Artificial intelligence chip startup SambaNova has secured $1 billion in new financing, underscoring investors’ growing appetite for companies seeking to challenge Nvidia’s dominance in one of the fastest-growing segments of the AI industry.
The latest funding round values the California-based company at $11 billion and comes as demand shifts from training large AI models to running them efficiently in real-world applications, an area known as AI inference.
The financing was led by General Atlantic, with participation from Seligman Ventures, T. Rowe Price and Capital Group, adding to a wave of investment flowing into semiconductor startups as enterprises expand their AI infrastructure.
The announcement follows another major capital raise earlier this year, when SambaNova secured more than $350 million from investors including Intel, which also entered into a strategic partnership with the company. Together, the two funding rounds provide SambaNova with more than $1.35 billion in fresh capital this year as it seeks to expand production, accelerate customer deployments and compete in a market largely dominated by Nvidia.
Speaking to CNBC at the Raise AI Summit in Paris, SambaNova co-founder and Chief Executive Officer Rodrigo Liang said the rapid expansion of AI inference has fundamentally changed the industry’s growth trajectory.
“Inference has broken everything open, and so what we’re seeing now is that as a standalone company, you have the ability to really move fast and drive the business across a broad range of sectors,” Liang said.
He added that customer demand has been accelerating rapidly.
“We’re scaling the business really, really fast, and so the capital allows us to really accelerate the deployments of the racks that customers really want.”
Liang also revealed that SambaNova is actively considering an initial public offering in 2027, with the United States emerging as the most likely listing venue. The comments suggest the company hopes to follow a growing list of AI infrastructure firms preparing to access public markets as investor enthusiasm for artificial intelligence remains strong.
AI Inference Becomes The Industry’s Next Battleground
For the past several years across the AI semiconductor market, Nvidia’s graphics processing units (GPUs) have dominated AI spending because they are essential for training massive language models such as ChatGPT and other generative AI systems. Today, however, many technology companies are directing increasing attention toward AI inference, the stage where trained models generate responses for users.
Unlike model training, which is typically performed once using enormous computing clusters, inference occurs every time an AI assistant answers a question, generates an image or completes a task.
As AI adoption spreads across businesses and consumers, inference workloads are expanding rapidly and are expected to become one of the industry’s largest long-term semiconductor markets. That transition has opened opportunities for startups developing chips specifically optimized for inference rather than training.
SambaNova is among several companies attempting to capitalize on that shift by designing specialized hardware capable of delivering faster AI performance while consuming less power and lowering operating costs. The company’s latest processor, known as the SN50, is sold as part of a complete server system designed for deployment inside data centers.
Unlike Nvidia’s traditional business model, which primarily supplies GPUs that customers integrate into larger computing systems, SambaNova offers integrated hardware solutions combining chips, servers and supporting infrastructure. The company believes this approach simplifies deployment while improving performance for enterprise AI applications.
Betting On On-Premises AI
Another pillar of SambaNova’s strategy is on-premises AI inference, allowing businesses to run AI models inside their own data centers instead of relying entirely on cloud providers. The approach appeals particularly to organizations handling sensitive information, including banks, governments and healthcare providers, where data privacy, regulatory compliance and security remain critical concerns.
Running AI locally can also reduce network latency, allowing AI systems to respond more quickly because data does not need to travel to external cloud servers. The strategy received a significant endorsement on Wednesday when JPMorgan Chase announced plans to deploy SambaNova’s systems for on-premises inference across its enterprise AI workloads.
The bank said the technology would support demanding internal AI applications, highlighting growing interest among large financial institutions in retaining greater control over AI infrastructure. SambaNova notes that on-premises deployments provide customers with faster performance, stronger security and greater operational control compared with cloud-based AI services managed by third-party providers.
Investors Continue Backing AI Chip Challengers
The funding round comes amid sustained investor enthusiasm for AI semiconductor companies. Public markets have continued to reward businesses supplying the hardware underpinning the AI boom.
The PHLX Semiconductor Index, which tracks major chipmakers, has climbed roughly 80% this year, indicating strong demand for AI infrastructure and continued optimism surrounding semiconductor earnings.
The sector is frequently described as the “picks and shovels” of the AI revolution because chipmakers supply the essential hardware required to build and operate artificial intelligence systems, regardless of which AI applications ultimately succeed.
Investors are now increasingly backing startups that hope to capture portions of the AI hardware market currently dominated by Nvidia. While Nvidia remains the clear leader in AI accelerators, rising demand for inference has created room for alternative architectures designed specifically for serving AI applications efficiently.
SambaNova and others seem to be in a race to prepare for the next phase of competition.
Also on Wednesday, South Korean AI chip startup Rebellions told CNBC it plans to pursue an initial public offering on South Korea’s Kospi exchange during the first or second quarter of 2027, adding to the growing pipeline of AI infrastructure companies preparing to enter public markets.
The competition has continued to intensify across the sector. Last year, Nvidia itself signed a licensing agreement with AI inference startup Groq, highlighting the strategic importance of specialized inference technologies even for the industry’s dominant player.
SambaNova’s latest fundraising illustrates how the AI hardware race is evolving. The first wave of generative AI investment centered on acquiring enormous computing power to train frontier models. The next phase is increasingly focused on making those models practical, affordable, and efficient to operate at scale.
As businesses deploy AI assistants, autonomous agents and enterprise automation across millions of daily interactions, inference chips are becoming as strategically important as training hardware.
That shift is creating opportunities beyond Nvidia for investors. For startups such as SambaNova, it offers a chance to compete in one of artificial intelligence’s fastest-growing markets by targeting the operational workloads that are expected to dominate AI computing over the coming decade.
