Back to Devoured - August 17, 2026
DEVOURED
Google reportedly taps AMD to design next-generation TPU — hybrid AI ASIC could integrate on-package CPU cores for reinforcement learning

Google reportedly taps AMD to design next-generation TPU — hybrid AI ASIC could integrate on-package CPU cores for reinforcement learning

AI infrastructurehardware Tom's Hardware
Google is reportedly partnering with AMD to develop a 10th-generation TPU that integrates x86 CPU cores directly on-package for reinforcement learning workloads.
What: Reports from SemiAnalysis suggest Google is shifting its custom TPU architecture to include general-purpose CPU compute via AMD, mirroring the design approach of AMD's Instinct MI300A. The shift responds to the growing demand for higher CPU-to-accelerator ratios, as seen in the transition from TPU v7 systems—which used one Xeon per four TPUs—to TPU v8 systems, which utilize one Google Axion CPU per two TPUs.
Why it matters: AI development is becoming increasingly reliant on reinforcement learning and reasoning agents, which require significantly more general-purpose compute than standard LLM training. Embedding CPU cores directly into the accelerator package reduces data latency and power consumption, marking a move toward specialized, hybrid silicon designed for specific algorithmic workflows.
Deep dive
  • Google’s 10th-generation TPU (v10) may pivot to include on-package x86 CPU cores.
  • AMD is a candidate to provide CPU IP and advanced packaging expertise (SoIC) for this hybrid design.
  • The design targets reinforcement learning and agentic AI models that are more CPU-intensive than typical transformer-based training.
  • Google has already increased its CPU-to-TPU ratio from 1:4 (Xeon to TPU v7) to 1:2 (Axion to TPU v8).
  • Integrating general-purpose cores with tensor accelerators reduces the physical distance between compute units, improving performance for complex reasoning loops.
  • This project would represent AMD’s first significant involvement in custom AI ASIC development for a cloud hyperscaler.
  • Broadcom remains Google’s primary silicon design partner for previous TPU generations, suggesting this specific hybrid project might require AMD's unique x86/accelerator packaging capabilities.
Decoder
  • TPU (Tensor Processing Unit): A custom ASIC (Application-Specific Integrated Circuit) designed by Google specifically for machine learning and neural network training.
  • Reinforcement Learning (RL): A training method where an AI agent learns to make decisions by performing actions in an environment to achieve a reward, requiring frequent CPU-based state calculations.
  • SoIC (System on Integrated Chips): A 3D advanced packaging technology that allows stacking different chiplets together for tighter integration and lower latency.
  • Chiplet: A small integrated circuit that contains a specific subset of functionality, meant to be combined with other chiplets in a single package to create a larger, complex processor.
  • Agentic Model: AI systems capable of executing multi-step tasks by reasoning and interacting with software tools, often requiring significant general-purpose compute beyond simple matrix multiplication.
Original article

Google has teamed up with AMD to develop one of its 10th-generation TPUs, according to a note by SemiAnalysis (via Sean). Analysts at SemiAnalysis believe Google may be interested in AMD's CPU cores for CPU-heavy workloads. If accurate, the collaboration would mark AMD's first major involvement in a custom AI ASIC project and could indicate that Google is exploring a new kind of TPU that combines its proprietary accelerator technology with on-board general-purpose cores for CPU-heavy workloads.

Market chatter suggests [Google] is working with AMD on a TPU project in the v10 generation," a SemiAnalysis note for clients cited by Sean reads. "AMD's involvement would be the first real involvement in a custom AI ASIC project, despite having a custom silicon team. AMD has strong IP, especially in advanced packaging and SoIC. Additionally, CPU IP could also be a draw given Google and its customers are pushing for TPUs with on-package CPU cores for RL workloads.

Having developed nine generations of its proprietary AI accelerators (with Broadcom acting as actual silicon designer) and possessing extensive expertise in accelerator architecture, Google hardly needs AMD to design a conventional TPU. Hence, chances that AMD will implement Google's TPU v10i for inference or V10t for training are low. Hence, Google might need something only a CPU maker like AMD could provide, including CPU IP, programmable logic, interconnects, or certain advanced packaging know-how.

Of these, the CPU angle is particularly noteworthy. SemiAnalysis claims that Google and its customers are pushing for TPUs with on-package CPU cores for reinforcement learning and potentially other CPU-heavy workloads. While conventional LLM training remains overwhelmingly accelerator-heavy, reinforcement learning for reasoning and agentic models can require considerably more general-purpose compute around accelerator operations.

Google has already begun to increase CPU resources around its latest inference-oriented TPUs. Its TPU 8i systems, designed for inference, reasoning, and RL workloads, feature one Google Axion CPU for every two TPUs. By contrast, servers running Google's 7th Generation TPUs used one Xeon 'Emerald Rapid' processor for every four TPUs. Furthermore, we are hearing that in some cases a 1:1 ratio of CPUs to accelerators is optimal, so the future of AI may be way more CPU-heavy than we think.

Meanwhile, bringing CPU cores directly into the TPU package could be a logical next step, as reducing the distance between general-purpose and tensor compute can improve performance and reduce power consumption. This is where AMD comes into play, as it already has experience developing a data center-grade design — the Instinct MI300A — that packs both x86 and accelerator chiplets. A hypothetical Google design could therefore combine Google-developed TPU compute chiplets with AMD CPU and HBM in a tightly integrated package built by AMD. Perhaps, Intel would appear as another potential candidate given that Google and Intel have multiple strategic collaborations. Yet Intel has no experience building hybrid x86+accelerator data center designs.

Note that for now we are speculating and our analysis may be inaccurate. For now, the nature of AMD's alleged involvement remains unclear. Yet, if the report is indeed accurate, the important development may not be that AMD is helping Google build another TPU. Instead, what matters is that Google is considering a new CPU-heavy member of its TPU v10 family, optimized specifically for RL and agentic workloads, and is using AMD as a provider of some of the building blocks needed to create it.

Explore all 72 articles in today's digest

Fresh dev news, AI-digested daily. Never miss a trend at devoured.fyi