Details on TensorOpera's Infrastructure Team-Up with Aethir

By Alex Passett August 13, 2024

Back in June, Silicon Valley-based AI company TensorOpera introduced Fox-1, its open-source small language model (SLM) with, according to the official press release, “1.6 billion parameters, outperforming other models in its class,” even models reportedly from industry giants like Google or Apple. And this certainly isn’t the company’s first rodeo, either; since early 2022, TensorOpera has provided GenAI-platform-at-scale solutions, offering users what they need to build and commercialize their own GenAI applications both easily and economically.

What’s more is that its Fox-1 features an innovative architecture that is, quote, “78% more advanced than many other models.”

This is what leads us to more recent news:

TensorOpera established a partnership with Aethir, a distributed cloud infrastructure provider with teams looking to spearhead the next evolution of cloud compute.

By partnering up, Aethir has equipped TensorOpera with the advanced GPU resources necessary for Fox-1’s additional training. Aethir's previous collaborations with NVIDIA Cloud Partners, Infrastructure Funds, and other enterprise-grade hardware providers has established a global, large-scale GPU cloud. This led (and still leads) to the delivery of cost-effective and scalable GPU resources; these are essential for high-throughput, substantial memory capacity, as well as efficient parallel processing capabilities.

Thus, with the support of Aethir’s decentralized cloud infrastructure, TensorOpera has the necessary tools for facilitating streamlined AI development that requires higher network bandwidth and ample amounts of GPU power.

“I am thrilled about our partnership with Aethir,” said Salman Avestimehr, CEO and co-founder of TensorOpera. “In the dynamic landscape of generative AI, the ability to efficiently scale up and down during various stages of model development and in-production deployment is essential. Aethir’s decentralized infrastructure offers this flexibility, combining cost-effectiveness with high-quality performance. Having experienced these benefits firsthand during the training of our Fox-1 model, we decided to integrate Aethir's GPU resources into TensorOpera's AI platform to empower developers with the resources necessary for pioneering the next generation of AI technologies."