Intel invented the microprocessor and dominated PC and server chips for 40 years. Its Core processors are in 80%+ of Windows laptops. Its data center CPUs (Xeon) run enterprise servers worldwide. Intel's foundry ambitions - manufacturing chips for others - and its 18A process node are existential bets to reclaim leadership after losing ground to TSMC and AMD. Pat Gelsinger's turnaround is the highest-stakes story in semiconductors.
Key relationships in depth
Intel and Microsoft have a foundational partnership spanning decades: Intel processors power the majority of Windows PCs and Azure server infrastructure, and in 2025 Intel's foundry division reportedly secured contracts to manufacture Microsoft's Maia 2 custom AI processor on Intel's 18A process node. Intel and Microsoft also collaborate on AI PC features through Intel's Core Ultra NPU integration with Windows ML and Microsoft Copilot+ PC certification requirements.
Why it matters: Microsoft's custom silicon strategy (Maia, Cobalt) requires foundry manufacturing capacity, and Intel's 18A process node offers a US-based advanced manufacturing alternative to TSMC that satisfies domestic supply chain requirements for Microsoft's government cloud contracts. Intel benefits enormously from a Maia 2 tape-out validating its 18A process, which would be a critical commercial proof point for Intel Foundry's turnaround narrative.
NVIDIA disclosed a roughly 4% strategic equity stake in Intel, purchasing approximately $5B worth of shares. This positions NVIDIA as a significant Intel shareholder at a time when Intel is undergoing major restructuring under CEO Lip-Bu Tan, spinning off its foundry business and refocusing on chip design.
Why it matters: A viable Intel foundry would reduce NVIDIA's dependence on TSMC for the most critical chips in the world and serves as geopolitical insurance given Taiwan-China tensions. NVIDIA benefits from Intel's success as a manufacturer without needing to operate a fab itself.
Intel and Google significantly expanded their chip partnership in April 2026, with Google committing to use multiple generations of Intel Xeon processors across Google Cloud infrastructure and expanding co-development of custom Infrastructure Processing Units (IPUs). The multiyear agreement covers Intel's latest Xeon 6 processors for Google Cloud compute and extends the two companies' joint work on IPUs that handle networking and storage offload tasks, freeing GPU and CPU resources for AI workloads.
Why it matters: Google's AI infrastructure requires specialized processing at every layer, and Intel's IPUs (developed jointly since 2022) offload network packet processing and storage I/O from general-purpose CPUs, improving the efficiency of Google's AI factory architecture. Intel gains Google as a multi-generation committed customer for both standard Xeon CPUs and custom silicon, providing revenue visibility during Intel's foundry and product transition.
Intel and Amazon have a major cloud and foundry partnership: Amazon uses Intel Xeon processors across a significant portion of AWS EC2 general-purpose and compute-optimized instances, and Intel's advanced packaging technology (EMIB-T) is in active discussions with Amazon for custom AI chip packaging. Amazon is reportedly in talks with Intel Foundry to leverage Intel's EMIB-T advanced packaging for Amazon's Trainium and Inferentia custom AI chips.
Why it matters: Amazon's Trainium 2 AI training chip requires advanced multi-die packaging that TSMC's CoWoS provides, but Intel's EMIB-T packaging could offer Amazon a US-based alternative with competitive latency and bandwidth for memory integration. Amazon benefits from supply chain diversification for its custom silicon program while Intel gains a landmark cloud AI chip packaging customer that validates its advanced packaging business.
Relationships are compiled by VexMap from company filings, announcements and reporting, and last reviewed September 2026. Spotted something wrong? Tell us. Not investment advice.