AI startups raise billions, but most of that money is devoted to model training. What happens after the model is trained? It must be carried out in production without acquiring property and sacrificing large sums of money. That’s where Baseten comes in. The San Francisco-based startup has closed the $75 million Series C round, co-led by IVP and Spark Capital, with support from Greylock, Conviction, South Park Commons, 01 Advisors and Lachy Groom.
With this fresh funding, Baseten has now raised $135 million to tackle one of AI’s biggest challenges: reasoning. The process of running AI models in real applications. This funding will drive product expansion, employment and global growth as businesses expand to meet the growing demand.
It was founded in 2019 by Amir Hahigat, Tuhin Srivastava (CEO), Philip Haus and Pankaj Gupta, who has grown to a team of about 60 employees for six years. The company also builds a strong customer base and serves over 100 companies and hundreds of small and medium-sized businesses, including writing, patreons and writers.
With AI efficiency becoming a top priority following a major breakthrough from China’s AI lab Deepseek in January, Baseten moved quickly to support Deepseek’s R1 Reasoning model, competing with Openai’s O1. The company promotes its ability to deliver top-tier performance at a significantly lower cost than Openai.
Deepseek argues that its model is trained at some cost compared to its US counterparts, increasing interest in alternatives to expensive AI solutions.
According to CEO Tuhin Srivastava, demand for Baseten is surged from organizations exploring a switch to Deepseek. The company is working hard to meet growing interest, ensuring the platform meets the needs of businesses looking for more affordable AI solutions.
“There are a lot of people who are paying millions of dollars per quarter to Openai and humanity who are thinking, ‘How can I save money?’,” he said. “And they flocked.”
Increased demand for AI inference
AI products are increasingly embedding models as core components rather than optional add-ons. This shift means inference, the process of querying a trained model and receiving the results. However, performing inferences on scale is expensive and often suffers from slow response times and GPU shortages.
“People who build AI products that aren’t worried about reasoning are not on the scale yet,” said Will Reed, general partner at Spark Capital. “All successful AI products require exceptional inference performance. Otherwise, no one wants to use it. And you bet the future of your product or your company with that performance. When you’re doing it, choosing the right partner is a make-or-break.”
How Baseten makes AI models work in production
For many companies, moving from AI development to deployment is a painful process. Models that work well in lab settings often struggle to produce, leading to delays, downtime and unexpected costs. Baseten’s platform is designed to handle layers of layers, allowing AI teams to focus on building products rather than worrying about infrastructure.
Instead of running its own data center, Baseten deploys software from leading cloud providers such as Amazon and Google into their infrastructure. While enterprise customers can integrate their own infrastructure through dedicated tiers, Baseten’s multi-cloud approach ensures access to a larger pool of GPUs than a particular provider can offer.
“In this market, your No. 1 differentiation is how fast you can travel. This is a central advantage for customers,” Srivastava told CNBC. “We can go to production without worrying about reliability, security or performance.”
Why investors are betting on Baseten
Baseten’s approach relies on AI-powered products and has already won over customers like Abridge, Gamma, Writer, and others who need a model to run under real-world conditions.
“We’re looking forward to seeing you in the future,” said Tuhin Srivastava, CEO and co-founder of Baseten. “Speed, reliability, and cost-effectiveness are non-negotiable, and that’s what spends 100% of our focus. It’s its commitment and incredible we’ve gathered billions and gathered billions. The trust we built with such a group of customers has essentially allowed us to grow five times over the past year with zero churn.”
Baseten’s team has 50 employees with talent from companies such as Github, Google, Uber, Amazon, Palantir, Atlassian and Airtable, and currently has 50 employees. The company also deploys multi-cloud support, hybrid cloud capabilities, and integration with Tensort, making it easier for AI teams to run their models efficiently.
“Baseten is continuously focused on creating performance, reliability, scalable and multi-cloud in AI inference,” says Sarah Guo, General Partner and Founder of Consucction. “Their growth is driven by delivering mature products at the right time and accelerating with strong tailwinds. Advances in model capabilities, more open source models, and interest from companies in shipping production AI applications. The rise of the world.”
The road ahead
When AI adoption surges, the demand for fast, reliable inference only increases. Baseten has established itself as a go-to provider for AI native companies that need to deploy models without encountering infrastructure bottlenecks. With fresh capital and growth industry momentum, startups are preparing for their next phase of growth.
Source link
