GPUs and beyond: DigitalOcean’s roadmap for AI platforms and applications

At DigitalOcean, we’re not just observing the AI revolution—we’re actively shaping it.

Last year we made a key acquisition to expand the AI capabilities of our platform, broadening access to AI/ML development tools that were once the domain of large enterprises. In July 2024 at Deploy, we unveiled parts of our new AI offerings: GPU Droplets, empowering users to harness unprecedented computational power for AI/ML workloads, and our GenAI platform, a comprehensive solution for deploying AI agents and fine-tuning models. These innovations are a leap forward in making advanced AI tools accessible to businesses of all sizes.

Today, I’m excited to share more details on our future roadmap and DigitalOcean’s guiding philosophy for advancing AI capabilities for our users, along with an update on the future of the Paperspace brand and products. Despite the unprecedented pace of AI, its evolution is still in its infancy and so we are just at the start of unlocking AI capabilities for our customers.

From AI infrastructure to AI platforms to AI applications

Technology has often followed a path of democratization, moving from the fingertips of a few to the hands of many. The growth of the Internet followed this trajectory. In 1993, the introduction of Mosaic, the first widely-used web browser, created a new platform for digital interaction. This and subsequent browsers, were the platforms for the development and distribution of web-based applications. As a result, millions could now engage with the digital world, from online encyclopedias to early social networks to e-commerce shopping sites.

The iPhone and subsequent mobile technologies introduced advanced platforms that transformed smartphones from luxury items into powerful, accessible devices. These platforms enabled the creation of apps and app stores, creating an ecosystem where developers could easily create and distribute applications. Consequently, consumers gained access to an array of tools and services—from navigation apps and mobile payment platforms to on-demand ride services—that fundamentally changed how we interact with mobile technology.

In both cases, the progression from an infrastructure to platform to application layer made these technologies increasingly accessible to everyday users.

We anticipate a similar evolution with artificial intelligence: from foundational models developed by companies like Anthropic, Cohere, and OpenAI, to platforms that will enable a broad range of developers to build AI-powered applications, and ultimately to applications that will allow end-users to interact with generative AI tools in their daily lives. This progression will extend AI’s reach from specialized research labs to businesses and individuals worldwide. Our aim is to play a key role in accelerating AI democratization.

From simple cloud computing to accessible AI

At DigitalOcean, we’re expanding our horizons beyond infrastructure. Rather than just providing the underlying infrastructure, we’re building solutions that span the entire stack so that our customers can harness GenAI without needing deep expertise or needing to build GenAI capabilities themselves. This aligns with our aim to serve the broadest possible audience of developers, startups, and fast-growing businesses. By extending our focus to these layers, we’re positioned to make AI more accessible and useful for those who have traditionally been underserved by enterprise-focused solutions from hyperscalers.

This is a natural progression of our work to date.

One of our core values at DigitalOcean is “DO Simple.” We prioritize simplicity and ease of use, making app development easy and ensuring our tools are never the blockers to innovation. Droplets let you deploy and scale virtual machines with ease, Spaces object storage allows you to store and serve large amounts of data effortlessly, and App Platform lets you build and run applications without managing complex infrastructure. These innovations, to name a few, have enabled countless developers and businesses to build, run, and scale applications in the cloud.

Now, we’re applying this same philosophy of simplicity to AI, focusing on key areas including GPUs to supercharge AI/ML workloads and a GenAI platform to accelerate app building.

Available now: GPUs to supercharge AI/ML workloads

Our primary focus is on making AI accessible to a broad range of users, and we’re starting with infrastructure—providing innovators with the building blocks for AI innovation. NVIDIA H100 GPUs are available for use today, with customers like Lepton AI and Supermaven leveraging these powerful GPUs to train their models and scale their businesses.

Additionally, our GPU Droplets, launching in October, will make it simple for anyone to access on-demand GPU servers for AI and machine learning tasks. GPU Droplets offer NVIDIA H100 GPUs, delivering exceptional performance for training and inference on AI/ML models, processing large datasets, and handling complex neural networks for deep learning use cases. With options ranging from single GPU configurations to high-capacity 8-GPU setups, users can scale their AI projects as needed.

If you’re a startup developing GenAI powered applications, training ML models, or other compute-intensive tasks, our GPU Droplets provide the compute power you need. These resources allow you to train large language models on large amounts of text data, reducing the time required to develop and refine your AI solution. Instead of investing in expensive on-premises hardware, our cost-effective, on-demand GPU resources allow you to iterate quickly and deploy your AI solution faster.

DigitalOcean Kubernetes is also introducing support for H100 GPUs, so that customers can deploy AI/ML workloads and other resource-intensive tasks on Kubernetes in both single-node and multi-node setups.

Coming Soon: GenAI platform to accelerate app building

Our GenAI platform aims to democratize AI application development, making it accessible to developers and startups without deep AI expertise. It will provide pre-built components like hosted LLMs, data ingestion pipelines, and knowledge bases, allowing our customers to easily create AI-powered applications. This platform is designed to bridge the gap between Gen AI’s promise and deployed GenAI applications, focusing on simplicity and ease of use.

For instance, imagine you’re running a ticketing platform for live entertainment venues and want to improve your client services with AI. With our GenAI platform, you could easily create a chatbot by simply uploading your in-house documentation and pointing the GenAI pipeline to it. Instead of grappling with complex AI infrastructure, you’d be able to focus on customizing the chatbot to your specific needs.

Folding Paperspace into DigitalOcean

As we launch DigitalOcean’s new AI products, we also want to provide an update on the Paperspace brand and existing Paperspace products to its users. In the coming weeks, we will fold the Paperspace brand and public-facing website into DigitalOcean, as we work towards a unified AI/ML product suite. Our teams are already operating as one, and this is the first step to bringing a simpler experience to those looking to leverage our AI/ML products.

Existing Paperspace customers can continue to access their products through the same console they currently use.

We’re building an AI roadmap with our customers in mind

I joined DigitalOcean to lead our teams into a new era, with a focus on simplifying cloud computing and harnessing the power of AI for developers and growing technology businesses. We’ve unveiled part of our AI vision, but the full picture is still taking shape. Behind the scenes, we’re working on additional platforms and applications that will make AI more accessible and empower the next generation of builders to turn their boldest ideas into reality. We’re building that AI roadmap with our customers in mind.

While formulating our AI strategy, we reached out to hundreds of DigitalOcean customers to understand their needs and aspirations. We heard from startups excited about lowering the barrier to entry for AI, from businesses looking to create chatbots to improve productivity, and from developers seeking AI-assisted cloud management tools. Our approach is grounded in these real-world insights.

We’re committed to developing AI-powered tools that will genuinely change the way our users interact with and leverage cloud technologies and AI. From Supermaven to Nomic AI to Lepton AI, we already have the next generation of AI innovators building on DigitalOcean. We’re looking forward to welcoming more of you.