AMNON SHASHUA DOUBLEAI LAUNCHES AN AI THAT SPEEDS UP CODE WHILE IT RUNS

What happened: Amnon Shashua's startup doubleAI launched Maxwell, an AI model that rewrites machine code inside running software to cut computing power use. It opened for early access on Monday. Key numbers: Data pipelines more than 38 times faster; AI inference kernels more than 4.5 times faster; one early customer's compute costs cut by 67%; LTX model inference doubled; Why it matters: doubleAI plans to charge a share of customer compute savings, and Shashua expects AI labs with large compute bills to benefit most. LTX said halving GPU time directly affects serving costs and margins.

Amnon Shashua's startup doubleAI launched Maxwell, an AI model that rewrites machine code inside running software to cut computing power use. It opened for early access on Monday.

doubleAI plans to charge a share of customer compute savings, and Shashua expects AI labs with large compute bills to benefit most. LTX said halving GPU time directly affects serving costs and margins.