san francisco: The world’s leading artificial intelligence companies are spending hundreds of billions of dollars on the computer data centers needed to create AI.
Big tech companies like Amazon and Google and well-funded startups like Anthropic and OpenAI are among the few organizations with the money and connections to access all the computing power.
As a result, other organizations become isolated.
Anjney Midha, a partner at venture capital firm Andreessen Horowitz and a serial technology entrepreneur, wants to change that dynamic with an unusual startup. His young company, Amp, is purchasing additional computing power from data center operators in the United States and other countries so that it can share that valuable resource with everyone who needs it.
Menlo Park, California-based Amp aims to create a global pool of specialized computer chips that can be used by startups, universities, and other organizations that don’t have access to the massive amounts of computing needed to train the most powerful AI models.
“Some companies just can’t get the computing power they need,” Mida said. “The world’s richest and most powerful corporations are hoarding infrastructure for themselves.”
Amp has raised more than US$1.3 billion (RM5.16 billion) from investors including Andreessen Horowitz, startup incubator Y Combinator and various cloud computing providers.
Several prominent startups have also agreed to help leverage and share the company’s pool of computing power, including Periodic Labs, an AI startup focused on scientific discovery, and Celebrity Labs, which builds AI systems for speech generation.
Amp is part of a broader effort to pool AI infrastructure. Chipmaker Nvidia and French startup Mistral announced earlier this year that they would pool computing power to build AI systems for European companies and nations, hoping to reduce dependence on big U.S. tech companies.
Midha likened his startup to building a power grid. Just as pools of electricity can be shared by homes and offices, pools of computing power can also be shared by start-ups and universities.
Investors contribute funds that Amp can use to purchase computing power from data center operators. AI startups can then join the coalition and use the computing power to build AI models. In return, these startups provide additional funding and perhaps other resources.
Some startups may share digital data needed to train AI models. Alternatively, you might share the model itself and distribute the core software code to other members of the federation. Companies may also work together to train models.
Periodic Labs CEO Liam Fedas said the coalition’s ultimate value was collective bargaining. On its own, his startup would struggle to get the computing power it needs. But Amp could gain even more leverage if it were able to negotiate with data center operators on behalf of many startups.
“If we pool demand, we can have more serious conversations about purchasing computing power,” Fedas said. – ©2026 The New York Times Company
This article was originally published in The New York Times.
