@nikogeorge Big inflection point. SCOTUS clarity + a Schedule III move would finally align policy with economic reality. 280E relief alone could reshape the entire sector.
1/4
⚖️ December 15, 2025 could be the biggest day for U.S. cannabis since Raich in 2005.
SCOTUS will decide whether to hear Canna Provisions v. Garland—a direct challenge to federal authority over 100% intrastate, state-legal cannabis.
Granting cert could mean the end of federal prohibition. 🧵⬇️
Been spending a lot of time digging into Nebius Group ($NBIS) over the past few weeks. It’s rare to find a company that is:
• Growing revenue 300–400% YoY
• Has tens of billions in anchor contracts already signed and being deployed
• Yet trades at very reasonable forward multiples
Key notes from my research (mostly public info + filings):
Core business = full-stack AI-centric cloud infrastructure, primarily large GPU clusters plus the developer platform.
Q3 2025 revenue $146M → +355% YoY. Guidance implies 2026 run-rate of $7–9B. Current clusters sold out into 2026.
Two contracts largely de-risk the story:
• $19.4B multi-year take-or-pay with Microsoft (Azure capacity from Nebius’s New Jersey site)
• $3B five-year deal with Meta
Management expects 20–30% EBIT margins on both — well above typical hyperscaler economics.
Nebius is one of the few independent providers able to deliver material volumes of NVIDIA GB300 racks in 2026 outside AWS/Azure/GCP, and the first in continental Europe to bring them online.
Valuation appears attractive on a sum-of-the-parts basis. The market assigns nearly all of the ~$22B enterprise value to infrastructure and close to zero to everything else.
Other notable assets:
• Avride — 83%-owned autonomous-driving company (Level-4 permit). Commercial robotaxi rides live in Dallas via Uber app. Early-2025 third-party round implied ~$6B standalone valuation.
• Toloka — profitable AI data-labeling / generative-AI data engine
• TripleTen — fast-growing edtech/bootcamp business
• ~28% stake in ClickHouse (open-source analytics DB, private valuation >$2B, frequent IPO candidate)
• >$1B net cash + small strategic Nvidia stake
Rough 2027 estimates emerging on the Street:
≈ $12–14B revenue
≈ 25% EBITDA margin → $3–3.5B EBITDA
15× (conservative) = $45–52B EV for infra alone
Add $7–10B for non-core assets → $55–60B total EV in 2–3 years
Today the company is ~$21B EV. That prices in much of the already-contracted growth.
Risks are large and obvious — multi-billion capex, power procurement, execution on new sites, likely future equity issuance — but Microsoft/Meta prepayments + current cash largely fund the 2026 build-out, so the balance sheet isn’t stretched today.
~3× 2026 guided ARR and ~6–7× 2027 EBITDA. Feels inexpensive vs. most private/public pure-play GPU cloud peers, especially with the European regulatory moat and extra optionality.
Not investment advice, just sharing my work. I’m long and plan to hold for years. The mix of (a) hyper-growth already under contract and (b) several billion dollars of additional assets currently valued near zero feels highly asymmetric.
Curious for pushback or extra color — especially from anyone close to European data centers or the autonomy ecosystem.
DYOR / NFA
#NBIS #AI #investing $NBIS
@pennycheck@Wambscapital I’m holding a pretty large amount of this as well as the January options? Is it time for me to give up on it or at this point stay with it. My buy price was around 9.80 average