TREE NEWS reports: OpenAI’s first self-developed ASIC chip, Jalapeño, has reportedly surpassed Nvidia’s GB300 and Blackwell in key performance metrics, according to tests by SemiAnalysis. The chip, developed with Broadcom, achieves higher throughput and lower latency while consuming less power, marking a significant milestone in AI hardware. However, experts note it is still in engineering sample stage, with mass production expected by 2027.
OpenAI’s In-House ASIC Chip ‘Jalapeño’ Outperforms Nvidia Blackwell in Tests
AI take
This signals a potential shift in AI infrastructure leverage, as a custom in-house chip could reduce dependency on dominant suppliers. The reported efficiency gains matter for scaling costs, but the engineering sample stage and 2027 timeline mean near-term market dynamics are unchanged. Watch whether performance holds through production and how suppliers respond competitively.
Generated by AI for reference only.
Share on WeChat
Open WeChat → Scan → then tap "…" to send to a chat or Moments.
Tap "…" in the top-right corner to send to a chat or share to Moments.
Related News
1h ago
Beijing centre launches space computing cloud platform covering 40+ satellites
2h ago
1.223M SOL Worth $117M Deposited to Coinbase Prime via Single Address
2h ago
Flap Says flap.sh Restored, Backup Domain flap.cool Stays Open
2h ago
Bitcoin tests short-term holder cost basis, shows resilience vs FTX crash: CryptoQuant
3h ago
ICE Deletes Trade.xyz Partnership Tweet, Keeps Retweet
3h ago
Anthropic S-1 Filing Lifts Vertiv 25%, Nvidia Just 2.1% in Three Sessions