B Kad

Why does the $NBIS stock valuation not clearly show this? First off Token Factory is buried inside Nebius AI Cloud and it's trading in the public markets which are more short term sentiment driven than the overtly euphoric far too forward looking private markets. Plus Nebius does not separately disclose Token Factory revenue, ARR, gross margin or total enterprise token volume which if disclosed would make a significant difference to how $NBIS itself is valued. Public OpenRouter data misses most dedicated/private workloads. The market still largely values $NBIS as a capital intensive GPU cloud rather than assigning Token Factory a Fireworks/Baseten style inference multiple although players like Fireworks/ Baseten are the ones who will be squeezed far more than $NBIS is a massive physical compute shortage which it seems like we are in. At a ~$45B total $NBIS market cap, an $11B Token Factory would represent only around 23% of Nebius. Though Token Factory could become one of the highest multiple assets inside the entire Nebius group. THE TOP END CASE/ POTENTIAL: If Nebius $NBIS becomes a category leader rather than merely reaching competitor parity: 70T tokens/day × $0.11/M × 0.365 = approximately $2.81B ARR. At 22×–25× ARR: $2.81B × 22–25× = approximately $62B-$70B. That makes $60B–$80B my credible three year top end for Token Factory though it will take a lot of work for Nebius to get there which I am confident they will. A $100B+ blue sky valuation becomes mathematically possible at approximately 90T tokens/day, $4B+ ARR and a sustained premium multiple but I would not make that the core case considering how wildly different public market valuations tend to be compared to current day private markets. Why could $NBIS still seriously reach these volumes? - OpenRouter token volume has grown approximately 250× in under two years. - Agentic workloads can consume 5×-15× more tokens than ordinary chatbot interactions as we all know now. - Token Factory production inference workloads more than tripled in Q2. - Nebius controls the infrastructure instead of depending entirely on rented third party capacity which is a massive advantage in a supply glut. - Eigen AI added model, kernel and inference optimization, Clarifai adds orchestration and system level optimization, Inferize targets cold starts, elasticity and GPU utilization. This acquisition pipeline just shows the amount of Focus Nebius as a company has towards optimizing Token Factory to the best of its potential. All these acquisitions happened within a span of months and this can easily continue. - Nebius already ranks as the fastest inference providers on several major open models as per Artificial Analysis. The remaining challenge is converting raw speed and infrastructure ownership into Fireworks/Baseten level reliability, enterprise adoption and recurring revenue. This is the real $NBIS Token Factory thesis: Current bullish value: $8B-$12B Three year competitor parity: $30B-$40B Credible category leader top end: $60B-$80B The market is currently treating Token Factory like an optional feature attached to the Nebius GPU cloud and when the markets realize it's true potential there is a significant chance of the stock rerating. If Nebius at some point starts disclosing Token Factory ARR, total enterprise tokens and dedicated endpoint growth, the market might be forced to value it as an inference platform instead. This could then absolutely be the next rerating event for $NBIS. Finally this is personal math and it' not unlikely for me to make mistakes. Please feel free to correct me if you think otherwise. And finally these are merely estimates and none of this is to be considered as financial advice. NFA

Here’s the link to the first ever Nebius weekly if you haven’t already watched it 👇 youtube.com/live/HQ4p_a6EpNM…