Two of the largest venture rounds of the week went to AI infrastructure companies Crusoe and Fluidstack, as the race to build compute capacity for artificial intelligence reaches multibillion-dollar scale. The weekly roundup, titled 'The Week’s 10 Biggest Funding Rounds: Crusoe And Fluidstack Lead Multibillion-Dollar AI Infrastructure Haul' on Crunchbase News, highlights how the U.S. Check out the full list of deals that reshaped funding landscapes.

What You Need to Know

AI infrastructure startups are attracting enormous capital because training and running advanced AI models requires massive computing power. Crusoe, once a crypto mining company, has pivoted to become a leading AI cloud provider. Fluidstack is building large-scale GPU data centers. Their combined $4.5 billion in new funding signals that investors see long-term demand for compute, not a bubble. Other big rounds spanned cybersecurity, food and HR software, but AI infrastructure dominated.

The New AI Infrastructure Giants

Denver-based Crusoe raised $3 billion in a Series F co-led by Atreides Management and Valor Equity Partners, with participation from Mubadala Capital. The company serves customers including OpenAI, Microsoft and Meta. Originally founded to capture stranded natural gas for Bitcoin mining, Crusoe now runs AI clouds and data centers. Its valuation has tripled to $30 billion in less than a year, according to Crunchbase. The company has raised nearly $7.2 billion to date.

New York-based Fluidstack secured $1.5 billion in a private equity round led by Jane Street Capital. The company provides GPU clusters and data center infrastructure for demanding AI workloads. Total funding now exceeds $2.6 billion, and the company is valued at $18 billion. Fluidstack represents a new breed of infrastructure providers racing to meet insatiable compute demand.

  • Crusoe: $3B Series F, co-led by Atreides Management and Valor Equity Partners.
  • Fluidstack: $1.5B private equity round led by Jane Street Capital.
  • Gimlet Labs: $300M Series B led by Andreessen Horowitz, building an inference cloud.
  • Upwind Security: $300M cybersecurity round, co-led by Bessemer Venture Partners and TCV.

Why This Matters

The sheer scale of these rounds reshapes the competitive landscape for AI compute. Crusoe and Fluidstack are not just raising capital; they are building physical infrastructure that will determine which AI models can be trained and deployed at scale. This creates a two-tier system: companies with access to massive compute clusters will dominate, while smaller players may struggle. The involvement of investors like Valor Equity Partners and Atreides Management signals that infrastructure is now a core asset class, not a niche. For cloud customers, this concentration of capital could drive down costs if capacity expands faster than demand, but it also risks creating bottlenecks around a few providers. The move by Gimlet Labs to optimize inference across chip types highlights a parallel trend: making compute more efficient is as critical as building more capacity.

Beyond AI Infrastructure

While AI infrastructure dominated, other sectors saw notable raises. Upwind Security, at $300 million, underscores the convergence of cloud security and real-time threat detection. Food company David raised $250 million for high-protein products, reflecting consumer trends toward protein-rich nutrition. The breadth of sectors shows that venture capital remains active, but the biggest checks are increasingly reserved for companies enabling the AI economy. The Crunchbase Megadeals Board tracks all venture deals over $100 million, and this week’s entries reinforce that infrastructure is the new frontier.

The Big Picture

Investors are betting that demand for AI compute will continue to grow exponentially. Crusoe’s transformation from crypto to AI mirrors a broader shift: the same energy and hardware assets now power machine learning instead of mining. Fluidstack’s rapid ascent shows that even late movers can attract multibillion-dollar valuations if they scale fast. The next few years will test whether these infrastructure bets pay off as AI models become more efficient and new architectures emerge.