# Danila Shtan

Markdown copy of https://realalpha.app/people/danila-shtan/ on Real Alpha, the best minutes from long interviews with the people building AI. Cite and link the page, not this copy.

Real Alpha has cut 8 moments of Danila Shtan (CTO, Nebius) talking, from an interview on Unsupervised Learning published Nov 2, 2026.

- Role: CTO, Nebius

## What Danila Shtan said

- [Chip scarcity gave AI clouds room to compete](https://realalpha.app/s/chip-scarcity-gave-ai-clouds-room-to-compete/)
  Unsupervised Learning · Nov 2, 2026 · 5 min
  > “this is a massive investment on their part that needs to be justified by the size of a potential business. And it's not that there yet.” — Danila Shtan · [42:19](https://realalpha.app/v/qJkcjURcAe8/?t=2539)
- [Open-model inference still has major room to improve](https://realalpha.app/s/open-model-inference-still-has-major-room-to-improve/)
  Unsupervised Learning · Nov 2, 2026 · 3 min
  > “95% of our customers do not run vanilla open source models. These are modified, fine-tuned” — Danila Shtan · [26:43](https://realalpha.app/v/qJkcjURcAe8/?t=1603)
- [Nebius lets engineers use AI but keeps accountability human](https://realalpha.app/s/nebius-lets-engineers-use-ai-but-keeps-accountability-human/)
  Unsupervised Learning · Nov 2, 2026 · 3 min
  > “whatever you've done with the agent, it's still your result. It's your responsibility. It's your pull request. It's your code.” — Danila Shtan · [54:52](https://realalpha.app/v/qJkcjURcAe8/?t=3292)
- [AI customers need shared storage built for extreme throughput](https://realalpha.app/s/ai-customers-need-shared-storage-built-for-extreme/)
  Unsupervised Learning · Nov 2, 2026 · 4 min
  > “five, six, seven hundred gigabytes per second of throughput.” — Danila Shtan · [16:47](https://realalpha.app/v/qJkcjURcAe8/?t=1007)
- [New Nvidia chips require constant iteration before launch](https://realalpha.app/s/new-nvidia-chips-require-constant-iteration-before-launch/)
  Unsupervised Learning · Nov 2, 2026 · 5 min
  > “they are constantly iterating while these chips are being produced.” — Danila Shtan · [11:51](https://realalpha.app/v/qJkcjURcAe8/?t=711)
- [Prefill/decode disaggregation remains a case-by-case tool](https://realalpha.app/s/prefill-decode-disaggregation-remains-a-case-by-case-tool/)
  Unsupervised Learning · Nov 2, 2026 · 3 min
  > “we are in the stage where everyone's talking about that, and no one's doing that properly.” — Danila Shtan · [22:02](https://realalpha.app/v/qJkcjURcAe8/?t=1322)
- [Execution capacity is the bottleneck across AI infrastructure](https://realalpha.app/s/execution-capacity-is-the-bottleneck-across-ai/)
  Unsupervised Learning · Nov 2, 2026 · 3 min
  > “an experienced project manager is probably in highest demand I've ever seen throughout my career.” — Danila Shtan · [36:34](https://realalpha.app/v/qJkcjURcAe8/?t=2194)
- [Agent reliability may matter more than model size](https://realalpha.app/s/agent-reliability-may-matter-more-than-model-size/)
  Unsupervised Learning · Nov 2, 2026 · 3 min
  > “I think a lot of future AI adoption lies in how reliable your agentic scenarios are and how reliably you reach those outcomes.” — Danila Shtan · [1:00:07](https://realalpha.app/v/qJkcjURcAe8/?t=3607)

## The episode

- [The Window Before AWS Wakes Up: Nebius CTO on Racing the Hyperscalers](https://realalpha.app/v/qJkcjURcAe8/) · Unsupervised Learning · Nov 2, 2026