Rackscale AI Accelerators
- Exponential Industry
- Technology Area
- 11 Milestones
Fully co-designed rack- and POD-scale AI systems that combine accelerators, host CPUs, scale-up fabric (NVLink-class domains), and scale-out networking in one thermal/mechanical package. Lineage runs from multi-GPU HGX/DGX boxes through Grace Hopper NVL32 domains to liquid-cooled 72-GPU racks (GB200, GB300, Vera Rubin) and competing racks such as AMD Helios. NVIDIA, AMD, and OEM rack builders (Dell, Supermicro) ship the current generation for AI-factory halls.
Hidden Factory
Rackscale AI Accelerators
NVIDIA and six institutions launch AI-factory financing platforms (>$500B)
NVIDIA announced MOUs with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR to set up independent compute-financing platforms designed to mobilize over $500 billion of third-party capital for AI-factory buildout. NVIDIA said the figure is not NVIDIA revenue, a single fund, or one customer commitment; partners independently underwrite each deal. Jensen Huang said NVIDIA may offer residual-value support on up to 25% of an opportunity. Final agreements were still outstanding /NVIDIA/ /Jensen Huang / X/.
AMD Helios rack-scale AI launch
AMD launches Helios rack-scale AI infrastructure (72× MI455X) at Advancing AI 2026 (AMD blog).
CoreWeave measures Vera Rubin NVL72 10x tokens per megawatt vs GB200
CoreWeave published first measured silicon performance on Vera Rubin NVL72. At a matched interactivity target, the rack generated 10x tokens-per-second per megawatt versus NVIDIA GB200 NVL72 on the same DeepSeek R1 workload, using TensorRT-LLM and Dynamo with NVFP4, expert parallelism, multi-token prediction, and disaggregated prefill/decode /CoreWeave/.
Intel–SambaNova–Foxconn rackscale AI infrastructure (Computex 2026)
intent and production-ready demos of rackscale AI infrastructure for inference/agentic workloads on Xeon + SambaNova SN-50 RDUs with Foxconn system integration /Intel Newsroom/.
CoreWeave completes first Vera Rubin NVL72 bring-up
CoreWeave completed the industry's first bring-up and validation of NVIDIA Vera Rubin NVL72 on CoreWeave Cloud, standing up the rack-scale system with power, cooling, networking, and compute verified and running /CoreWeave/.
First Dell PowerEdge XE9812 Vera Rubin NVL72 delivery to CoreWeave
Michael Dell announced delivery of the world's first working liquid-cooled Dell PowerEdge XE9812 NVIDIA Vera Rubin NVL72 server rack for CoreWeave (X, May 31, 2026).
Nebius achieves NVIDIA Exemplar Cloud on GB300 NVL72 training
Nebius reported NVIDIA Exemplar Cloud on GB300 NVL72 for training, meeting the 95% threshold across NVIDIA Performance Benchmarking recipes. Nebius said it is among the first clouds to hold Exemplar on GB300 and across multiple GPU generations (after H200) /Nebius/.
NVIDIA Vera Rubin platform announcement
NVIDIA announces Vera Rubin platform and related racks at GTC; seven chips in full production.
NVIDIA invests $2B in Nebius and partners on full-stack AI cloud
NVIDIA and Nebius announced a strategic partnership. NVIDIA will invest $2 billion in Nebius and support early adoption of Rubin, Vera CPUs, and BlueField so Nebius can deploy more than 5 GW of NVIDIA systems by the end of 2030 /Nebius / NVIDIA/.
NVIDIA Blackwell platform and GB200 NVL72 announcement (GTC 2024)
GTC 2024 (Mar 18) launch of the NVIDIA Blackwell platform: Blackwell GPU architecture, GB200 Grace Blackwell Superchip, liquid-cooled GB200 NVL72 rack-scale system (72 Blackwell GPUs + 36 Grace CPUs on fifth-gen NVLink), HGX B200, and broad cloud/OEM adoption commitments. Foundational rackscale-AI milestone that established the NVL72 rack product line later continued by GB300 and Vera Rubin NVL72.
NVIDIA DGX GH200 AI supercomputer announcement
DGX GH200 large-memory AI supercomputer class with Grace Hopper Superchips and NVLink Switch System—predecessor NVLink domain architecture before Blackwell NVL72.