Nvidia bars neoclouds from GTC Berlin over AMD and TPU use
Nvidia denied several smaller GPU‑cloud providers (“neoclouds”) entry to its GTC Berlin conference (Oct 20‑22, 2026) after they incorporated AMD Instinct GPUs and Google TPUs into their compute fleets, a SemiAnalysis report released on Oct 16, 2026 disclosed.
Key Takeaways
- Nvidia barred neoclouds that used AMD GPUs or Google TPUs from GTC Berlin, while hyperscalers such as AWS and Azure faced no comparable restriction.
- Neoclouds have been splitting GPU orders roughly 1:1 between AMD and Nvidia, with at least two firms deploying 10,000 units of each.
- The neocloud sector generated >$25 B in 2025 revenue and is projected to approach $400 B by 2031.
- Nvidia’s AI‑accelerator market share is expected at ~75 % in 2026, down from 87 % in 2024.
- More than 300 neocloud providers operated in September 2026, a 55 % rise from the previous year.
| Metric | Value | Why it matters |
|---|---|---|
| Event | GTC Berlin 2026 | Nvidia’s flagship developer summit, a key venue for AI‑infrastructure announcements. |
| Dates | Oct 20‑22, 2026 | Timing left affected neoclouds only days to react before the conference opened. |
| Nvidia AI‑accelerator share (2026) | ~75 % | Indicates Nvidia’s continued dominance despite a modest decline. |
| Neocloud revenue (2025) | >$25 B | Shows the financial weight of the segment Nvidia is pressuring. |
| Neocloud providers (Sep 2026) | >300 | Reflects a 55 % expansion, intensifying competition for GPU capacity. |
What triggered the exclusion at GTC Berlin?
According to SemiAnalysis, Nvidia’s gatekeeping targeted neoclouds that had integrated AMD Instinct accelerators or Google’s custom Tensor Processing Units (TPUs) into their AI‑compute fleets. The restriction was applied only to these smaller providers; hyperscale cloud operators such as Amazon Web Services and Microsoft Azure, which command far larger purchasing power, were not subjected to the same condition.
Why Nvidia is pressuring smaller cloud players
Nvidia’s stated objective is to keep “neocloud” operators on Nvidia‑only hardware, thereby preserving a homogeneous ecosystem that maximizes the company’s AI‑accelerator revenue. By excluding rivals’ chips from its premier conference, Nvidia signals that access to high‑visibility platforms may be contingent on hardware exclusivity.
Neocloud diversification: economics and scale
By August 2026, many neoclouds were allocating GPU purchases in an almost equal split between AMD and Nvidia, with at least two firms each operating 10,000 GPUs from both vendors. The driver is primarily cost: AMD’s Instinct line offers a price advantage, while Nvidia retains the performance lead for many AI workloads. The sector’s rapid growth—up 55 % to more than 300 providers from roughly 209 a year earlier—has intensified the search for competitive pricing.
Responses from the neocloud community
SemiAnalysis reports that some affected providers are evaluating a switch to alternative suppliers to sidestep Nvidia’s pressure. Others are contemplating a “quiet” rollout of non‑Nvidia chips, hoping to avoid a public showdown while still diversifying their hardware stacks.
Broader market implications
Nvidia still commands roughly three‑quarters of AI‑accelerator revenue in 2026, down from an 87 % share in 2024. The company’s leverage at GTC underscores its role as a gatekeeper in the AI‑infrastructure value chain. However, the growing financial clout of neoclouds—projected to near $400 B by 2031—suggests a potential shift in bargaining power if providers collectively pursue multi‑vendor strategies.
Why it matters
For traders and investors tracking AI infrastructure, the episode highlights three risk vectors: (1) whether Nvidia can sustain its ~75 % market share or see further erosion; (2) the likelihood that neoclouds will publicly disclose AMD or TPU deployments, which could reshape market sentiment; and (3) Nvidia’s post‑GTC stance—whether it will intensify, soften, or abandon the exclusionary policy. Shifts in any of these areas could reverberate through GPU pricing, on‑chain DeFi projects that rely on AI compute, and the broader TVL of AI‑related protocols.
Frequently asked questions
Which neoclouds were barred from GTC Berlin?
The SemiAnalysis report does not name specific firms, but it identifies the affected providers as those that had incorporated AMD Instinct GPUs or Google TPUs into their service offerings.
Did hyperscale cloud providers face the same restriction?
No. Nvidia’s pressure was selective; Amazon Web Services, Microsoft Azure, and similar hyperscalers were not barred from attending GTC Berlin despite also offering diverse hardware options.
What is the financial significance of the neocloud sector?
Neoclouds generated over $25 B in revenue in 2025 and are projected to approach $400 B by 2031, underscoring a rapidly expanding market that could challenge Nvidia’s dominance.
How might this affect Nvidia’s market share?
If neoclouds increasingly adopt AMD or TPU solutions, Nvidia’s AI‑accelerator share could slip further below the projected 75 % for 2026, especially if the exclusionary policy provokes a coordinated multi‑vendor response.
What should investors monitor after GTC Berlin?
Key signals include any public disclosures of AMD or TPU deployments by neoclouds, Nvidia’s official response to the backlash, and shifts in AI‑accelerator revenue share reported in subsequent earnings releases.
Bottom Line
Nvidia’s decision to bar neoclouds using AMD GPUs or Google TPUs from GTC Berlin underscores a strategic push to preserve its AI‑hardware monopoly, yet the growing scale and diversification of the neocloud market—already worth >$25 B and projected toward $400 B—creates a credible counter‑force that could erode Nvidia’s ~75 % share if providers collectively embrace multi‑vendor stacks.
Crypto is highly volatile and risky. This is not financial advice — do your own research.
Reporting based on: original source.
Comments
Post a Comment