Positron AI, a chip startup based in Reno, Nevada, has raised $875 million in new funding at a post-money valuation of $5 billion, in an announcement reported on September 10, 2026. The money is intended to carry Positron’s next processor, called Asimov, from design into production. The company builds chips for AI inference, the stage at which a trained model generates responses, rather than the training of models that has driven much of the demand for Nvidia’s hardware.
The financing came in two parts. A $375 million Series C round valued the company at $3.5 billion before the new money, and a further $500 million Series C-1 followed. New Enterprise Associates and Jim Clark led the later tranche, with additional backing from Andra Capital, Atreides Management, Valor Equity Partners and SemiAnalysis Capital. The round followed a $230 million Series B in February 2026 that had valued Positron at $1 billion, a fivefold increase in valuation over roughly seven months.
Positron’s central design choice is to use standard LPDDR5X memory rather than the high-bandwidth memory, or HBM, that sits alongside most AI accelerators. HBM has been in short supply and commands high prices, and the company argues that building around more widely available memory can lower the cost of running large models. The approach is aimed at data center operators looking for an alternative to Nvidia’s graphics processors for inference workloads.
The Asimov chip is to be manufactured on TSMC’s N3P process and will carry between 288 gigabytes and 2,304 gigabytes of memory per chip, with a realizable bandwidth of about 2.76 terabytes per second and power draw near 400 watts. Positron says the design can deliver far more output per dollar than Nvidia’s GB300 systems on certain inference tasks, though the company notes those figures come from simulations because the chip has not yet been manufactured.
Positron plans to complete the Asimov design, a step known as tapeout, by the end of 2026, with production expected in the second half of 2027. The raise adds to a run of large investments in companies trying to build inference-focused silicon as spending on AI infrastructure continues to climb. Whether Positron’s claims hold will depend on how the chip performs once it reaches physical hardware and real workloads rather than internal projections.
