Nvidia Just Invited a Rival Chipmaker Into Its Own Racks. Here’s Why
Nvidia (NVDA) opened its NVLink Fusion interconnect to d-Matrix, a Microsoft (MSFT)-backed startup, allowing its chips to integrate with Nvidia systems. This move could help Nvidia collect fees on networking and switches even if competitors supply GPUs. Nvidia's data-center revenue grew 138% YoY to $40.31B last quarter, with total revenue at $96.22B, up 105.8% YoY. However, if d-Matrix's Raptor chips prove more efficient, hyperscalers may shift inference workloads, potentially reducing Nvidia's
How this was made
The 30-second read
Why it matters
The move creates a new revenue stream but introduces competitive risk; it signals a broader industry trend toward modular AI racks.
Market read
The partnership could reshape AI hardware revenue dynamics and affect valuation models for Nvidia, its partners, and competing chipmakers.
What to watch
Potential regulatory scrutiny over dominant AI infrastructure control and the speed of third‑party chip adoption.
Background
Nvidia is expanding its AI data‑center ecosystem by licensing its NVLink interconnect to third‑party inference chip makers, a shift from a pure hardware sales model.
Ticker impact
Nvidia announced NVLink Fusion partnership with d-Matrix, allowing third‑party inference chips to plug into Nvidia racks.
Short‑term upside on partnership news, medium‑term pressure if third‑party chips gain market share.
The deal opens a toll‑collector model, adding non‑GPU revenue, yet it may enable competitors to erode Nvidia's high‑margin GPU sales.
Microsoft-backed d‑Matrix will use Nvidia's NVLink Fusion, highlighting Microsoft’s role in supporting alternative inference silicon.
Minor positive impact as the partnership showcases Microsoft’s AI strategy.
Microsoft’s backing of d‑Matrix aligns with its cloud AI push, but the financial effect is limited.
Marvell previously joined Nvidia’s NVLink Fusion, reinforcing its position as a networking partner in Nvidia’s AI racks.
Slight upside as the partnership validates Marvell’s role in AI infrastructure.
Marvell benefits from increased demand for high‑speed interconnect components in AI racks.
Market effects
Accelerates competition in AI inference hardware and expands the data‑center networking market.
Primarily U.S. AI and cloud providers, with global ripple as hyperscalers adopt the standard.
High, given Nvidia’s central role in AI infrastructure worldwide.
Counterpoint
The partnership may cannibalize Nvidia’s own GPU margins faster than anticipated, leading to longer‑term downside.
Key entities
- CompanyNvidia
Designer of GPUs and AI infrastructure, launching NVLink Fusion.
- Startupd-Matrix
Microsoft‑backed inference chip startup integrating with Nvidia racks.
- CompanyMicrosoft
Backer of d-Matrix, supporting AI hardware diversification.
