VDURA launches V12 as a multi‑tenant, API‑driven storage service for AI factories, scaling from 8 to 100,000 GPUs on Supermicro hardware.
NoteOossa1 min read
VDURA announced that its Data Platform V12 is now generally available. The new version runs on Supermicro Building Block Solutions and can support anywhere from 8 to 100,000 GPUs. It adds multi‑tenant, API‑driven storage services for GPU clouds and AI factories. The release turns the platform into a shared storage service that AI teams can use without building their own hardware stack.
It lets more AI developers use hyperscale‑grade storage without huge upfront hardware costs.
Tools & agents
CodeScene used AI agents and its CodeHealth MCP server to clean up a 300,000‑line C codebase of Street Fighter III in three weeks, at a token cost of about $4,000.
Tools & agents
AWS and OpenAI launched a limited preview of Bedrock Managed Agents, letting enterprises build and run OpenAI‑powered agents with AWS security and governance.
Tools & agents
On Feb 19 2026 Google added Gemini 3.1 Pro Preview, a new version of its Gemini 3 series with better tool handling and a custom‑tools endpoint.
Oossa · Newsletter
The week in AI, explained
Every Monday: the stories worth knowing, in plain language. Free, no spam.
| # | Source | Outlet | Date | Key takeaway |
|---|---|---|---|---|
| 1 | VDURA Data Platform V12 Now Generally Available, the Hyperscaler Storage Playbook to AI Factories and Neoclouds on Supermicro ↗ | Business Wire | Sep 30, 2026 | PITTSBURGH & ABU DHABI, United Arab Emirates--(BUSINESS WIRE)--VDURA, provider of high-performance data storage infrastructure for AI factor |
1 sources
