OpenAI reports its custom AI silicon beats Nvidia processors in tests, threatening Nvidia's AI chip leadership and pressuring semiconductor ETFs.
The article states OpenAI claims its new chips outperform Nvidia processors in tests, directly threatening Nvidia's position as the leading AI accelerator supplier and potential revenue from a key AI lab customer.
NVDA may slip as investors price in increased competition from OpenAI's custom silicon, though the lack of third-party benchmarks could limit the downside.
OpenAI is widely reported to rely on Nvidia GPUs for training and inference. The new chip claim suggests OpenAI aims to reduce that dependence.
OpenAI is a major AI lab and a large consumer of compute; losing its business to in-house silicon would cut into Nvidia's data-center demand.
SMH holds Nvidia as one of its top holdings. If NVDA trades lower on the OpenAI chip claim, the ETF likely declines due to direct exposure.
SMH is heavily weighted to Nvidia, so a decline in NVDA shares typically pulls the ETF lower. The drag depends on Nvidia's current weighting and offsetting moves in other holdings.
The news is based on internal OpenAI tests and lacks independent verification. Investors should monitor Nvidia's stock reaction and any official benchmark releases before making changes.
OpenAI said its custom AI chips outperformed Nvidia processors in internal tests, according to a Bloomberg report. The claim challenges Nvidia's lock on AI accelerator demand from major cloud and AI labs. If OpenAI shifts workloads to in-house silicon, Nvidia's data-center revenue faces a direct competitive threat. Investors await independent benchmark data to verify the results.
OpenAI said its new custom chips outperformed Nvidia processors in tests, according to a Bloomberg report on August 25, 2026.
Nvidia dominates the AI accelerator market and counts OpenAI among major customers. A credible claim of superior in-house silicon could pressure Nvidia's demand and valuation.
The article reports OpenAI's internal test claims. Independent benchmark data or third-party validation was not provided.