Speaking about the Nebius cloud infrastructure business. The company talks about their technical advantage from custom-built services and modern architecture developed after 2022. Do you think that's a legitimate advantage they have, or is it just rhetoric?

The first legitimate aspect is their close relationship with Nvidia, which works well because their technical team is excellent. Nvidia is very particular about their relationships; they want them to be mutual. If their R&D sees value in the low-level technical feedback that improves their products, they foster those relationships. The talent that Nebius has allows them to maintain this important relationship.

Speaking about the Nebius cloud infrastructure business. The company talks about their technical advantage from custom-built services and modern architecture developed after 2022. Do you think that's a legitimate advantage they have, or is it just rhetoric?

On the customer side, they have some promising products that are spinoffs of internal solutions for managing cloud infrastructure. In particular, I'm a co-founder of a startup, where we train open-source LLMs. For our startup, we use Nebius infrastructure because we found it to be the best on the market, even though it's not cheap. It allows rapid prototyping for ML engineers, and neither Azure nor Google TPU services offer a similar level of ease for prototyping generative solutions.