7GC invests in Fluidstack Marking it’s 8th Investment in FUND II

We're excited to announce our 8th investment from Fund II into Fluidstack, which falls squarely within our generative AI investment theme.

Founded by Gary Wu and Cesar Maklary, London-based Fluidstack is uniquely positioned to capitalize on the rapidly growing demand for AI computing resources. Their strategic approach and vision aligned perfectly with our current AI thesis, making them a perfect partner for 7GC.

If we think about the AI technology stack, there are three layers – infrastructure, platform, and application companies. Like prior computing cycles, infrastructure is the first beneficiary before traversing to the platform, and finally the application layer.

The infrastructure layer remains squarely owned by Nvidia GPU technology, with hyperscalers (Amazon, Google, Microsoft) receiving the most substantial GPU allocations from Nvidia and offering the hardware through their cloud offerings to their customers. The hyperscalers are rigorously competing for GPU supply to maintain an advantage in this new compute cycle race, with OpenAI/Microsoft, xAI, and Meta all building GPU clusters with over 100,000 GPUs. These individual training clusters cost in excess of $4 billion of server capital expenditures alone. This competitive dynamic has created a challenging supply and demand disequilibrium for the hardware. It has forced almost all leading AI research labs needing GPUs to access compute directly through the hyperscalers, allowing these businesses to become the default middleman.

As evidenced above, Nvidia GPU demand is not slowing but is continuing to grow as computing clusters scale. These clusters, as they scale in size, become incredibly complex software problems as companies look to ensure the highest utilization rate efficiency by reducing failure rates. With hyperscalers continuing to build, Fluidstack is becoming a powerful alternative partner for AI research labs in their own build process.

Fluidstack is an alternative for many AI labs that do not get full access to their GPU computing needs and need 24/7 live customer support to support their research capabilities. In other words, they help manage, source, and monitor customers' GPU clusters. This differentiated product offering is working – with some of the leading research labs worldwide tagged as customers, including Poolside, Character.AI, Mistral AI, and Reka.

We will update our LP base more significantly once the company formalizes the round announcement later in Q4 2024.