AI Training Slowdown May Shift Chip Spending, Not Shrink Nvidia’s Market

November 18, 2026
AI Training Slowdown May Shift Chip Spending, Not Shrink Nvidia’s Market
  • Nvidia reported $107.6 billion in data center revenue for the three months ended October 25, up 110% from a year earlier, while its new Vera Rubin platform generated nearly $20 billion in its first quarter of production shipments.

  • The company expects major frontier labs to account for about 15% of its data center revenue this fiscal year and raised its revenue-growth forecast for next fiscal year from 70% to 75%.

  • A brief pause in frontier AI training sparked concerns that labs would cut chip spending, but the slowdown was short-lived: OpenAI resumed its large training run after 10 days, and Anthropic released upgraded frontier models in November.

  • So far, there is little evidence that the pause has reduced chip orders, with TSMC’s latest sales figures remaining strong; existing compute and memory procurement agreements could delay any financial impact by roughly six months.

  • AMD also posted a record quarter, with revenue up 45% to $13.4 billion and data center sales nearly doubling to $8.5 billion; OpenAI and Anthropic were among the first customers using its Helios systems with MI450 GPUs.

  • Meanwhile, Ramp data showed frontier model families made up about 60% of business token usage in early November, roughly three times their share at the start of the year, as falling token prices encouraged more usage even while corporate AI spending grew only modestly.

  • That shift points to rising inference demand—the computing required to run existing models—which could offset weaker spending on frontier training and may require even more total compute infrastructure.

  • At the same time, competition is driving AI companies toward more efficient, lower-cost computing, including custom chips; Counterpoint expects custom AI server processor shipments to exceed data center GPU shipments by 2028.

  • The takeaway is that a frontier slowdown may redirect chip spending rather than shrink the market, with cheaper custom chips posing a greater challenge to Nvidia’s dominance and margins than a broad decline in AI investment.

Summary based on 1 source


Get a daily email with more Tech stories

More Stories