Provider news · April 2026
What the providers shipped
April 2026 brought a smorgasbord of updates for DigitalOcean, but it's the avalanche of bling with their new AI-Native Cloud rebrand and the launch of multi-node GPU worker nodes for their already overpriced GPU plans that screams "who’s buying this?" Meanwhile, Linode stuck to incremental improvements with Kubernetes v1.32 support and Linode interfaces, barely raising an eyebrow. UpCloud, on the other hand, released Kubernetes v1.35, a nod to CNCF and AI Conformance, and offered the usual new server families to pad the catalog. Vultr rounded out the month with a flashy new enterprise portal and the introduction of Clusters, but it’s their low-cost archival storage that actually seems worth taking a look at. For developers, the incremental progress and over-the-top marketing battles don't change much on the ground.
DigitalOcean
- Apr 30 DOKS NVIDIA B300 GPUs (multi-node) — Multi-node GPU worker nodes available by contract.
- Apr 28 Vector Databases launch / Managed Weaviate private preview — New Data Services grouping; Weaviate private preview in TOR1.
- Apr 28 Remote MCP servers for Volumes and NFS — API-based management access for block storage and network file storage.
- Apr 28 PostgreSQL Advanced Edition (public preview) — Higher-performance managed PostgreSQL clusters in preview.
- Apr 28 MySQL Advanced Edition (public preview) — Higher-performance managed MySQL clusters in preview.
- Apr 28 Launchpad Starter Kits GA — Preconfigured application stacks now generally available.
- Apr 28 Knowledge Bases GA — Managed RAG knowledge bases GA with reranking, hybrid retrieval, RAG playground.
- Apr 28 Data Services GA — Managed databases and knowledge bases generally available.
- Apr 28 Managed Databases in control panel — Centralized management interface for all database engines.
- Apr 28 Dedicated Inference GA — Managed LLM hosting service generally available.
- Apr 28 Rebrand to AI-Native Cloud — Gradient/AI Platform, GPU Droplets and Bare Metal GPUs product names updated.
- Apr 27 Container Registry larger images — Supports container images up to 100 GB and layers up to 20 GB.
- Apr 27 Remote MCP server for Container Registry — API-based registry management access.
- Apr 24 App Platform request-based autoscaling (preview) — Traffic-based scaling for service components in public preview.
- Apr 15 AMD Instinct MI325X GPUs in TOR1 — New GPU Droplet option available by contract.
- Apr 13 Droplet snapshot transfer between accounts — Snapshots transferable between accounts via API and control panel.
- Apr 9 NAT Gateway bandwidth increase — VPC NAT Gateway throughput raised to 2 Gbps per size.
- Apr 8 Spaces over private VPC network — VPC-local internal DNS can reach Spaces over the internal network.
Linode
- Apr 21 LKE updates — Updates and improvements to the Kubernetes Engine service.
- Apr 10 LKE Enterprise Kubernetes v1.32 support — LKE Enterprise adds support for Kubernetes v1.32.
- Apr 9 Linode interfaces and firewall templates — New Linode network interface and firewall template capabilities.
- Apr 7 Cloud logs updates — Enhancements and new features for the cloud logging service.
UpCloud
- Apr 28 MySQL v8.4 LTS now available — MySQL 8.4 Long-Term Support version available on Managed Databases.
- Apr 16 Kubernetes v1.35 now available — Kubernetes 1.35 released with CNCF and AI Conformance certification.
- Apr 15 Starter & Premium servers available — Two new Cloud Server plan families launched for different workload types.
Vultr
- Introducing Clusters — Deploy and manage groups of compute as one unit with automated fabric networking.
- Milan Region Now Available — New Milan data center expands the European footprint with lower latency and capacity.
- Introducing console.vultr.com — New enterprise portal with built-in IAM: policies, roles, groups, service accounts, orgs.
- Archival Object Storage Tier Released — Low-cost $0.006/GB-mo archival tier with lifecycle management and on-demand retrieval.
- Vultr Serverless Inference Improvements — KV caching, tool-calling, throughput gains, and usage-based per-token billing.
- NVIDIA Build Partnership — Vultr Serverless Inference models now accessible via build.nvidia.com.
- Load Balancer Scaling Fix — Fixed load balancers scaling down to zero nodes for consistent availability.
- UDP Support for Load Balancers — Load Balancers now support UDP for gaming, streaming, and DNS alongside TCP.