Equinix and Fujitsu are among the data center operators that joined Nvidia’s artificial-intelligence (AI) partner program, which gives companies access to AI infrastructure inside a network of 19 colocation partners’ facilities. The DGX-Ready Data Center program is now global, and it added new services including a “try-and-buy” option and GPU-as-a-service.

The other new colocation partners are Cyxtera, Interxion, Data Dock, NTT Communications, NTT Data, IDCFrontier, Vantage Data Centers, and Verne Global.

The chipmaker originally launched the AI-ready program in February with nine North American data centers. “And since we’ve announced it, we have more and more colo providers approaching us,” said Tony Paikeday, director of product marketing for the Nvidia DGX portfolio, at the time of the launch. Nvidia’s goal, he said, is to expand the program globally — and five months later it has expanded to 24 markets around the world.

It added three new colocation partners in Europe, five in Asia, and two in North America bringing its network to 19 validated data center operators. “And the interesting thing here is that all of [the data center operators] came to us,” Paikeday said. Several more will be added to the list after today, he said. “They didn’t get here in time for today’s announcement.”

These partners provide their facilities to host Nvidia’s DGX AI compute infrastructure, and their customers can use the infrastructure without having to buy it upfront. Companies expressing interest in this AI-ready program run the gamut, from Silicon Valley startups, consumer internet companies, auto parts suppliers, health care firms, rideshare companies, online retailers, and research labs, Paikeday said. “The challenge is many organizations need proven facilities to host this kind of infrastructure now, not later, because of the urgency of AI workloads,” he added. “But they need it to be accessible as a cost-effective operating expense.”

In addition to offering the AI-ready infrastructure, two North American data center partners — Core Scientific and Flexential — are also adding an option to let enterprises try the DGX infrastructure before buying these systems and deploying them in their own data center or in a colocation facility.

Partners across all three geographic regions are offering GPU-as-a-service options that let businesses use DGX-powered compute on a pay-per-use basis without buying a full system.

Mobile game developer Jam City is one of these GPU-as-a-service customers. It’s using Core Scientific’s AI-Optimized Cloud product, powered by Nvidia DGX. Machine learning and AI are guiding the company’s game design, said Rami Safadi, chief data officer at Jam City, in a prepared statement. “Core Scientific’s cloud has enhanced how we utilize data and allowed us to analyze billions of rows of data per day,” he said. “We’ve seen an 8x increase in speed, enabling us to train an entirely new set of winning AI business models.”