Baseten Secures $150 Million in Series D Funding, Reaching $2.15 Billion Valuation

Exclusive: Baseten, AI inference unicorn, raises $150 million at $2.15 billion valuation

In some sense, Baseten started with building blocks -- literally. "So, I grew up in Australia," said Tuhin Srivastava, CEO and cofounder of AI inference unicorn Baseten. "And, in Australia, you learn to count with these 'base ten' blocks. They're the building blocks of how you count, how you operate." Years after his days of stacking yellow and green blocks in the classroom, Srivastava now spends his time playing with a very different set of building blocks. The Baseten CEO and cofounders Amir Haghighat, Philip Howes, and Pankaj Gupta provide the infrastructure that AI models run on. It's a classic "picks and shovels" business that, right in the middle of a rollicking AI boom, is helping companies deploy, manage, and scale AI applications. Or, to use Srivastava's analogy, the company is laying "the train tracks so the models can run." The trains that run on Baseten's tracks can be anything -- from large language models to AI models that generate video or that turn voice into text. Srivastava told Fortune that Baseten's revenue has grown by "more than 10x" over the last 12 months, and the company's big-name customers include Abridge, OpenEvidence, Clay, Patreon, and Writer. Now, six months after raising a $75 million Series C, Baseten is re-upping with a $150 million Series D, Fortune has exclusively learned. The deal, which almost triples Baseten's valuation to $2.15 billion, was led by Bond, with participation from new investors including CapitalG, Premji, and Scribble as well as existing investors Conviction, 01a, IVP, Spark, and Greylock. Baseten focuses on the inference side of AI -- the process by which trained AI models use their knowledge to generate predictions and decisions. "Inference will be one of the biggest markets in AI," said Sarah Guo, Conviction founder and an early Baseten investor, via email. "It has become clear that open source AI and custom models are here to stay. The ongoing struggles of research labs to provide performance, scale and reliability in the face of massive demand shows just how technically hard this problem is, and Baseten has the leading product and most sophisticated, stickiest and happiest customers in the category." In his role at Baseten, Srivastava has a front seat to the state of the AI application layer, and he says it's moving as fast as conventional wisdom states, "very dynamic, very savvy, kind of no sunk cost, and off to the races very, very quickly." "The biggest challenge [and] opportunity are the same: rapid growth," said Jay Simons, Bond general partner, via email. "Growing and scaling as quickly as they are is a challenge for any company... Baseten carries an incredible amount of responsibility for its customers... It's a critical service that needs to be absolutely bullet-proof." What happens if the AI wave crashes? I asked Jill Chase, CapitalG partner. "It's not like they need tons of unique companies building AI products, which I think there will be," said Chase. "But even if there weren't, all AI is pegged to inference, which is usage of AI products. I think that's a bet everyone is very comfortable making -- that AI usage will continue to grow massively over time, even if it's consolidated to three or hits three billion customers." There's certainly precedent for startups growing into behemoths by serving other startups (take Stripe, for example). The question, of course, is how big this market ultimately is. Srivastava's placed his bets. "We think AI applications are just the last great market," he said. "This will be bigger than anything we've ever seen... We think that's the end state of the world from a technology perspective, at least as far as we see it... We want to be the index [for that economic growth]. If the market as a whole wins, we win" In short, Srivastava believes that the building blocks, the foundation, will hold. Joey Abrams curated the deals section of today's newsletter. Subscribe here.

AI inference startup Baseten closes $150M investment backed by CapitalG - SiliconANGLE

AI inference startup Baseten closes $150M investment backed by CapitalG Artificial intelligence startup Baseten Labs Inc. today announced that it closed a $150 million investment at a $2.15 billion valuation. BOND led the Series D round, which comes about six months after the company's previous raise. It was joined by Alphabet Inc.'s CapitalG fund, Conviction, Premji Invest, 01A, IVP, Spark, Greylock and Scribble Ventures. Baseten offers a software platform that promises to help enterprises speed up their AI inference workloads. According to the company, its technology provides up to 50% faster performance than competing products. Customers can deploy the software on their own infrastructure or use a cloud-based version managed by Baseten. The cloud version runs across 10 different infrastructure-as-a-service platforms. When traffic to an AI application increases, Baseten automatically provisions additional graphics cards on the most suitable platform. If one of the public clouds used by a customer experiences technical issues, the software can automatically switch to another. Baseten uses an approach called topology-aware parallelism to optimize customers' infrastructure. When an AI model runs across multiple graphics cards, those graphics cards must coordinate their work by exchanging data over the network. Topology-aware parallelism reduces data traffic and thereby lowers hardware usage. Baseten says that it optimizes not only the hardware on which customers' AI models run but also the models themselves. It uses an approach called operator fusion to combine calculations that are usually done separately into a single computation, which saves time. A quantization tool shrinks neural networks to reduce their memory requirements. Baseten offers its core inference features alongside a set of developer tools. Software teams can use those tools to automate some of the work involved in releasing AI models to production. Baseten also makes it easier to deploy a model's dependencies, or the software module on which it relies to run. Once a model is deployed on Baseten's platform, developers can use a built-in observability tool to monitor its performance. The software tracks the number of requests sent to the model, response times, hardware usage and related metrics. In May, Baseten expanded its focus beyond inference by launching an AI training service. It provides access to infrastructure that companies can use to build new AI models. Baseten allows software teams to periodically save AI models during training for backup purposes. If the training process is interrupted, developers can recover the most recent backup instead of starting from scratch. "Every breakout AI application depends on fast, reliable, and cost-effective inference, the same way the last 15 years of companies depended on the cloud," said Baseten co-founder and Chief Executive Officer Tuhin Srivastava.

Baseten Raises $150 Million to Scale AI Inference Platform | PYMNTS.com

This round nearly triples the company's valuation from where it was six months ago, bringing it to $2.15 billion, according to a Friday (Sept. 5) press release. Baseten's platform is purpose-built for inference and is used to power AI products and applications across industries, according to the release. It brings together applied AI research, flexible infrastructure and seamless developer tooling to help companies bring models into production, per the release. "Every breakout AI application depends on fast, reliable and cost-effective inference, the same way the last 15 years of companies depended on the cloud," Baseten Co-founder and CEO Tuhin Srivastava said in the release. "Baseten makes that possible." Baseten will use its new funding to invest in model performance research, infrastructure and developer tooling and to expand its customer teams, according to the release. Jay Simons, general partner at BOND, which led the Series D round, said in the release: "Inference is a major bottleneck, and solving it is key to unleashing the full potential of AI. Baseten is doing this today." PYMNTS reported Aug. 25 that inference is the stage where an AI model is used to generate predictions, responses or insights and that the costs of inference are recurring and can add up fast. Every AI-enabled workflow involves inference, and the more queries or predictions a company's AI systems make, the bigger the bill. Baseten's Series D funding round came about seven months after its Series C, which was announced on Feb. 19. In the earlier round, the company raised $75 million. In May, the company announced two new inference products powered by its Baseten Inference Stack: Baseten Model APIs for open-source AI models and Baseten Training features for training models to improve inference performance. "In the AI market, your number one differentiator is how fast you can move," Srivastava said in a press release announcing the products. "Model APIs give developers the speed and confidence to ship AI features knowing that we've handled the heavy lifting on performance and scale."