Halfwild

Halfwild leases dedicated GPU servers to a small number of AI teams. The machines are ours, they sit in a data center in Tampa, Florida, and they run only your work.

Who we work with

Teams that keep GPUs busy most of the day: startups fine-tuning or serving open models, and research groups with steady training runs. Usually they have outgrown renting by the hour, but don't need hundreds of GPUs on a multi-year contract.

If you only need GPUs a few hours a week, an on-demand cloud will serve you better, and we'll tell you so.

How an engagement works

  1. A conversation. You tell us what you run and when you need it. We send a written quote within one business day.
  2. A trial week. You run your real work on an identical 8× A100 server for seven days, at no charge.
  3. A term. Month to month or twelve months, at a price fixed for the term. Your server is set up in Tampa and we move your work across with you.
  4. Renewal or exit. At the end of a term you renew, go month to month or leave. Nothing renews without your agreement.

How we work, in full

Tyler Waitt

When something needs attention, you write to a person who knows your servers by name. Usually that is me.

Tyler Waitt, Founder
[email protected] · (727) 519-3729