Baseten to Raise $1.5 Billion Amid Ongoing Inference Gold Rush

AI inference-focused startup Baseten is in final negotiations to secure a massive $1.5 billion funding round at a valuation of approximately $13 billion, according to a report from the Wall Street Journal. A June 18, 2026 article from TechCrunch noted that this round follows the company’s Series E funding just five months earlier, where it raised $300 million at a valuation of $5 billion—marking an extraordinary 160% increase in valuation.

Founded in 2019, Baseten offers a platform designed to execute inference processes—where AI models interpret and act on user inputs—at high speeds and low costs. A key feature of the company’s technology is its ability to automatically route tasks to the most appropriate model, leveraging cost-efficient and high-performing open-source models to minimize inference costs for its clients.

The funding round is reportedly being led by four prominent firms: Spark Capital, Sands Capital, Altimeter Capital, and Wellington Management. However, this is not a standard funding round. It is a “split-priced round,” meaning some investors may participate at the $13 billion valuation while others may join at a lower $11 billion valuation. According to the Wall Street Journal, this strategy is increasingly being used by startups to inflate their headline valuation.

Startups Specializing in Inference

Baseten is at the forefront of what The Next Wave has called the “inference gold rush.” While model training has historically received more attention, the increasing deployment of large language models (LLMs) has led to an explosion in demand for inference processing.

The company provides an API-based service that selects the optimal model based on the user’s input and executes the inference. For example, it automatically assigns a high-performance closed model for complex math problems, while routing simpler queries to lightweight open-source models—effectively optimizing total costs for clients. This approach has gained traction, particularly among enterprise clients concerned about inference costs.

Baseten raised $150 million in a Series D round in 2024, followed by $300 million in a Series E round in early 2025. With this latest funding round, the company’s total funding is expected to approach $2 billion. The rapid rise in valuation over just 18 months reflects the broader trend of AI moving into operational phases.

160% Growth in Valuation Within Five Months

Assuming the Series E round was announced in January 2026, this latest funding round comes just five months later. The leap from a $5 billion valuation to $13 billion (or $11 billion for some investors) underscores investors’ high expectations for the AI inference market.

However, the split-priced nature of the round deserves scrutiny. The participation of some investors at an $11 billion valuation suggests that not all parties agreed with the $13 billion figure. This approach benefits lead investors by boosting the perceived valuation of their portfolios while allowing other participants to invest at a lower risk. However, if Baseten fails to maintain its valuation in subsequent rounds, it could create friction among investors.

Baseten’s rapid growth illustrates the concentrated capital investment in the “inference layer” of the AI industry. Competitors like Together AI, Fireworks AI, and Anyscale have also secured significant funding. Investors appear to be betting that the demand for inference will soon surpass that for model training, driving this surge in financial backing.

The Reality of Split-Priced Rounds

Split-priced rounds have become a popular strategy for startups navigating a challenging funding environment. By allowing some investors to participate at higher valuations and others at lower valuations, startups can inflate their headline valuation while offering lower-risk participation to certain investors.

In Baseten’s case, a $2 billion difference in valuation between the two groups of investors suggests that different terms and conditions might have been applied. This strategy, increasingly common in large funding rounds, is often used by startups to maintain high valuations.

However, split-priced rounds come with risks. If a startup cannot sustain its valuation in subsequent funding rounds, early investors who participated at higher valuations may face challenges, potentially causing tension among stakeholders. Baseten’s ability to justify its lofty valuation through revenue growth will be a critical factor moving forward.

Investors and Round Structure

The latest funding round is co-led by Spark Capital, Sands Capital, Altimeter Capital, and Wellington Management—firms known for their expertise in the technology sector. Altimeter Capital, in particular, has been actively investing in AI-related companies, while Wellington Management is recognized for its large crossover investments.

Other investors from previous rounds include Coatue Management, Lightspeed Venture Partners, and Madrona Venture Group. While it is unclear which new investors are participating in this round, the high valuation suggests the possible involvement of large institutional investors, such as sovereign wealth funds or pension funds.

The Inference Gold Rush

Within the AI industry, there is a growing shift of GPU resources from model training to inference processing. Although inference costs for large-scale models like OpenAI’s GPT-4 and Anthropic’s Claude 3.5 remain high, advancements in open-source models have made more affordable options viable for many tasks.

Baseten claims its platform can reduce inference costs by up to 90% by efficiently routing tasks between open-source models like Llama 3, Mistral, and Falcon, and closed models like GPT-4 and Claude, depending on the task requirements. This cost-optimization strategy has proven particularly attractive to large SaaS companies and enterprise clients struggling with high inference costs.

The rapid growth of the inference market could also help alleviate GPU shortages. Since inference requires less GPU memory compared to training, it allows for more simultaneous processing of requests. If inference-focused platforms like Baseten become more widespread, more efficient utilization of GPU resources is expected.

Baseten’s Technology and Differentiation

Baseten’s core technology lies in model routing and inference optimization. Its platform provides the infrastructure necessary for AI inference, automating routing across multiple model providers. This enables users to access various models through a single API.

The company is also focused on accelerating inference. Its system optimizes the entire inference pipeline through GPU memory management and batch processing. Additionally, it automatically selects the most suitable model and hardware configuration based on latency requirements.

Baseten’s competitive edge is its accuracy in model routing and cost optimization capabilities. By leveraging open-source models, which are significantly more cost-effective than closed models, the company lowers the barriers to AI adoption for businesses. Its customer base is growing rapidly, particularly among companies in fintech, healthcare, and e-commerce.

Challenges Ahead

Baseten’s most significant challenge will be meeting revenue growth expectations to justify its valuation. At $13 billion, its valuation far exceeds that of competitors like Together AI ($3 billion) and Fireworks AI ($2 billion). To sustain this level, the company must rapidly expand its customer base and revenue streams.

The increasingly competitive landscape is another hurdle. Major cloud providers like AWS, Microsoft Azure, and Google Cloud Platform are bolstering their inference services, putting pressure on independent startups like Baseten to differentiate and compete. Additionally, as AI models evolve, the computational requirements for inference may also change, presenting further challenges.

The risks of split-priced rounds also loom large. If Baseten fails to maintain its valuation in future funding rounds, it could create conflicts among its investors, particularly those who joined at the higher $13 billion valuation.

Editorial Opinion

In the short term, this funding round is likely to act as a catalyst for increased capital flow into the AI inference market. The announcement of Baseten’s massive fundraising could encourage rival startups to announce similar rounds. Moreover, the growing adoption of split-priced rounds could influence how startups strategize to maintain high valuations, especially in later funding stages.

In the long term, the inference layer is expected to grow in importance within the AI industry’s value chain. While model training may become dominated by a few major players, inference processing is poised to become a competitive market with numerous entrants. Startups specializing in inference, like Baseten, are likely to gain prominence, potentially sparking industry consolidation. However, the sustainability of growth in the inference market remains a question. Baseten’s valuation assumes a far more advanced adoption of AI across enterprises than what is currently observed.

References

• TechCrunch AI — Published 2026-06-18