CScale has raised $145 million to build out its AI networking business, with both Nvidia and Intel participating in the round, according to a report from WSAU.

The $145 million figure is the single confirmed number in the round. The report names Nvidia and Intel as backers but does not disclose the size of either company's stake, the round's valuation, the identities of other participating investors, or whether the deal includes strategic components such as technology licensing or co-development agreements. CScale itself has not, in the material available, published a breakdown of how it intends to allocate the capital across R&D, hiring, and product delivery.

What the participation of Nvidia and Intel signals, however, is concrete. Nvidia dominates the AI accelerator market, and networking has become a central bottleneck in large-scale AI cluster deployment; the company has pushed its own interconnect portfolio aggressively as system-level bandwidth demands scale. A strategic investment from Nvidia in a networking startup therefore carries commercial weight beyond the dollars involved — it aligns a supplier at the heart of AI compute with a company building connectivity technology for the same workloads.

Intel's involvement points in a related but distinct direction. Intel has spent years working to expand its position in AI infrastructure through its data-center silicon portfolio and its contract manufacturing ambitions, and it has a track record of taking equity stakes in companies it views as strategically adjacent to its platform roadmap. The report does not specify which Intel entity participated — the corporate venture arm or another vehicle — nor whether the investment comes with any commercial strings attached.

The funding lands amid sustained capital concentration in AI infrastructure. Startups building networking, interconnect, and memory-bandwidth solutions have attracted disproportionate investor attention over the past two years, because cluster-scale training and inference workloads have exposed the limits of conventional Ethernet and PCIe topologies. Investors and hyperscalers alike are funding alternatives that promise higher aggregate bandwidth at lower power and latency.

For CScale, the immediate significance of the round is validation. A $145 million raise with two of the most consequential names in semiconductors on the cap table gives the company both balance-sheet runway and credibility with the data-center customers that dominate AI infrastructure purchasing. The round also insulates CScale, at least partially, from the tighter private-market conditions that have compressed valuations across later-stage hardware ventures.

Open questions remain. The report does not detail CScale's product portfolio, its target market segment within AI networking — switch silicon, interconnect IP, optical or copper interconnects, or software-defined fabrics — or any revenue or deployment figures. Without those specifics, the strategic logic of Nvidia's and Intel's participation rests on inference rather than disclosure, and the two backers' own competing networking interests raise the possibility that the investment is as much about optionality and visibility as about product.

The next milestones to watch are whether CScale discloses how the $145 million will be deployed, whether either strategic investor takes a board seat or signs a commercial agreement, and how the company's technology positions against the incumbent networking portfolios that Nvidia already sells into every major AI cluster.