Nvidia and AMD sales rise after OpenAI’s brief frontier AI training pause

Executive Briefing
“Nvidia and AMD both reported strong revenue growth in results cited by The Verge, following...”
Key Takeaways
- 1.OpenAI restarted the training run after 10 days, before its announced two-week pause ended, while continuing to hold back some smaller experimental runs. Anthropic released upgraded frontier models in November.
- 2.In figures cited by the article, Nvidia’s data-center revenue rose 110% year over year to $107.6 billion in the three months ended October 25.
- 3.AMD reported a 45% revenue increase to $13.4 billion in the quarter covered by the article. Data-center sales nearly doubled to $8.5 billion.
- 4.Counterpoint Research’s Neil Shah told The Verge that existing computing and memory procurement agreements could delay the financial effects of a training slowdown by about six months.
- 5.Shah argued that demand to run existing AI models could sustain investment in computing infrastructure even if frontier training slows. That is his forecast, not a conclusion established by the revenue figures above.
Nvidia and AMD both reported strong revenue growth in results cited by The Verge, following OpenAI’s brief pause in a frontier AI training run. Those results do not establish how a prolonged slowdown in training would affect chip demand.
Key Developments & Primary Facts
OpenAI restarted the training run after 10 days, before its announced two-week pause ended, while continuing to hold back some smaller experimental runs. Anthropic released upgraded frontier models in November.
In figures cited by the article, Nvidia’s data-center revenue rose 110% year over year to $107.6 billion in the three months ended October 25.
AMD reported a 45% revenue increase to $13.4 billion in the quarter covered by the article. Data-center sales nearly doubled to $8.5 billion.
Counterpoint Research’s Neil Shah told The Verge that existing computing and memory procurement agreements could delay the financial effects of a training slowdown by about six months.
Shah argued that demand to run existing AI models could sustain investment in computing infrastructure even if frontier training slows. That is his forecast, not a conclusion established by the revenue figures above.
Strategic Analysis & Implications
Current chip revenue can reflect earlier orders. Strong results and a short training pause can therefore coexist without resolving the longer-term effect.
If businesses devote more computing capacity to serving users, chip demand could shift from training toward running models rather than disappear. Later orders and revenue would be needed to measure that shift.
Outlook & Factors to Watch
It remains unclear whether AI labs will reduce frontier-model training for an extended period.
The supplied figures do not separate chip revenue from training and model-serving workloads, so they cannot show whether one is offsetting the other.







