OpenAI Offers Guaranteed Compute as AI Capacity Tightens
The program lets customers lock in long-term access to OpenAI compute through one- to three-year commitments.
Topics
News
- Anthropic Hires Andrej Karpathy for Claude Pretraining Team
- Google Overhauls Search With AI Agents, Interactive Results
- OpenAI Offers Guaranteed Compute as AI Capacity Tightens
- Analog Devices to Buy Empower Semiconductor for $1.5 Billion
- Advait Energy Signs Norway Pact for Hydrogen Fuel Cell Manufacturing
- Anthropic Buys Stainless to Strengthen AI Developer Tools
OpenAI on Tuesday introduced a new offering called Guaranteed Capacity, allowing eligible customers to secure long-term access to computing power as demand for artificial intelligence infrastructure rises.
The program lets customers make one- to three-year commitments, with discounts increasing based on annual commitment levels, OpenAI said.
Customers can draw down from those commitments across OpenAI products, with capacity planned across supported cloud providers and model families.
The offering is designed for production systems, customer-facing applications and AI agents that need predictable access to OpenAI compute, the company said.
OpenAI’s request form asks customers to specify workload type, endpoint or data-residency needs, and estimated capacity requirements measured in tokens per minute.
Chief Executive Officer Sam Altman said the move comes in response to rising demand for reliable access to compute infrastructure.
“Customers are increasingly asking us for certainty on capacity. As models get better, we expect that the world will be capacity-constrained for some time,” Altman said in a post on X.
Altman said OpenAI is offering discounted tokens for one- to three-year commitments, adding that the structure should also help the company plan its infrastructure needs.
The launch highlights growing pressure on compute availability across the AI industry, where large-scale infrastructure is required to train and operate advanced models. OpenAI previously told investors it is targeting around $600 billion in total compute spending by 2030.
OpenAI is also exploring ways to sell computing capacity directly to companies and individuals, a service Altman has described as an “AI cloud,” Reuters reported.
Such a move would place OpenAI more directly in competition with large cloud providers and emerging AI infrastructure firms.
The company has been looking for new revenue streams as it funds a costly infrastructure buildout.
