Andromeda, an AI infrastructure startup building on-demand access to large-scale GPU compute, has raised a new funding round as it accelerates its mission to create a global marketplace for artificial intelligence computing power. The company operates in the rapidly expanding AI infrastructure sector and focuses on enabling developers, startups, and AI labs to rent and manage high-performance compute resources without owning physical hardware.

The latest round values Andromeda at approximately $1.5 billion and includes $60 million in new funding led by Paradigm, bringing Paradigm’s total investment in the company to about $60 million. The financing reflects continued investor conviction in the long-term demand for GPU infrastructure as AI model training and deployment scale globally.

The investment was led by Paradigm, a venture capital firm known for backing emerging technologies across crypto, fintech, and AI infrastructure markets. Paradigm’s participation builds on its earlier investment in Andromeda and signals ongoing support for the company’s vision of turning compute into a standardized, tradable commodity similar to energy or other global markets.

Andromeda’s platform is designed to simplify access to high-performance computing clusters by aggregating GPU supply from multiple sources and offering it through a unified interface. Customers can rent compute capacity on demand, avoiding long-term infrastructure commitments and reducing the complexity of sourcing GPUs in a market where demand has significantly outpaced supply in recent years.

The company’s origins trace back to AI-focused investors Nat Friedman and Daniel Gross, who initially developed internal GPU infrastructure to support their portfolio companies. That effort evolved into a large-scale compute cluster before being spun out into a standalone company under CEO Wil Moushey, who now leads its commercialization and expansion strategy.

Early versions of Andromeda’s infrastructure included thousands of NVIDIA H100 GPUs deployed to support AI startups facing severe compute shortages. The system was originally built to solve internal bottlenecks in AI model training, but it has since expanded into a broader platform targeting enterprise customers and AI labs seeking flexible access to compute resources at scale.

Investor interest in Andromeda reflects a broader shift in venture capital toward “AI infrastructure as a market,” where compute is increasingly treated as a financialized asset class. Rather than simply selling cloud services, Andromeda aims to standardize GPU access and pricing, effectively building a marketplace layer for global compute distribution.

The company’s strategy centers on scaling its compute network to as many as 100,000 GPUs, positioning itself as a neutral infrastructure provider that can serve AI labs, startups, and enterprise customers alike. It also aims to improve liquidity in GPU markets by matching supply and demand more efficiently across providers and geographies.

With continued expansion of its infrastructure footprint, Andromeda is positioning itself as a foundational layer in the emerging AI economy. As demand for compute continues to surge, the company’s approach reflects a broader industry belief that access to GPUs will become one of the most critical constraints—and opportunities—in the next phase of artificial intelligence development.

Related Articles