Intel Backs SambaNova With $350M as Tan Juggles Dual Leadership Roles

Intel Backs SambaNova With $350M as Tan Juggles Dual Leadership Roles

Intel is investing in SambaNova Systems as part of a $350 million funding round, deepening a relationship that puts the chipmaker's CEO in the unusual position of simultaneously running one company while chairing another that competes in the same market.

The partnership, announced Tuesday, comes as Intel attempts to claw back relevance in the artificial intelligence chip market that Nvidia has dominated since ChatGPT's launch in late 2022. Under the multiyear collaboration, SambaNova—which makes chips specifically for running generative AI models—will adopt Intel's server processors and graphics cards in its systems.

The arrangement is particularly notable given the overlapping leadership: Lip-Bu Tan, who became Intel's CEO, has served as SambaNova's chairman since 2017 and was an early financial backer of the startup. Intel itself first invested in SambaNova back in 2019, years before Tan took the helm at the struggling semiconductor giant.

For Intel, the deal represents a pragmatic acknowledgment of its position in the AI chip hierarchy. Rather than going head-to-head with Nvidia's dominance in graphics processing units—the silicon that powers AI model training at companies like Anthropic and OpenAI—Intel is effectively betting on multiple horses. If SambaNova's technology gains traction, Intel benefits both as an investor and as a supplier of the underlying server infrastructure.

The startup is making aggressive performance claims about its new SN50 chip, saying it outperforms the graphics cards in Nvidia's B200 systems. Whether enterprise buyers—who have spent the past two years standardizing on Nvidia's architecture—will be convinced remains an open question.

The timing is revealing. Intel has posted four consecutive years of revenue declines while Nvidia has become the world's most valuable publicly traded company on the back of AI demand. Intel's graphics card business, despite years of investment, has failed to meaningfully challenge Nvidia's stranglehold on AI workloads.

The dual-role arrangement raises questions about how Tan navigates potential conflicts between Intel's own AI ambitions and SambaNova's competitive positioning. Intel is developing its own AI accelerators—Gaudi chips that directly compete with Nvidia's offerings. Now its CEO chairs a company that also makes competing AI silicon, albeit with a different architectural approach.

For CFOs tracking the AI infrastructure landscape, the partnership signals that Intel is pursuing a hedging strategy rather than betting everything on its in-house development. The company appears to be acknowledging that the AI chip market may support multiple winners, and that owning a piece of various approaches—through investment, partnership, and internal development—may be smarter than trying to replicate Nvidia's success directly.

The $350 million round itself suggests SambaNova is still burning capital to compete in a market where Nvidia's first-mover advantage and software ecosystem create formidable barriers to entry. Whether Intel's backing provides enough runway for SambaNova to gain meaningful market share will likely depend on whether enterprises are willing to diversify away from Nvidia's de facto standard architecture.