Google and Intel announced an expanded multiyear partnership on Thursday, extending a collaboration that began in 2021. The stated goal is for Google Cloud to continue using Intel AI infrastructure and for both companies to continue developing processors together, including custom hardware aimed at accelerating data center workloads.
What’s changing in the partnership
According to the announcement reported by TechCrunch, Google Cloud will continue using Intel’s Xeon processors for AI, cloud, and inference tasks. The update includes use of Intel’s latest Xeon 6 chips. Google has used Intel’s Xeon processors for decades.
The companies will also expand their co-development of custom infrastructure processing units (IPUs). IPUs are chips that help accelerate and manage data center tasks by offloading them from CPUs. The IPU co-development work, which started in 2021, focuses on custom ASIC-based IPUs.
Intel declined to share information regarding pricing for the deal. The exact cost structure and commercial terms remain undisclosed.
Why CPUs and IPUs matter for AI workloads
The partnership reflects a broader infrastructure reality: the industry is experiencing a growing demand for CPUs. While GPUs are widely used for developing and training AI models, CPUs are crucial for running AI models and supporting general AI infrastructure.
IPUs are positioned as a way to improve how data center tasks are handled. By offloading certain workloads from CPUs, IPUs can reduce CPU bottlenecks and change how systems are balanced across different compute components. The approach involves accelerating and managing data center tasks through specialized silicon.
Intel CEO Lip-Bu Tan said in a company press release: “AI is reshaping how infrastructure is built and scaled,” adding that scaling AI requires “more than accelerators” and that “CPUs and IPUs are central to delivering the performance, efficiency and flexibility modern AI workloads demand.”
Chip shortages and the broader CPU push
The timing of the expanded partnership aligns with supply pressure. More companies have been focusing on CPUs in recent months due to a growing shortage of chips. SoftBank-owned Arm Holdings recently announced the Arm AGI CPU, described as the first chip Arm produced itself, amid a worldwide shortage for CPUs.
This suggests that AI infrastructure planning is increasingly shaped by availability as much as by architectural preferences. If CPUs are scarce, providers may need closer alignment between software workloads and the processors available for inference and infrastructure operations.
The partnership reflects a multiyear hardware strategy: continued Xeon usage for AI and inference, plus expanded work on custom ASIC-based IPUs. The industry may see how such custom accelerators and infrastructure chips fit into future AI deployment stacks, particularly if CPU supply constraints persist.
What to watch next
The structure of the announcement points to a system-level approach rather than a single-chip focus. Intel’s statement emphasizes “balanced systems,” and the deal expands both general-purpose CPU capacity and specialized offload hardware for data center tasks.
The industry could see more AI infrastructure designs that assume CPUs will remain central for inference and orchestration, while IPUs handle additional layers of data center processing. This could represent a shift in how vendors evaluate performance and efficiency tradeoffs across CPU and custom processing units. However, the available reporting does not include benchmark results, timelines for IPU availability, or details on how Google Cloud will integrate these custom units into its existing infrastructure.
What is clear from the reported material is that the Google-Intel relationship is deepening in both processor usage (Xeon, including Xeon 6) and custom infrastructure silicon (ASIC-based IPUs). Intel declined to disclose pricing, so the commercial impact cannot be assessed from the available details alone.
Source: TechCrunch