launch by end of 2025, followed by Nova Lake and Granite Rapids in 2026. The Core Ultra 3 rollout is expected to drive PC market recovery to approximately 290 million units for the year.