AI Infrastructure: Clouds, Data Centers and Alternatives to Nvidia

On October 7, 2026, CNews is holding a meetup on infrastructure for artificial intelligence: clouds, in-house data centers and alternatives to Nvidia GPUs, for customer companies, cloud service providers and IT integrators.

Main discussion topics

The meetup will examine three approaches to AI infrastructure:

• Building in-house data centers: capital investment, hardware requirements, payback periods
• Cloud providers: renting ready-made computing resources with flexible scalability
• Alternatives to Nvidia GPUs: a search for Chinese-made processors (Ascend, Huawei), domestic developments, and joint investment in regional data centers

Why the issue matters in Russia

The country has run into an acute shortage of Nvidia hardware because of export restrictions and sanctions. Companies investing in their own AI systems have to look for independent solutions. Choosing a path, whether an in-house data center, the cloud or a hybrid approach, requires an understanding of the economics, the technical options and a strategic outlook spanning 3–5 years.

The meetup gives companies that have already faced this problem a chance to swap experience with those just starting to plan their AI infrastructure.

What this means

The shortage of Western hardware is changing the game: Russian businesses are speeding up their move to independent infrastructure, choosing among Chinese alternatives, cloud services and consortium investment in regional data centers. The discussion of infrastructure is moving out of technical back rooms and onto the business agenda, and it is becoming critical for any company with AI ambitions.

How to choose between the cloud and an in-house data center for AI?

The CNews meetup will examine three approaches: cloud providers with ready-made computing resources and flexible scalability, in-house data centers with their capital investment and hardware requirements, and alternatives to Nvidia GPUs.