AI labs are losing money on cheap tokens as GPU costs surge 82%
AI labs are cutting token prices to win enterprise customers while the GPUs powering those tokens keep getting more expensive to rent. That math does not work forever.
The GPU cost squeeze
Labs lock in GPUs on multiyear contracts at $2 to $4 per hour. When that reserved capacity runs dry, spot pricing runs as high as $13 per hour. One investor’s analysis estimates Anthropic’s average GPU cost rose 82% in two quarters. A $5 billion short-term compute bill in Q3 was enough to flip an operating profit back into a loss.
Top spenders already pulling back
Separately, the heaviest AI buyers are showing signs of fatigue. The top 1% of firms on Ramp spent less on AI per employee in September than in July. Anthropic added $5.4 billion in revenue in Q3, down from $6.8 billion in Q2. Adoption across other firms is still climbing, but the biggest spenders have stalled.
Where this ends up for enterprise buyers
The investor analysis predicts labs will push enterprise customers into multiyear minimum commitments, structured to mirror the length of the GPU leases behind them. If you’re negotiating AI contracts for your business, expect pricing and commitment structures to get more rigid, not less. The era of flexible, month-to-month AI spend at scale may be shorter than it looked.
I’m founder of BizStack at Cagri Sarigoz LLC and a passionate advocate for entrepreneurs.
With over 14 years in tech, marketing, and AI, including my role as Head of SEO at CitizenShipper and co-founder of TaleBot at Intale AI, I’m dedicated to sharing genuine, useful product insights and tips.
At BizStack, I aim to cut through the digital noise to provide clear, actionable advice.
And more than all else, I’m a father to a (always) little girl and a husband.
Contact me at [email protected] for assistance.
