SAN FRANCISCO, Feb. 4, 2026, 09:18 (PST)
- Cerebras Systems secured $1 billion in a late-stage funding round, pushing its valuation to roughly $23 billion.
- The financing was led by Tiger Global, with Benchmark, Coatue, and 1789 Capital also joining in.
- Big AI customers are pushing to diversify their chip sources for “inference,” the stage when models respond to user queries, driving this deal.
Cerebras Systems announced it secured $1 billion in a late-stage funding round, pushing its valuation to roughly $23 billion—almost three times what it was just over four months ago. Tiger Global led the round, the company confirmed.
The fundraising highlights the rapid influx of capital into AI infrastructure, despite a more cautious stance from some venture investors in other areas. Chips, servers, and data centers have become the key bottleneck for launching new AI products, with investors seeing capacity as a limited resource.
Timing is crucial as major AI players aim to diversify away from a single supplier. Reuters revealed this week that OpenAI is exploring alternatives to Nvidia for AI “inference” chips—the specialized hardware that powers AI responses—bringing startups like Cerebras into the spotlight in this supply battle.
Cerebras announced a $1 billion Series H funding round, valuing the company at roughly $23 billion post-money. The investor lineup featured Benchmark, Fidelity Management & Research, Atreides Management, Alpha Wave Global, Altimeter, AMD, Coatue, and 1789 Capital, which is backed by Donald Trump Jr., among others.
This funding round marks Cerebras’ second billion-dollar raise since September, when Reuters reported its valuation at $8.1 billion. The sharp increase highlights just how eagerly investors are backing ventures linked to expanding AI computing infrastructure.
This marks Cerebras’ first fundraising effort since it pulled its U.S. IPO filing back in October. This trend isn’t unique—tech firms able to secure sizable private funding are delaying public offerings, at least while private capital is still flowing.
Cerebras, located in Sunnyvale, California, manufactures “wafer-scale” chips—processors that cover an entire silicon wafer instead of being sliced into smaller dies like typical chips. The company offers systems for training AI models as well as inference, the workload that’s gaining traction since it powers every chatbot query in real time.
The fresh funding boosts Cerebras as it takes on a market largely controlled by Nvidia’s data-center GPUs, with AMD also pushing to grab market share and several startups offering specialized inference chips. Reuters reported Nvidia reached out to Cerebras and Groq about possible acquisitions; Cerebras turned down the offers and instead sealed a commercial deal with OpenAI, revealed last month.
The move toward “Nvidia alternatives” isn’t a total departure. Nvidia stated this week that “Customers continue to choose NVIDIA for inference” thanks to its performance and cost-effectiveness at scale. OpenAI confirmed it still depends on Nvidia for most of its inference fleet. CEO Sam Altman even tweeted that Nvidia produces “the best AI chips in the world.” Reuters
For Cerebras, this funding round signals both trust and challenge. The company is banking on growing demand for faster inference and expects major customers to invest in diversifying their suppliers, despite Nvidia and AMD continuing to roll out fresh hardware and competitive pricing within the same data centers.
