DistrAi connects AI companies with idle GPUs in universities, research institutes and gamers' PCs. Each job is matched to the right machine and is designed to run in isolation, and the owner is paid.
Global AI compute market, growing 25%+ a year
Typical saving against AWS and Google
Running today on gamer GPUs, with institutional pilots next
Goal: large models split across institutional GPUs. First milestone: 70B
A share of our income supports Israeli reserve soldiers, and reservists have priority to work with us.
Read moreDistrAi is a marketplace with two sides. Owners of idle GPUs supply compute, and AI companies use it. We sit in the middle: we match, we isolate and we pay.
Institutions and gamers register idle GPUs and decide when they are available.
Our scheduler sends each job to the machine that fits it best, by hardware, reliability, location and price. Every job is designed to run isolated from the owner's computer.
AI companies pay for the hours they use. Owners are paid automatically for the hours their GPUs run.
Gamers go online at random hours. Universities, call centers and offices work fixed hours, so we know when their machines are free, and capacity becomes something a customer can plan around.
The customer describes the job and we choose the hardware. Not the biggest machine, the right one, so there is less waste and a lower cost.
Some models are too big for one GPU. Our scheduling algorithm is designed to split them across several machines on one institutional network, using the fewest machines that fit and the fastest links between them. This is in development.
Rent GPU compute by the hour and choose a reliability tier: institutional hardware with a 99.9% uptime target, or our lower-cost community network. Typically 40 to 80% below AWS and Google.
Pricing and cost estimateUniversities, colleges and research institutes earn from GPU hours they are not using. You choose which GPUs and which hours, researchers keep priority, and you receive 70% of what customers pay.
Request a pilotTurn an idle graphics card into income while you are not playing. What you earn depends on your card's memory and the current rate.
See what you could earnAn institution will only share its machines if it is safe. So the process starts with their IT team, and the owner stays in control.
Jobs are designed to run isolated from the owner's computer, so the customer does not see the owner's machine and the owner does not see the job.
With institutions we install nothing until their IT team has reviewed the software and approved the scope.
Owners choose which GPUs and hours to offer and can stop at any time. Our jobs are designed to yield to the owner's own work.
Audits and SOC 2 readiness are part of our plan. We will not claim a certification before we hold it.
We are an early-stage company. This is what exists and what is planned.
A working proof of concept that sends AI jobs to GPU machines on our network.
A first AI company runs models of up to 10 billion parameters on the network, followed by a first institutional pilot, targeted for early 2027.
Large models split across institutional GPUs: first 70B parameters on a campus cluster, later 400B on institutional HPC clusters.
People who have built, run and sold technology companies, and who work in AI and large-scale data.
20+ years leading R&D, NLP, machine learning and large-scale data systems, including analytics supporting 50+ billion requests per day. Former CTO and VP of Data Science at Leadspace, and CTO at Amenity Analytics (acquired by Symphony).
PhD in artificial intelligence. 25+ years in AI, machine learning, NLP and data science. Has built enterprise-scale AI platforms for intelligence and analytics.
Full Professor and Head of Data Science at the Hebrew University Business School. Expert in text and data mining, NLP, algorithmic trading and sentiment analysis. Co-founded ProntoNLP, acquired by S&P Global.
Founder and CEO of Bring, a Web3 rewards platform. Co-founded eXelate, acquired by Nielsen for about $200M (2015). Founded Oridian and Freshub, acquired by Brightcom and IKAN.
Reserve soldiers carry a heavy load for the country, and then go back to work and to their families. We built this into the company.
Part of DistrAi's income goes to supporting Israeli reserve soldiers.
Reservists have priority when we hire and when we assign work in the company.
Choose the page that fits you. Each one explains the next step.