What happened

  • doubleAI disclosed a $200 million Series A funding round on 19 October 2026.
  • Lightspeed and Bessemer co-led the round; Nvidia Corp., Dell Technologies Inc. and other investors participated.
  • CTech reported a post-money valuation of more than $1 billion.
  • The Tel-Aviv-based firm unveiled its AI model Maxwell, which can automatically optimise software binaries and lower hardware usage.
  • Maxwell rewrote Nvidia's cuGraph library, speeding up 55% of its algorithms by at least a factor of two.
  • The model also accelerated pandas on the PDS-H benchmark by a factor of 38 and doubled video generation speed for the LTX-2.5 world model.
  • doubleAI plans to sell Maxwell to enterprises seeking to reduce AI workload costs.
  • The Wall Street Journal said half the funding will buy 1,000 Nvidia B300 Blackwell graphics cards.

Why it matters

Enterprises face rising costs and energy use from large AI models. Maxwell’s ability to optimise binaries without changing source code lets organisations improve performance while the application runs in production. The funding gives doubleAI the resources to acquire a large fleet of graphics cards and scale its service, putting it in a better position to compete with larger model providers such as OpenAI and Anthropic. By reducing the compute required for the same tasks, firms can extend the life of existing hardware and lower their carbon footprint.

What this means for enterprise IT teams

  • Teams can evaluate Maxwell as a drop-in optimisation layer that works on existing binaries, avoiding lengthy development cycles.
  • Deploying the model may reduce the need for additional graphics cards, helping to control capital expenditure.
  • The planned purchase of 1,000 B300 Blackwell graphics cards will increase processing capacity, but IT must plan for power and cooling requirements.
  • As regulators and cloud providers tighten controls around AI agents, solutions that optimise workloads without altering source code provide clearer audit trails and align with emerging compliance expectations.

In the words of doubleAI: "The model sped up 55% of the algorithms in the library by a factor of two or more."

The bottom line

doubleAI’s $200 million raise backs a technology that directly tackles enterprise concerns about AI cost, performance and regulatory pressure. With a large hardware investment and a model that improves efficiency at the binary level, Maxwell could become a key component of many organisations’ AI strategies.

Sources

  • siliconangle.com