- AI model training costs could skyrocket to $100 billion by 2026.
- Anthropic CEO predicts few players able to afford top-scale training.
- Specialization may prevent rapid commoditization of AI models.
The race to develop cutting-edge artificial intelligence models is heating up, but the CEO of AI powerhouse Anthropic isn’t breaking a sweat.
Dario Amodei predicts the astronomical cost of training advanced language models could soon balloon to a staggering $100 billion.
Keeping the competition at bay
In a recent interview with CNBC, Amodei brushed off concerns about rapid commoditization, stating the number of players financially capable of training professional-level AI models at top scale “is going to be relatively small.”
He foresees current $100 million models escalating to a whopping $10 billion in just two years.
Gazing into the AI crystal ball
Amodei believes diversity in development techniques could help stave off commoditization, likening different models to the unique traits of human beings.
Some AI models may specialize in areas like law or national security, while others gain expertise in biochemistry, leading providers to “specialize in different things, even as the base model they made is the same.”
