OpenAI's new chip outperforms Nvidia on efficiency but won't stop buying Nvidia.
The chip is for inference only.
Nvidia's Q2 report comes tomorrow.
What happened
On August 25, 2026, OpenAI unveiled Jalapeño, its first custom inference chip, claiming 1.5–1.9x more AI work per watt and 1.7–3.6x lower latency than Nvidia's GB200 and GB300 on the InferenceX benchmark.
Why it matters
Despite the headline-grabbing result, OpenAI explicitly stated it will continue to buy Nvidia accelerators for both training and inference, indicating this is a cost-saving move rather than a full supplier replacement.
What to watch
OpenAI plans to deploy Jalapeño in its own infrastructure by end of 2026, but Nvidia's Q2 results due August 26 will shed light on broader demand beyond OpenAI.
Ask the AI about this article →
OpenAI's announcement of Jalapeño is a notable engineering feat, but the company's simultaneous commitment to Nvidia underscores the complexity of the AI hardware landscape. While the chip shows impressive efficiency gains on public benchmarks, it is designed only for inference, not training. This suggests OpenAI is optimizing costs for its own workloads rather than seeking to replace Nvidia entirely. Nvidia's data center revenue has been strong, growing 92% year over year in Q1 FY2027, and its supply commitments remain high. The real threat to Nvidia would be if external customers start buying Jalapeño or if a training-capable version emerges, but neither is on the horizon. Investors should watch Nvidia's Q2 results for signs of broader demand.
AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.
Free · takes 30 seconds · unsubscribe anytime
Ask AI anything about this article. Q&As are published on this page for other readers too.
The Linux Foundation announced the contribution of TRACE (Trust, Runtime Attestation and Compliance Evidence)…
A new Stanford research paper warns that AI chatbots and agents may let advertising dollars influence what the…
The Wall Street Journal published an op-ed by billionaire investor Stanley Druckenmiller that was written with…
OpenAI unveiled benchmarks for its Jalapeño inference chip on Tuesday, showing it competes with Nvidia’s best…
Researchers from Bytedance Seed developed EdgeBench, a benchmark measuring how AIs improve on tasks over multi…
Lovable introduced a vision where apps become 'capabilities' that AI agents can call directly, bypassing the h…
