Today, Together AI and Y Combinator (YC) are announcing a partnership to deliver the first dedicated YC GPU cluster, giving YC's portfolio of AI-native startups easier access to the compute they need to build and scale.
Compute has become the biggest bottleneck #
Breakthroughs in AI have led to a new generation of companies building AI applications, and a key input to building any of these apps is access to compute. More and more startups struggle to get this access in a timely, cost-effective way.
It used to be simple. A startup could spin up GPU instances on demand and scale from there. Today, as model quality and token value keep rising, just securing capacity is one of the hardest problems a young company faces, let alone getting good pricing on it.
For many startups, the upfront cost to secure two years of compute exceeds their entire cash balance, forcing a choice: Raise a round just to fund a compute contract, or go without the capacity to compete.
Flexible, dedicated access to compute #
Together AI and Y Combinator have partnered to solve that specific problem. Together AI has built a dedicated cluster and developer experience just for YC Portfolio companies to quickly and cost-effectively get access to compute for their AI needs across inference and training. Instead of long-term commitments, startups can spin up GPUs for short-term sprints while benefiting from long-term rates.
The cluster supports the full range of needs across YC's portfolio, from teams requiring single-node compute to companies scaling up as their needs grow. Additionally, YC start up founders benefit from the learnings and best practices that Together researchers, engineers, and customer experience teams bring from working with leading AI-native companies.
Startups reserve, provision, and manage their own GPUs directly through Together’s self-service portal, with their own billing support, so scaling compute never has to route through YC. Founders get GPUs ready in minutes, and stay in control of their own usage from day one.
The cluster is running at full utilization today, while individual companies can still plan their compute needs months in advance.
A natural fit #
Together AI was founded four years ago on the belief that generative AI would become foundational, and built a cloud service spanning the full generative AI lifecycle. It now works with over 8,000 customers, including Cursor, Decagon, Cognition, and Eleven Labs.
As models improve, the value of every token produced keeps climbing, and so does the cost of compute behind it. Together's research, including work on attention mechanisms and the Mamba architecture now used in models like NVIDIA’s Nemotron, is aimed squarely at that problem: faster inference and better unit economics for every workload on the cluster, from early-stage teams to companies already operating at scale.
YC has long been one of the largest seed funders of research-driven companies, and securing compute has become essential to attracting and supporting the best founders.
Both organizations share a belief that founders do their best work with access to the same resources a much bigger company would have. Compute is simply the latest one.
What's next #
Together AI and YC plan to expand the cluster, and Together will keep supporting companies as they graduate from the batch and scale with whatever setup makes sense long term. As those needs grow, founders can also tap into the rest of Together’s platform, from inference to fine-tuning and training, all on the same infrastructure they already know.
If you're building an AI-native startup and want to learn more, contact us. YC's next batch is currently taking applications for their Fall 2026 cycle. YC is especially excited about funding companies developing new research breakthroughs that require GPU compute. Apply at* ycombinator.com/apply**.*