Gradient Platform is now Generally Available
DigitalOcean’s Gradient Platform (previously GenAI Platform) is now generally available, offering a fully managed, developer-first platform for building, scaling, and deploying AI-powered applications. It includes tools for creating intelligent agents, integrating top LLMs with a single API via serverless inference—no infrastructure management required. With powerful new features like external data integration, agent traceability, and evaluations, the Platform helps developers go from idea to production faster and with more control. Learn more ->
AMD Instinct™ MI325X GPU Droplets are now available
Unlock exceptional AI performance with new GPU Droplets powered by AMD. These GPUs provide 256GB of HBM3E memory and 6.0TB/s of bandwidth, enabling faster AI inference and the ability to handle massive models entirely in memory. Learn about the release on our blog and watch the DigitalOcean and AMD webinar for more information.
NVIDIA RTX 6000 Ada Generation GPU Droplets are now $1.57 GPU/hr
NVIDIA RTX 6000 Ada Generation GPU on DigitalOcean is now $1.57 GPU/hr, reduced from $1.89 GPU/hour. This GPU offers cost-efficient inference capabilities for graphical processing, rendering, virtual workstations, compute, and media & gaming. This versatility means you can tackle a wide array of demanding tasks, all while optimizing your budget. Unlock new possibilities with the NVIDIA RTX 6000 Ada Generation today.
ERNIE 4.5-21B 1-Click Model: is now available on DigitalOcean
Now you can deploy ERNIE 4.5-21B-A3B-Base, Baidu's powerful open-source model, as a DigitalOcean Gradient GPU Droplet 1-Click Model. Leverage ERNIE 4.5 21B’s robust performance in complex natural language processing tasks, advanced multimodal capabilities, and computational efficiency with instant deployment on cost-effective, scalable, GPU Droplets. Try it today to elevate your AI development with DigitalOcean GPU Droplets.
Deploy AI Faster: Introducing DigitalOcean's Optimized Inference Image
Deploying LLMs just got a whole lot easier with DigitalOcean's Optimized Inference Image for GPU Droplets. This pre-configured OS image eliminates the hassle of manual setup, allowing you to instantly access a production-grade environment with built-in optimizations like CUDA and FlashAttention. Now, you can go from Droplet launch to live inference — whether you’re deploying a single model or multiple models simultaneously on the same GPU.
Four Powerful, New Features to Help You Build and Deploy More Efficient Apps On DigitalOcean Kubernetes [DOKS]
This past month, we released four new Kubernetes features that enable you to build, deploy, and scale Kubernetes apps more efficiently on DOKS. They include:
- Support for new Droplet types: NVIDIA RTX 4000 Ada Generation GPU, NVIDIA RTX 6000 Ada Generation GPU, NVIDIA L40S GPU, and AMD MI300X GPU.
- Nodepool scale-to-zero: Cost-optimization tool that automatically scales down idle node pools to zero nodes.
- Support for DOKS in our new Atlanta datacenter [ATL1]: You can now deploy fully-managed Kubernetes clusters in DigitalOcean’s newest AI-optimized data center, ATL1 in Atlanta-Douglasville.
- Routing Agent now available: With support for Kubernetes custom resources, DOKS Routing Agent makes it easy to define custom routes, helping to simplify static route configuration in your Kubernetes clusters.
Learn more about the new features and get started by reading our blog post.
Now Available: Kafka Schema Registry
With Kafka Schema Registry, you can improve data integrity and simplify Kafka integration for event-driven applications through centralized schema management. By implementing this tool, you can eliminate data interpretation issues between producers and consumers, fostering a more reliable and efficient data flow. Learn more ->
