Intel Joins Musk’s 1 TW Terafab Project and Seals AI Infrastructure Deal with Google
Intel has joined Elon Musk’s Terafab project to design and fabricate chips at a planned capacity of 1 TW per year—equivalent to 70% of TSMC’s current global output—bringing its scaling expertise to Musk’s mega‐fab initiative. The company also signed a multiyear deal with Google to integrate Xeon CPUs and infrastructure processing units for more efficient, large-scale AI and cloud workloads.
1. Terafab Partnership Details
Intel announced it will leverage its design, fabrication and packaging capabilities to support Elon Musk’s Terafab semiconductor project, aiming to produce 1 TW of compute power annually for AI, robotics, Tesla vehicles and space-based data centers. Once fully operational, the facility seeks to match 70% of TSMC’s current global output, marking a potential record in chip production scale.
2. Google AI Infrastructure Collaboration
Separately, Intel signed a multiyear agreement with Google to integrate its Xeon processors with custom infrastructure processing units, offloading networking, storage and security tasks from CPUs. This collaboration is intended to accelerate AI model training and inference by improving data center efficiency, power consumption and overall performance.
3. Strategic Implications for Intel
These partnerships reinforce Intel’s push to reclaim leadership in advanced manufacturing and AI infrastructure, bolstering its competitive position against TSMC, AMD and Nvidia. By combining fab expertise with cloud-scale AI optimization, Intel aims to drive new revenue streams and restore investor confidence in its data center roadmap.