The artificial intelligence investment frenzy is increasingly shifting away from the companies building large language models and toward the firms supplying the infrastructure that allows those models to operate at scale.
That trend was underscored this week when California-based AI startup Baseten announced a $1.5 billion funding round that values the company at $13 billion, one of the largest private funding deals in the AI infrastructure sector this year.
The round was led by Sands Capital and Wellington Management, while Australian venture capital giant Blackbird VC participated with what it described as the largest investment in its history. Although Blackbird did not disclose the amount invested, the firm said the deal represents its biggest financial commitment to date.
The financing highlights the enormous investor appetite for companies positioned in the less glamorous but increasingly critical layer of the AI ecosystem: inference infrastructure. While firms such as OpenAI, Anthropic, and Google have attracted attention for developing advanced AI models, a growing number of investors are focusing on the infrastructure providers that help enterprises deploy, customize, and run those models efficiently.
Baseten sells software and computing infrastructure that enables companies to build and deploy customized AI applications. Rather than creating foundational models itself, the company provides the tools that allow businesses to operationalize AI systems, often at lower costs than relying directly on major model providers.
That positioning appears to be resonating strongly with customers.
Baseten said revenue has increased 20-fold over the past year, driven largely by surging demand for inference services. Inference refers to the stage where trained AI models are put into real-world use, generating responses, recommendations, images, code, or other outputs for customers.
The distinction is becoming increasingly important across the AI industry. Training frontier AI models requires enormous computing resources and billions of dollars in investment. However, many analysts believe inference could ultimately become the larger market because every AI interaction, query, and application depends on inference infrastructure after a model has been trained.
As enterprises increasingly embed AI into products, workflows, and customer services, demand for efficient inference platforms is expected to rise sharply.
Baseten’s latest fundraising is its fourth capital raise in just 18 months, reflecting how rapidly investors are pouring money into infrastructure providers that support the commercialization of generative AI.
“It’s a signal of conviction,” Blackbird partner Michael Tolo said, explaining the firm’s decision to deepen its investment.
Tolo argued that the economics of AI deployment are beginning to change in ways that favor infrastructure specialists.
“For companies building AI into their tech systems, Baseten competes with companies like OpenAI and Anthropic at a lower price, and this is the biggest shift that we’ve seen in both unit economics and competitive leverage within the AI market so far,” he said.
However, enterprises are becoming increasingly sensitive to costs associated with running AI workloads as competition intensifies. While foundation model providers have spent heavily building cutting-edge systems, a parallel race has emerged among infrastructure companies seeking to reduce deployment costs and improve performance.
Investors now see that layer of the market as potentially more durable and profitable than model development itself.
The funding round also bolsters a wider investment narrative that has dominated technology markets over the past two years. Much of the capital flowing into AI has focused on the infrastructure stack rather than end-user applications. Chipmakers, cloud providers, data center operators, networking companies, and inference platforms have attracted massive valuations as investors bet they will benefit regardless of which AI models ultimately dominate the market.
The strategy mirrors previous technology cycles where infrastructure providers often emerged as some of the biggest winners. During the internet boom, for example, companies supplying networking equipment, cloud infrastructure, and software platforms frequently generated more sustainable returns than many consumer-facing startups.
Baseten now joins a growing list of AI infrastructure firms attracting multibillion-dollar valuations as investors take positions for what many believe will be years of rising enterprise AI adoption. The company said it plans to use the new capital to expand computing capacity, enhance its software platform, and hire additional staff.
The fundraising also highlights Australia’s growing footprint in the global AI ecosystem. Baseten was co-founded by Australians, while Blackbird’s participation demonstrates how local venture capital firms are increasingly gaining exposure to some of Silicon Valley’s most valuable private technology companies.
