$NBIS signs a 5-year AI infrastructure deal with $META worth up to $27B.
The agreement includes $12B of dedicated capacity starting in 2027 plus a commitment from Meta to purchase up to $15B of additional compute across future Nebius clusters.
Why founder-led businesses tend to win
And why $NBIS fits the bill
Markets focus on quarterly earnings prints which is exactly why we saw $NBIS drop 8% after their Q4 print.
Long-term returns usually come from structure.
One of the most overlooked structural advantages in public markets is founder leadership.
Over multi-year periods, founder-led companies have historically outperformed professionally managed peers. This is because incentives, time-horizon and capital allocation look fundamentally different(Zuck is a great example)
Time Horizon Changes Everything
AI infrastructure requires:
- Heavy upfront capex
- Multi multi year power planning and site planning
- Long contract cycles(think $MSFT 5 year deal)
- Extremely delayed margin realization( $NBIS wont even hit EBIT positive this year)
Which is something that cannot be optimized on a quarter by quarter basis right.
Usually hired in executives have incentives tied to much shorter windows like a 2 or 3 year window, which fundamentally changes how they are going to operate in order to achieve those targets sooner.
A founder optimizes the shape of the business not for the reporting cycle.
Capital Allocation Is apart of the Moat
In AI Infra, power procurement and the time to value from a cluster build out perspectrive are all that matter
Questions that define long-term outcomes:
- When do you secure capacity?
- How aggressively do you expand? ( $NBIS going from 7 to 16 sites)
- Do you prioritize immediate margins or long-term positioning?
- How do you structure contract duration?
$NBIS is taking a disciplined approach:
- Significant cash buffer (60%+ of capex will be funded from cashflow)
- Long-term contracted power roadmap (Guided for 3GW+)
The executive team is treating the balance sheet like a tool.
Previous Experience Matters
Arkady Volozh previously built Yandex into a scaled technology platform by investing early in infrastructure and engineering depth.
This was the biggest IPO in 2011 after google. Arkady has faced the pressure.
He has led a big company in public.
He has navigated extremely competitive markets
He has managed MASSIVE engineering teams
He has allocated capital through cycles.
The experience matters for the AI cycle we are in.
Why This Matters Over the Next Few Years
The current phase for $NBIS is the buildout(duh)
The next phase, if executed correctly, is utilization stabilization and operating leverage expansion.
Founder-led companies often look expensive during the "build" phase. The financial profile changes once the infrastructure maturity is reached(who knows when that will be)
This is when margins expand and free cash flow finally peaks its head.
Most investors arrive at that stage AFTER the rerating(this is our alpha).
Final Thought
Founder-led businesses do not guarantee success. Do not blindly follow a management team ever.
But in capital-intensive industries where timing, discipline, and strategic consistency determine outcomes, leadership structure becomes a material variable.
AI infrastructure is entering a multi-year expansion cycle.
In that environment, aligned ownership and long-term capital allocation matters FAR more than quarterly narratives.
That is the lens through which I evaluate Nebius.
A Colombian pilot flying at 12,500 feet captured what is being described as the best UFO footage ever recorded, and its authenticity has been confirmed.