
Vathys builds deep learning processors that address the data movement bottleneck in current deep learning architectures. By focusing on the data movement rather than just compute, Vathys offers a unique solution that is approximately 10 times better than existing deep learning processors. Their innovative architecture is designed specifically for deep learning applications, positioning them as a differentiated player in the market with significant potential for scale and traction.

Vathys builds deep learning processors that address the data movement bottleneck in current deep learning architectures. By focusing on the data movement rather than just compute, Vathys offers a unique solution that is approximately 10 times better than existing deep learning processors. Their innovative architecture is designed specifically for deep learning applications, positioning them as a differentiated player in the market with significant potential for scale and traction.
What they do: Design energy-efficient deep-learning processors that reduce data-movement overhead
Founded: 2015
Headquarters: Portland, Oregon
YC: Y Combinator Winter 2018
Reported total funding: $3.37M (disclosed)
Deep learning hardware — reducing data-movement overhead in neural network acceleration
2015
DeepTech
Latest disclosed seed round date listed as 2018-09-01; multiple seed rounds reported in 2018.
“YC W18, investors reported include Haystack, Brightstone Venture Capital, DCVC, FundersClub, MAGIC Fund”