Baseten raises $1.5 billion at a $13 billion valuation, with Blackbird making its biggest bet ever

Baseten has closed a $1.5 billion Series F at a valuation of up to $13 billion, a record for the AI infrastructure segment. The biggest surprise was the participation of Australian fund Blackbird VC, with an amount that may turn out to be the largest single venture bet in Australian history.

What Baseten is

Baseten is a platform for deploying and serving AI models in production. Founded in 2019 in San Francisco, the company specializes in "AI inference": running trained models in real products with guarantees on speed, stability and cost.

In essence, Baseten does for neural networks what AWS Lambda does for serverless functions. An engineer uploads a model, configures the environment and gets a ready endpoint with autoscaling, monitoring and contractual SLAs. There is no need to build and maintain a GPU cluster yourself.

Over the past two years the company has shifted its focus from working with individual ML teams to serving large enterprise customers. Industrial companies, fintech and medtech are moving to their own AI deployments, and they need a reliable infrastructure-level contractor. This shift from prototypes to production created explosive demand and let Baseten's valuation climb so quickly.

Blackbird's record bet

Blackbird VC is one of Australia's best-known venture funds, with a portfolio that includes Canva, Airtasker and other successful companies from the Pacific region. Its participation in the Baseten round is unusual: Blackbird very rarely invests in American startups, let alone at this scale.

According to sources, the fund's investment may exceed any other single investment in the history of the Australian venture market.

This is not just one fund's record. It signals a strategic shift: large Asia-Pacific funds are increasingly seeking a stake specifically in American AI infrastructure, rather than in applied products or the models themselves.

The structure of the round is telling too: $1.5 billion is a substantial sum even for the American market. Amid a partial cooling toward public AI valuations, a round like this means institutional investors remain convinced of multi-year growth in the infrastructure segment.

Why "shovels" beat "gold"

A thesis is getting louder in venture circles: in the AI era, the biggest money is made not on models but on the tools that serve them. Several reasons why it works:

• Models become obsolete quickly; deployment platforms do not
• The enterprise market demands SLAs, auditing and multi-cloud support, which pure API providers do not offer
• Price competition among LLM providers cuts model margins but increases demand for infrastructure
• Companies are moving from the OpenAI API to their own deployments for control over data and economics
• Valuations of infrastructure companies are growing faster than those of applied AI products

This context explains why Baseten reached a $13 billion valuation in just 18 months, a pace unusual even by the standards of the red-hot AI market.

What this means

The Baseten round is further evidence that the race for AI leadership has moved from "who builds the best model" to "who provides reliable infrastructure to run them." Amazon, Google, Microsoft and independent players are now competing not over development but over the right to serve enterprise AI at industrial scale. Whoever builds the most reliable "engine" will take this market.