SambaNova Systems valued at $11B, raises $1B in Series F for inference chips
The AMW Read
SambaNova's $1B raise at $11B valuation (5x in 5 months) and JPMorgan adoption update the inference silicon player map and signal a structural shift in enterprise AI infrastructure, warranting cross.§D for the explicit $1B round and cross.§H for the inference chip focus.
SambaNova Systems valued at $11B, raises $1B in Series F for inference chips
AI chip startup SambaNova Systems has raised $1 billion in a Series F round led by General Atlantic, reaching an $11 billion valuation — a fivefold increase from its $2.2 billion valuation just five months ago. The round includes Seligman Ventures, T. Rowe Price, Capital Group, BlackRock, Intel Capital, SoftBank Group, Qatar Investment Authority (QIA), and Vista Equity Partners. The company also announced JPMorgan Chase as a new customer, deploying SambaNova's SN40L and next-generation SN50 systems for on-premises inference infrastructure to process sensitive financial data.
Why it matters: SambaNova's valuation surge — from $2.2B in February to $11B in July — and its $1B raise signal a structural shift in the AI compute market. The company's inference-optimized chips (SN40L/SN50) claim 5-10x decode performance over NVIDIA GPUs for large language models, directly challenging the GPU-centric paradigm. More importantly, JPMorgan's adoption of on-premises SambaNova systems for sensitive financial workloads validates a growing enterprise preference for private AI infrastructure over public cloud, a pattern that could reshape inference economics across regulated industries. The round also includes sovereign capital (QIA) and strategic investors (Intel Capital, SoftBank), reflecting the geopolitical dimension of inference silicon supply chains.
SambaNova's trajectory — from a $2.2B valuation in February to $11B now, after reportedly rejecting a $1.6B Intel acquisition offer last year — illustrates the capital-compression arc in inference silicon. The company claims its SN40L and SN50 chips deliver 5-10x decode performance over NVIDIA GPUs for LLM inference, reducing token costs by eliminating the need for massive GPU clusters. With SoftBank as the first deployment partner for SN50 and JPMorgan building on-premises inference infrastructure, SambaNova is positioning itself as the leading alternative to NVIDIA in the inference market, targeting sovereign cloud, AI-dedicated cloud providers (NeoCloud), and enterprise private infrastructure. The company is developing its fifth-generation chip and plans an IPO in the long term.
