AI technology is no longer limited to large enterprises; it is increasingly being adopted by small and mid-sized businesses (SMBs) as well. DigitalOcean, a cloud service provider with a developer-focused approach catering to SMBs, has now expanded its offerings by introducing support for the industry-leading Nvidia H100 GPU AI accelerator.
Enabling AI for SMBs
Paperspace, a platform that provides access to GPUs and offers a data science platform called Gradient, is now hosting the Nvidia H100 GPUs. This virtualization approach makes these highly sought-after GPUs more accessible to developers. The H100 GPUs offer exceptional performance, significantly reducing training and inference times for AI models.
“For these companies, when they need very powerful GPUs, the way things stand now it’s actually quite tough to get access to them and so they find their own business growth gets restricted right and that’s where we come in,” says Kanishka Roychoudhury, GM of AI/ML at DigitalOcean.
While not every small business or developer needs a full H100 GPU, the Paperspace platform allows for provisioning a fractional amount or utilizing the full power of the H100 with its eight separate GPU instances.
Expanding Possibilities
SMBs are already leveraging the DigitalOcean Paperspace platform to build various AI applications beyond just chatbots. One example is the development of a video augmentation service for translating videos between languages. The introduction of the Nvidia H100 support is expected to further extend and diversify the range of applications being built.
“We want to actually make this very easily available and affordable to the smaller companies,” emphasizes Roychoudhury.
As for the usage of the H100 GPUs, they are primarily expected to be utilized for AI model training due to their high performance. However, as needs evolve, they may also be deployed to support inference workloads.
Integration and Future Plans
The acquisition of Paperspace by DigitalOcean is still in progress, with integration efforts ongoing. The two platforms already work together, allowing GPU resources to be deployed via Paperspace while accessing storage resources on DigitalOcean compute. In the near future, the H100 GPUs will not only be available as infrastructure resources but will also be integrated into the Paperspace Gradient data science platform, providing data scientists and engineers with a comprehensive environment for building, training, testing, and deploying AI models.
“We want to actually make this very easily available and affordable to the smaller companies,” emphasizes Roychoudhury.
DigitalOcean’s overall goal is to reduce the friction involved in accessing high-end AI tools and hardware like the Nvidia H100, making these resources readily available and affordable for smaller companies.
