Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Clear out junk files and repair common Windows errors3Fix the driver behind crashes, sound loss and screen glitchesiTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Nvidia is making selected technology available to outside silicon partners, but the public evidence does not show a general program for licensing its graphics cores to any chip vendor. The clearest example is NVLink Fusion, announced on March 31, 2026: it lets companies building custom XPUs integrate those chips into infrastructure compatible with Nvidia systems. Nvidia’s wider licensing strategy appears to span interconnects, networking and other platform technology, not just graphics.
What Nvidia has actually made available
Nvidia’s 2026 annual-review materials describe working with partners through license and development agreements when they want to build Nvidia capabilities into their products. That establishes a licensing practice, but it does not identify a standard catalog of graphics IP blocks, say that every block is available, or set out who can apply.
The most concrete public mechanism for outside custom silicon is NVLink Fusion. Nvidia announced it on March 31, 2026, describing it as a way for customers developing custom XPUs to build heterogeneous AI infrastructure compatible with Nvidia systems. In practical terms, a partner can design an accelerator of its own while connecting it to a broader Nvidia-centered platform. That is different from evidence that the partner has licensed an Nvidia GPU core for use inside its chip.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →What NVLink Fusion means for custom AI chips
An XPU is a broad term for a specialized processor, such as an accelerator designed for a particular AI workload. NVLink Fusion addresses how custom processors fit into a larger system: it offers an integration path into infrastructure built around Nvidia technology. The value is therefore at the system-connection layer, rather than necessarily in the accelerator’s graphics or compute core.
#1 Best Overall
- AI Performance: 767 AI TOPS
- OC mode: 2632 MHz (OC mode)/ 2602 MHz (Default mode)
- Powered by the NVIDIA Blackwell architecture and DLSS 4
- Axial-tech fan design features a smaller fan hub that facilitates longer blades and a barrier ring that increases downward air pressure
- A 2.5-slot design maximizes compatibility and cooling efficiency for superior performance in small chassis
Nvidia’s public descriptions position this as heterogeneous infrastructure: custom silicon and Nvidia components can be used together. The customer may supply the custom XPU, while Nvidia remains part of the surrounding system. Compatibility should not be mistaken for independence from Nvidia’s platform, nor does the announcement establish that every chip vendor, system configuration or product is eligible.
Which technology layers are in scope?
The public picture is broader than graphics, but it is not a published bill of materials for a license. On August 31, 2026, The Register reported Nvidia was inviting competitors to license or use parts of its IP holdings, particularly networking technology. It quoted Dion Harris, Nvidia’s senior director of HPC and AI Hyperscale Infrastructure, describing potential adoption across the CPU, GPU, scale-up fabric, Spectrum-X and InfiniBand networking, and DPU functions.
Rank #2
- Powered by the NVIDIA Blackwell architecture and DLSS 4
- Powered by GeForce RTX 5070 Ti
- Integrated with 16GB GDDR7 256bit memory interface
- PCIe 5.0
- WINDFORCE cooling system
That account suggests Nvidia sees value in making multiple layers of its platform available. It does not establish that all of those technologies are offered under identical licensing terms, or that a partner receives ownership of the underlying IP. Nvidia’s SEC filing for 2026 says the company uses patents, trademarks, trade secrets, confidentiality agreements and licensing arrangements to protect its IP, and expects to continue entering license agreements.
How the partner examples differ
| Partner | Publicly described work | What the announcement does not establish |
|---|---|---|
| Marvell | Linked to NVLink Fusion and custom-XPU/networking integration in the available public description. | A general license to Nvidia graphics cores or published commercial terms. |
| MediaTek | A deeper collaboration announced August 31, 2026, spanning AI infrastructure, local AI computing and automotive, with joint work involving Nvidia accelerated computing, AI, graphics and software. | That MediaTek received a general Nvidia graphics-core license. The announcement describes collaboration, not a standard graphics-IP licensing program. |
These examples should not be treated as interchangeable. Marvell is associated with the custom-XPU integration route; MediaTek’s announcement describes a broader, multi-generation collaboration across several markets. Neither public description supplies a complete contract or proves that a partner can take Nvidia GPU IP and build a competing graphics processor.
Rank #3
- Powered by the NVIDIA Blackwell architecture and DLSS 4. System Requirements: Minimum 850W PSU with 16-pin 12V-2x6 (12VHPWR) connector required. Verify before purchasing.
- Military-grade components deliver rock-solid power and longer lifespan for ultimate durability. Compatibility: 348mm (13.7") length, 3.6 slots, 4.3 lbs. Confirm case clearance and slot spacing. GPU bracket included.
- Protective PCB coating helps protect against short circuits caused by moisture, dust, or debris
- 3.6-slot design with massive fin array optimized for airflow from three Axial-tech fans
- Phase-change GPU thermal pad helps ensure optimal thermal performance and longevity, outlasting traditional thermal paste for graphics cards under heavy loads
Why Nvidia would work with potential rivals
A custom accelerator does not have to replace Nvidia’s role in the system. If a partner’s chip connects through Nvidia’s interconnect and works within Nvidia infrastructure, Nvidia can remain relevant to the surrounding compute, networking, software or system deployment. This may let Nvidia participate in projects even when a customer wants a processor tailored to its own needs.
There is also a strategic trade-off for the customer: custom silicon can coexist with Nvidia technology, but integration may preserve dependence on Nvidia’s platform. The extent of that dependence will vary by implementation; public announcements do not quantify it or disclose which components a particular customer must use.
Rank #4
- Powered by the NVIDIA Blackwell architecture and DLSS 4
- Powered by GeForce RTX 5060
- Integrated with 8GB GDDR7 128bit memory interface
- PCIe 5.0
- WINDFORCE cooling system
What remains unknown
Nvidia has not published, in the cited materials, a public application portal or a complete set of eligibility and deal terms for these arrangements. Royalty rates, license fees, exclusivity, minimum commitments and the precise scope of rights are not disclosed. There are also no authoritative initiative-level figures for the number of licensees, licensing revenue, adoption or market share.
Recommended Free Tools
So the accurate reading of the headline is limited but meaningful: Nvidia is opening selected IP and platform capabilities to partners, and NVLink Fusion gives custom-XPU designers a concrete integration route. A broad, openly available program for licensing Nvidia graphics cores to chip vendors has not been established by the public evidence described here.
Quick Recap
Best Value
- Powered by the NVIDIA Blackwell architecture and DLSS 4 OC mode: 2640MHz/Default mode: 2610MHz (Boost Clock)
- Military-grade components deliver rock-solid power and longer lifespan for ultimate durability
- Protective PCB coating helps protect against short circuits caused by moisture, dust, or debris
- 3.125-slot design with massive fin array optimized for airflow from three Axial-tech fans
- Phase-change GPU thermal pad helps ensure optimal thermal performance and longevity, outlasting traditional thermal paste for graphics cards under heavy loads
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
