TPUv7 Ironwood posts first long-context AgentX results
SemiAnalysis published the first official long-context, multi-turn AgentX results for Google’s TPUv7 Ironwood across Muse Spark Open, GLM-5.2, Qwen3.5, and Kimi K3. The report says Ironwood was competitive or achieved the best performance-per-dollar result in these submissions, including 1.2x to 2.7x the tokens per dollar of NVIDIA B200, B300, and AMD MI355X for GLM-5.2 at FP8. Results varied by precision and workload: NVIDIA’s FP4 offerings led some comparisons, while Rubin figures remain preview submissions that SemiAnalysis could not independently reproduce.
The benchmark results offer buyers new evidence about TPUv7’s cost efficiency for realistic agentic inference and show that Google’s TPU efforts extend beyond hardware into external customers and open-source serving integrations.
In November 2025, we published a newsletter saying that Google is selling TPUs to external customers. Since then, we have been proven right over and over again with Oracle, OpenAI, Anthropic, and Meta all becoming customers of TPUs just as our Accelerator Model reported.
Hardware demand is only half the story, as NVIDIA’s moat has historically been its software and large developer ecosystem. This is where TPU is also making headway, partnering with upstream maintainers across the entire ecosystem, including the Meta PyTorch team, vLLM (Inferact, RedHat), SGLang (RadixArk), llm-d, Ray, etc. This is not a surprise to folks that have read our September 2026 article laying out Google TPU’s externalization strategy.
Back in our September article, we published the first third-party TPUv7 results on InferenceX on single-turn workloads, showing that in an apples-to-apples FP8 comparison on Qwen 3.5 397B, TPUv7 Ironwood) delivered up to 50% better performance per dollar than B200 and B300.
We also reported that many Google TPU customers were asking for AgentX results, since long-context, multi-turn agentic serving is where inference spend is heading.
#Google#TPUv7 Ironwood#InferenceX#AgentX#NVIDIA#AMD#Muse Spark Open#GLM-5.2#Qwen3.5#Kimi K3#vLLM#SGLang