Nvidia's $5B Intel bet: chipmakers join forces on AI infrastructure

Nvidia is investing $5B in Intel's common stock at $23.28 per share, marking a strategic collaboration between the GPU giant and the legacy chipmaker. Under the deal, Intel will design and manufacture custom x86 CPUs for Nvidia's AI infrastructure platforms and develop system-on-chips that incorporate Nvidia RTX GPU chiplets for personal computing. The partnership aims to connect Nvidia's accelerated computing capabilities with Intel's x86 architecture using Nvidia's NVLink technology.

The deal follows growing interest in Intel's CPU offerings for AI workloads — a shift from the GPU-dominated AI trade. The US government already holds a significant stake in Intel, having invested $8.9B at $20.47 per share in August 2025. Intel's market capitalisation has since climbed to $172.67B, reflecting a recovery from earlier financial struggles. The company has also secured partnerships with Tesla and Google, reinforcing its comeback narrative.

Intel's turnaround is far from guaranteed. The company has spent years losing ground to AMD and Nvidia in high-performance computing. Integrating two fundamentally different chip architectures — Nvidia's GPUs and Intel's CPUs — into seamless AI platforms is technically complex. And at $23.28 per share, Nvidia is paying a premium over the US government's entry price, raising questions about valuation discipline if Intel's recovery stalls.

Nvidia's role as a corporate investor has historically centred on strategic bets that extend its platform reach — and a $5B equity stake in a direct competitor underscores how critical CPU-GPU integration has become for the next phase of AI infrastructure. With both the US government and Nvidia now holding significant Intel positions, the chipmaker's comeback has effectively been underwritten by two of the most powerful actors in American technology, raising the stakes — and the scrutiny — on Intel's ability to deliver.