Purdue sets January access target for Google AI chips on campus

Purdue University is bringing Google’s custom AI chips into its campus research computing program, with initial researcher access expected in January 2027. Announced October 29, the collaboration with Google Public Sector, Cirrascale and DDN will increase Purdue’s AI computing capacity by 30%, the university says, while keeping institutional control over data, security and governance.

From promised access to a campus installation

The project follows a broader Purdue–Google Public Sector partnership announced January 12, 2026. That agreement already included access to Tensor Processing Units, or TPUs—Google’s custom processors for AI computing. The October announcement adds concrete deployment details: the campus location, an expected capacity increase and a target for the first researchers to begin using the system.

Purdue announced the deployment collaboration at AI Impact: From Possible to Proven, the third annual AI summit co-sponsored by the university and Google. The summit focused on moving physical AI from experiments into real-world use. That field applies AI to machines and systems that interact with their surroundings, including robotics, autonomous vehicles, computer vision and advanced manufacturing.

The university says it will be the first U.S. academic institution to provide this custom AI hardware to researchers through its campus computing program.

Working with Google Public Sector, Cirrascale and DDN to integrate TPUs within the framework of our long-running community cluster high-performance computing program gives us the performance we need for ambitious training and inference work, while operating it within our existing business.

Preston Smith, executive director of Purdue’s Rosen Center for Advanced Computing

An addition to Purdue’s existing research systems

The TPUs will be integrated into Purdue’s newly renovated and expanded Mathematical Sciences Building Data Center. Faculty and students are the intended users, with the hardware supporting research that requires substantial AI computing resources.

Smith described two intended uses: training AI models and inference, the work of running trained models. He also framed access as an educational benefit, saying that experience with advanced computing resources would help students develop AI skills for the workforce.

The university says it will retain institutional control over data, security and governance as it integrates the Google accelerators.

The projected 30% increase describes Purdue’s overall AI computing capacity, not a measured speed improvement for an individual research task.