Supporting managed services keeps Neocloud’s GPU capacity healthy

Neoclouds are proliferating as AI labs and enterprises look for dedicated graphics processing unit capacity, but forming a cluster is just the beginning. For smaller providers, managed services support from hardware partners is increasingly determining whether that capacity remains healthy and profitable after go-live.

Specialized AI clouds are gaining a role alongside hyperscalers, particularly where sovereign cloud requirements dictate where data is stored and who controls it. According to Tom Richer (pictured), executive vice president of Alpha Compute Corp., a young vendor cannot own every layer of a complex GPU deployment.

“You can cobble together a lot of different vendors, but then when something goes wrong, the blame game starts,” he said. “You just want an integrated solution provider to ensure your success, to help you with the whole build and then with all the…managed services so that everything runs smoothly and the performance is right.” [service-level agreements].”

Richer spoke with Dave Vellante of theCUBE Research at the Dell Technologies New York Symposium during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed managed service support, sovereign AI infrastructure and Neocloud financing. (*Disclosure below.)

Managed services support anchors a sovereign neocloud

Alpha Compute entered AI infrastructure less than a year ago after starting with digital assets and is now a leader in sovereign and confidential computing running on hardware from Dell Technologies Inc. and Intel Corp. based. Its approach, which the company calls “sovereign by design,” starts with the silicon and stores the data for customers in-country, subject to rules such as the EU AI law and the General Data Protection Regulation, Richer explained.

“We start with the hardware. Imagine a stack. All the way down to the chip level, we create enclaves,” Richer said. “We are a Dell partner and therefore use both Dell technology and Intel technology.”

This stack is already running in production. The company’s 72-node St. Louis cluster combines Dell’s Nvidia Blackwell systems with InfiniBand networking and direct liquid cooling in a retrofitted brownfield data center, and Alpha Compute has outlined plans for additional Blackwell clusters. Alpha Compute announced a binding term sheet for a $31.9 million non-recourse loan facility secured by Nvidia B300 hardware.

“There’s a kind of trinity there, though. You need the buyer, a good buyer,” said Richer. “But you also need so much land and power [and] sleeve.”

Once a cluster is live, Alpha Compute owns the customer relationship. Tickets, remediation and hardware returns are handled by a joint team with Dell, which has made AI factories a core part of its corporate strategy. Alpha Compute’s previous customers are AI labs that want the provider to manage everything below the operating system. Richer says companies will likely charge more.

“With Dell as our managed services partner, we can go above and beyond [operating system] on the second day,” Richer said. “The great thing about Dell and its integrated approach is that we can reach any level.”

This depth of managed services support also shapes the way Richer views Neoclouds’ competition. Instead of replacing the public cloud, purpose-built providers without “noisy neighbors” or multi-tenant sharing would take on the most demanding and sensitive GPU workloads, he explained.

“I really see neoclouds now coming together and merging in a hybrid way with the public cloud for enterprise decision-making,” Richer said. “For the superintelligence workloads that are really demanding, enterprises can send these sensitive, intensive, high-performance computing workloads to the Neocloud so they can work together.”

Here’s the full video interview, part of SiliconANGLE and theCUBE’s coverage of the Dell Technologies New York Symposium:

(*Disclosure: TheCUBE is a paid media partner for the Dell Technologies New York Symposium. Neither Dell, the sponsor of theCUBE’s event coverage, nor other sponsors have editorial control over theCUBE or SiliconANGLE content.)

Photo: SiliconANGLE