Start your AI Journey

The use of FPGAs in AI inference is rather unique–Positron is utilizing these programmable chips to achieve a surprising result, better than enterprise-class GPUs both in raw performance per watt and cost. What’s more is the power savings give you multiple 8-card servers in under a 10 Kw rack versus the competition’s 6 Kw which would limit it to a single server in a rack. Watch this company as they further refine their performance. Looking for pricing? Contact us at BittWare.com/positron.

source