Together AI and Y Combinator launch first dedicated GPU cluster for YC startups

Together AI and Y Combinator announced in July 2026 the launch of the first dedicated GPU cluster for startups in the Y Combinator community. It gives them access to compute without the usual two-year capacity-rental contracts.

What YC startups get

Startups in the Y Combinator portfolio get direct access to GPU capacity on Together AI's cloud platform, with no long-term commitments. Instead of signing a contract two years in advance, a team takes compute when it actually needs it: for training a model, running an experiment, or launching a product.

Together AI specializes in cloud infrastructure for training and inference of AI models, and this cluster is the first one reserved specifically for the YC community. Key details of the partnership:

• The first GPU cluster dedicated specifically to the Y Combinator community
• The partners are Together AI (cloud AI infrastructure) and the Y Combinator accelerator
• The key change is dropping two-year contracts for renting compute
• Fast GPU access instead of queues and lengthy approvals

Why two-year contracts are a problem

A two-year GPU contract locks up capital that an early-stage startup barely has. A young team that needs to test a hypothesis within a few weeks has two options: pay for capacity that sits idle, or skip its own infrastructure altogether and slow down development.

The Together AI and Y Combinator partnership removes that barrier. A startup pays for what it uses and does not tie itself to commitments years ahead. In a community where iteration speed decides everything, the difference between "GPU access in a day" and "months of contract negotiations" is the difference between launching a product and stalling.

What the partnership changes

Y Combinator is one of the largest startup accelerators in the world, and Together AI serves as the infrastructure layer for AI development. Their joint cluster addresses a specific pain point: young AI teams find it harder and harder to get GPUs while large labs and corporations snap up capacity.

"No more two-year compute contracts.

Together AI and YC just gave YC startups a faster way to get GPUs," Together AI said in its blog.

A dedicated cluster means YC startups will not have to compete with the rest of the market for shared resources, because the capacity is reserved for the community. It is a first step toward making access to scarce compute independent of check size and contract length.

What this means

GPUs stop being a bottleneck at the start: instead of long contracts and frozen capital, AI startups get compute on demand. For early teams, this removes one of the main infrastructure barriers between an idea and a working model.