On this page · 9 sections
Summary. NVIDIA said in January 2026 that Rubin was in full production, said at GTC Taipei on 31 May that the platform was ramping into full production, and told investors on 26 August that production shipments had begun earlier that month. CoreWeave announced availability on 30 September, Oracle Cloud announced general availability on 27 October, and Google Cloud's A5X instances are due in November. NVIDIA's March release also names AWS, Microsoft Azure, Crusoe, Lambda, Nebius, Nscale and Together AI. The jump is large. A Rubin GPU delivers 50 petaflops of FP4 inference, 5 times Blackwell, and 35 petaflops for training, 3.5 times Blackwell, with HBM4 memory bandwidth of 22 terabytes per second at launch, 2.8 times higher, though NVIDIA's current product page lists 19.2. NVIDIA says the platform cuts cost per token up to 10x versus Blackwell for large mixture-of-experts inference, the economics that sit under model prices like $1 to $6 per million tokens. For infrastructure and ML platform teams, the question is not whether Rubin is faster but which workloads justify moving and when capacity will actually be reachable.
Update, 2 November 2026. This article originally said Vera Rubin entered full production on 1 June and would ship this fall. NVIDIA's own sequence is different: it said Rubin was in full production in January, said on 31 May that the platform was ramping into full production, and told investors that production shipments began in August. Cloud access has started, NVIDIA now counts seven chips and names ten cloud partners, and its current product page lists lower per-GPU memory and NVLink bandwidth than its launch figures. See our GTC 2027 buyer briefing for rack power and the roadmap.
What NVIDIA shipped
Vera Rubin is a platform of seven chips, not a single GPU: the Vera CPU, the Rubin GPU, the NVLink 6 Switch, the ConnectX-9 SuperNIC, the BlueField-4 DPU and the Spectrum-6 switch, plus the Groq 3 LPU added in March 2026. NVIDIA called Rubin in full production in January and the platform ramping into full production on 31 May, and CoreWeave said on 1 June that it was the first AI cloud provider to bring up and validate an NVL72. The rack-scale unit, the NVL72, combines 72 Rubin GPUs and 36 Vera CPUs into one system, delivering 3.6 exaflops of inference compute and 2.5 exaflops for training.
Where and when you can get it
NVIDIA's March release names ten cloud partners: the hyperscalers most teams already use, AWS, Google Cloud, Azure and Oracle Cloud, plus the specialist NVIDIA Cloud Partners CoreWeave, Crusoe, Lambda, Nebius, Nscale and Together AI. Production shipments began in August 2026. CoreWeave announced availability on 30 September, with Cognition as its first production customer, and Oracle announced general availability on 27 October. NVIDIA's 21 October post carries a general availability headline for Google Cloud's A5X instances, but its text says that is due in November. We found no general availability announcement from the others on 2 November, so confirm directly. As with every NVIDIA generation, the constraint will be capacity, not catalogue availability, so early reservation matters more than which provider lists it first.
| Metric | Rubin | Improvement vs Blackwell |
|---|---|---|
| FP4 inference per GPU | 50 petaflops | 5x |
| FP4 training per GPU | 35 petaflops | 3.5x |
| Memory bandwidth (HBM4) | 19.2 TB/s on current product page; 22 TB/s at launch | 2.8x on launch figure |
| NVLink per GPU | 3 TB/s on current product page; 3.6 TB/s at launch | 2x |
| NVL72 rack inference | 3.6 exaflops | About 5x at rack scale |
| Cost per token (MoE inference) | Up to 10x lower | 10x |
Why it matters for your workloads
The gains are not uniform, so the migration decision should not be either. Rubin's advantage is largest where memory bandwidth and interconnect dominate: large mixture-of-experts inference, advanced reasoning, and agentic workloads that fan out many model calls. Higher HBM4 bandwidth and doubled NVLink are what deliver the up-to-10x cost-per-token improvement, and that is the number that changes a FinOps case. For steady, small-batch inference that already runs cheaply, the benefit is marginal, and the sensible move is to wait. Our guide to GPU spend as a FinOps concern frames the tradeoff.
| Workload | Rubin advantage | What to do |
|---|---|---|
| Large MoE and reasoning inference | Up to 10x lower cost per token | Prioritise for migration |
| Frontier model training | 3.5x training throughput | Reserve capacity early |
| Memory-bound models | 19.2 TB/s HBM4 bandwidth per GPU | Strong fit, plan a move |
| Standard small-batch inference | Marginal at current cost | Stay put for now |
| Small or steady workloads | Limited benefit | No urgency to migrate |
Planning the move
Treat Rubin as a capacity-reservation exercise, not an upgrade you click. Identify the workloads whose cost per token or training time actually gates the business, model the saving against reserved-instance pricing when providers publish it, and get in the reservation queue early because demand will outrun supply. For everything else, Blackwell and cost-optimised options remain the right call, a point we make in our FinOps guide for cloud teams.
India-specific considerations
For Indian teams, Rubin capacity will arrive first through the global hyperscalers and specialist clouds rather than local infrastructure, so latency and data-residency questions matter for any workload touching personal data under the Digital Personal Data Protection Act, 2023. Weigh the cost-per-token saving against egress and residency, and reserve early, because Indian demand competes in the same global capacity pool. The broader planning frame is in our enterprise AI strategy guide.
What to do this quarter
Shortlist the two or three workloads where cost per token or training time is a real constraint, and size the Rubin saving against your current Blackwell spend. Open reservation conversations with your cloud provider now, since capacity, not availability, is the bottleneck. Leave everything that runs cheaply today where it is, and revisit once reserved pricing is public.
FAQ
How eCorpIT can help
eCorpIT is a Gurugram-based, CMMI Level 5 consultancy and an AWS, Microsoft and Google partner. Our senior infrastructure engineers help teams decide which workloads justify Rubin, model the cost-per-token saving against current spend, and plan capacity reservations and migration without over-committing. To plan your AI infrastructure for 2026, talk to our team.
References
- NVIDIA kicks off the next generation of AI with Rubin: six new chips, one AI supercomputer - NVIDIA Newsroom.
- Inside the NVIDIA Rubin platform: six new chips, one AI supercomputer - NVIDIA Technical Blog.
- NVIDIA kicks off the next generation of AI with Rubin - NVIDIA Investor Relations.
- NVIDIA Vera Rubin ships this fall: 8 cloud partners, 10x lower token cost, HBM4 triples bandwidth - Tech Times.
- NVIDIA launches next-generation Rubin AI compute platform - ServeTheHome.
Last updated: November 2, 2026.
