The company's platform is designed to make it possible to design personalized AI chips in the cloud with just a single click. With this approach, says the company, companies are able to build and deploy AI solutions in weeks instead of 12 to 18 months, and at a fraction of the cost of traditional hardware development.
"With AI model complexity doubling every three and a half months, the current way of building the AI infrastructure is not sustainable," says Jovan Stojanovic, co-founder and CEO of Anari AI. "We believe that every company deserves to have a personalized chip for their AI in just one click. We're supporting the rapid progress of AI in ways that haven't been possible until now."
The platform is built on a unique Python-based programming framework - developed by the company for chip design - that opens up new possibilities for hardware engineers and turns software engineers into hardware designers, says the company. The first configurable AI cloud chip from Anari is Thor X, which delivers 100x more efficient processing of 3D Point Cloud/Graph data structures compared to a high-end GPU.
The funding round was led by Earlybird, a venture capital investor focused on European technology companies, with participation from Acequia Capital, Serbian Entrepreneurs, and Eric Ries, author and founder of The Lean Startup.
"AI continues to be growing in size and complexity, and this rate of growth is outpacing standard CPU and GPU performance gains," says Roland Manger, co-founder and Partner at Earlybird. "Dedicated designs or flexible architectures are expensive, time-consuming and require hardware skills (in short: take lots of effort). Anari is a game changer in two ways: it removes the trade-off between performance and effort, and it makes AI-problem-specific chip solutions available instantaneously in the cloud."
Manger and Sasha Ostojic, ex-VP at NVIDIA, ex-SVP at Cruise Automation, Advisor at Samsung Electronics and Zoox, and Operating Partner at Playground Global, will be joining Anari's