About AiSilicon

Our dream is to build GPU computing tools that help people get work done, without wasting money.

We noticed that there was a gap in the GPU compute market between the more affordable hosts that don't provide many tools to help you, and the more premium platforms with proprietary features that subtly lock you in. We decided to build our own platform that fills this gap by providing on-demand GPU compute resources that are affordable and simple to use.

  • The issue with other affordable GPU providers

    Almost all other affordable GPU providers provide a very bare-bones "rent a server with a GPU" solution that, even if priced hourly, drives users to keep the servers they rent running 24/7 to avoid the overhead of setting things back up each and every day. These servers typically end up sitting idle for large amounts of time, burning money unnecessarily.

  • The issue with other premium GPU compute platforms

    Other premium GPU compute platforms cost substantially more than alternatives, and provide proprietary tools that are very effective at locking you in to using their platforms long-term, unless you want to make substantial changes to the way you work.

Where is the provider that is actually affordable in practice, easy to use, and doesn't lock you in with proprietary tools?

AiSilicon is the platform that fills this gap.

As a company founded by engineers, we knew that we could derive both power and flexibility from simplicity. Whether you're a team of seasoned machine learning professionals, or just an enthusiast tinkering with exciting new technologies, we think we can help you.

On this platform, you simply use the open-source tool git to push your code to AiSilicon, which starts GPU compute jobs on Docker containers in our secure Tier 4 datacenter automatically. Have your jobs write their output straight to the disk, and then use git again to pull the output back to your computer after a job completes. If you're curious about the details, check out our documentation to get the full picture.

You're only billed for the time your jobs spend running, plus a flat storage fee at a rate of $0.10/GB per month.

Reasonable rates, zero friction, and no lock-in. It really is that simple.