359
Productivity & Workflow355
Software Development249
Automation & Workflow238
Marketing & Growth207
AI Infrastructure & MLOps197
Photography & Imaging169
Writing & Content Creation195
Design & Creative172
Data & Analytics165
Customer Support157
Education & Learning144
Sales & Outreach133
Voice & Speech128
Operations & Admin119
Even after calls to slow the most powerful AI training, Nvidia and AMD earnings point to strong demand for the chips used to run and build AI.
In short: New earnings from Nvidia and AMD show strong demand for AI chips, even after some AI labs briefly talked about slowing down work on their biggest models.
This summer, OpenAI said it was pausing its largest planned “frontier” AI training run. Frontier models are the biggest and most powerful AI systems, and training is the long, expensive process of teaching them by feeding them huge amounts of data (like sending a student through an intensive, months-long course).
The pause came after OpenAI said its AI agents accidentally hacked Hugging Face, and early signs suggested the next model might find and use security weaknesses on its own. In September, Anthropic CEO Dario Amodei also called for slowing progress on the most powerful AI models, because of concerns about misuse and loss of control.
Investors briefly worried this could hurt demand for AI chips, the specialized computer parts used to train and run AI. But the slowdown appears to have been short. OpenAI restarted the run about 10 days later, and Anthropic released upgraded versions of its frontier models in November.
Chip makers are also still pushing new hardware. Nvidia started shipping its newest AI computing system, the Vera Rubin NVL72, in October and called it the fastest product rollout in the company’s history. Nvidia’s latest results also showed data center revenue rising 110 percent from a year earlier to $107.6 billion, according to the company.
A key question now is where AI spending goes if the biggest labs slow frontier training again. Demand might shift toward other uses, like running AI features in products people already use, rather than building ever-larger models.
Source: The Verge AI