Unverified50% confidenceReleaseExact time
Alibaba's T-Head announced the PPU (True 5V900) training and inference chip, with the company claiming performance reaches three times that of the previous generation; mass production and commercial sales are planned for Q1 2027.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/30/2026
First Seen
Valid until: 12/29/2026
Sources
Related Entities
Related Claims
UnverifiedFor projects involving 3–5 drones, PyBullet's serial simulation speed on CPU is sufficient to support the training requirements of on-policy algorithms such as PPO, and the acceleration advantage of GPU parallelization is not significant.67% similarUnverifiedGoogle's in-house TPU v5p was released in late 2023, optimized for large-scale matrix operations, and achieves high-speed interconnection of thousands of chips through ICI technology.67% similarUnverifiedPingtouge has released a new-generation AI chip, the Zhenwu V900, featuring a unified training-inference architecture, with performance approximately three times that of the previous generation Zhenwu M890, equipped with 216GB of video memory, scalable to 500,000 cards in a single cluster, and planned for mass production in Q1 2027.66% similarUnverifiedGrok 4.5 is based on the all-new V9 base model, with parameters reaching 1.5 trillion—three times those of the previous-generation V8—and was trained using tens of thousands of NVIDIA GB300 GPUs.63% similarUnverifiedNVIDIA H100 的 Transformer Engine 将 FP8 训练和推理的吞吐量相比上一代提升近 4 倍62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/961493API
curl https://kongchang.com/api/v1/knowledge/claims/961493MCP
get_claim(id=961493)