doubleAI Raises $200M Series A From Lightspeed, Bessemer, Nvidia and Dell for Maxwell, a 'Code Execution Model'
Amnon Shashua's doubleAI announced a $200M Series A co-led by Lightspeed and Bessemer, with Dell Technologies Capital, Greenfield, NVIDIA, Pitango and BRM participating. It unveiled Maxwell, which it calls a Code Execution Model that optimises running binaries. Speed-ups are company claims; valuation and GPU-spend figures are second-hand.
Squeezing more work out of existing chips has become a fundable category. doubleAI, co-founded by Mobileye founder Amnon Shashua, announced a $200 million Series A co-led by Lightspeed and Bessemer, with Dell Technologies Capital, Greenfield, NVIDIA, Pitango and BRM participating, the company said on Business Wire.
It unveiled Maxwell, which it calls a "Code Execution Model" — AI that works inside running code to make it faster, rewriting the binary while leaving source unchanged, with every change "formally proven correct," doubleAI says. It claims more than 38x faster data pipelines and more than 4.5x faster inference kernels. Those speed-ups are company claims. SiliconANGLE adds that CTech put the valuation above $1 billion and that, per the Wall Street Journal, about half the round will pay for 1,000 Nvidia B300 GPUs; we did not read CTech or the Journal.
Limits
- Valuation and B300 spend are second-hand via SiliconANGLE. Formal-proof and speed-up claims are doubleAI's.