NeuroFrame raises $42M Series B to scale real-time neural rendering for AR headsets

Tech · 4 min read

NeuroFrame, an eight-year-old startup focused on neural rendering, announced a $42 million Series B today to commercialize its low-latency rendering stack for augmented reality headsets. The round was led by Greylock with participation from NVIDIA and Waves Capital, and brings total funding to over $80 million.

The company’s pitch is hardware-agnostic neural upscaling and relighting that runs within tight power envelopes. NeuroFrame plans to ship a developer SDK supporting Unity and Unreal in the next quarter and has an early partnership to integrate with a major AR OEM’s upcoming reference design.

Investors noted that the funding will also accelerate NeuroFrame’s tooling for mixed-reality compositing and a cloud service for scene priors. NeuroFrame expects to expand its engineering team across graphics, compression, and mobile systems in the coming months.