Releasesby GlyphInference Router now uses cache-aware routing to maximize prompt cache reuse and automatically applies prompt...
Inference Router now uses cache-aware routing to maximize prompt cache reuse and automatically applies prompt caching to eligible Anthropic requests. You can opt out of prompt caching for supported models with X-Model-Affinity: none and …
Also shipped Aug 19, 2026DigitalOcean in August 2026 →
Data movement policies (General availability)
With this release, data movement policies are now generally available for Enterprise Edition (or higher). Data movement policies are part of Snowflake’s data exfiltration protection controls. These SQL-style conditional policies can help you limit, alert on, or block data movement operations based on specific movement types.
Organization Features in Organization Command Center (Preview)
Organization Features is now in preview in Organization Command Center. From the organization account, a user with the GLOBALORGADMIN role can open Organization Hub » Command center , then the Organization Features tile, to see which organization-level capabilities are enabled. In preview, the main control is premium views for cross-account telemetry.
Organization Command Center 3rd party access configuration (Preview)
3rd party access configuration is now in preview in Organization Command Center. From the organization account, a user with the GLOBALORGADMIN role can open Organization Hub » Command center , then the 3rd party access configuration tile, to classify member accounts as internal or external, set the default tenant type for new accounts, and maintain allowed...
Role-based access control (RBAC) is generally available
Role-based access control (RBAC) is now generally available. RBAC lets users assume a role in Databricks. When a user assumes a role, only that role's permissions apply, rather than the user's accumulated permissions. This enables exclusive access.
ai_extract precision mode is now generally available
You can now use ai_extract precision mode for advanced extraction workloads, including multi-page documents, schemas with 50 or more fields, many extracted items per document, and complex conditional field logic. See ai_extract function .
[Launched] Generally Available: Azure Databricks Lakebase in four additional regions
Lakebase is now generally available in four additional Azure regions: North Central US, France Central, Germany West Central, and East Asia. This expansion provides customers with additional deployment options and broader regional availability for Lakebas
More from DigitalOceanEvery DigitalOcean release →
Qwen3.8-2.4T-A95B model available on DigitalOcean Inference has been upgraded to Qwen3.8-Max
Qwen3.8-2.4T-A95B model available on DigitalOcean Inference has been upgraded to Qwen3.8-Max. Qwen3.8-Max supports vision and video input and a 1M-token context window. For more information, see the Available Models page.
The following open-source models are deprecated from DigitalOcean Inference as of 18 August 2026: Llama 3.3...
The following open-source models are deprecated from DigitalOcean Inference as of 18 August 2026: Llama 3.3 Instruct-70B DeepSeek R1 Distill Llama 70B Qwen3-32B Qwen3 Coder Flash Migrate Llama 3.3 Instruct-70B to Llama 4 Maverick 17B 128E …
Updated NVIDIA AI/ML Ready (gpu-h100x1-base, gpu-h100x8-base) Droplet base images are now available in the...
Updated NVIDIA AI/ML Ready (gpu-h100x1-base, gpu-h100x8-base) Droplet base images are now available in the Control Panel and through the API. The images are based on Ubuntu 24.04 (upgraded from Ubuntu 22.04) and include NVIDIA driver …
The following DeepSeek model is now available on DigitalOcean Inference for serverless inference, Agent...
The following DeepSeek model is now available on DigitalOcean Inference for serverless inference, Agent Development Kit, and agents: DeepSeek V4 Pro 0813 For more information, see the Available Models page.
v5 Droplet configurations are now available in the atl1, ric1, mkc1, and mem1 datacenters
v5 Droplet configurations are now available in the atl1, ric1, mkc1, and mem1 datacenters. They run on 5th Generation AMD EPYC processors for higher per-core performance and let you size compute, memory, storage, and networking …
DigitalOcean Kubernetes (DOKS) now supports Dynamic Resource Allocation (DRA) drivers for NVIDIA and AMD GPU...
DigitalOcean Kubernetes (DOKS) now supports Dynamic Resource Allocation (DRA) drivers for NVIDIA and AMD GPU node pools in public preview, starting with DOKS 1.36.3-do.0. DRA is an opt-in alternative to the managed GPU device plugins. …
Cloud Security Posture Management (CSPM) paid plans support Managed Rules, which let you enable or disable...
Cloud Security Posture Management (CSPM) paid plans support Managed Rules, which let you enable or disable CSPM rules for your entire team. To exclude specific resources while keeping a rule enabled for the rest of the team, suppress …
The following Alibaba model is now available on DigitalOcean Inference for serverless inference, Agent...
The following Alibaba model is now available on DigitalOcean Inference for serverless inference, Agent Development Kit, and agents: Qwen3.8-2.4T-A95B For more information, see the Available Models page.