Why I Can't Stop Buying Neocloud Leader Nebius Group

Nebius reported Q4 2025 revenue of $227.7 million, up 503.6% year over year, and full-year revenue of $529.8 million, up 350.9%, while adjusted EBITDA turned positive at $15 million for the first time. Management guided to $7 billion-$9 billion in ARR by end-2026 and about 40% adjusted EBITDA margin, backed by major long-term contracts including a $17.4 billion-$19.4 billion Microsoft deal and roughly $3 billion with Meta. The stock has already gained 681% over the past year, with 11 of 14 analysts rating it Buy or Strong Buy and an average target of $162.

Analysis

NBIS is transitioning from a “story stock” into a scarce-capacity utility for frontier AI demand, and that matters because the winners in this phase are the providers that can pre-sell power, compute, and rack deployment before supply is built. The second-order effect is that the real bottleneck shifts upstream to GPU allocation, transformers, power interconnects, and land permitting, which should keep adjacent infrastructure vendors and NVIDIA in a structurally advantaged position even if NBIS volatility remains high.

The more important signal is not just revenue acceleration, but that the business appears to be crossing the threshold where incremental capacity can be monetized faster than it is installed. That creates a reflexive loop: contracted backlog lowers financing risk, financing lowers execution risk, execution expands backlog. The market may still be underestimating how quickly this can rerate a “pre-profit” infrastructure platform toward a software-like multiple if margins hold near current trajectory into 2026.

The main risk is a capital intensity trap, not demand collapse. If power delivery slips, GPU supply tightens, or customer concentration translates into renegotiation pressure, the equity can de-rate violently because the market is paying for a very specific sequence: capacity first, monetization immediately after. The catalyst window is the next two earnings prints; a miss on deployment cadence or EBITDA margin guidance would matter more than a small revenue beat because the stock is already pricing a steep acceleration.

Consensus may be too focused on the size of the contracts and not enough on their embedded optionality and financing durability. If management continues to convert pipeline into pre-sold capacity, the upside is less about near-term EPS and more about the market granting a higher-quality infrastructure multiple earlier than expected. Conversely, if customers are simply reserving scarce capacity without full utilization, the headline ARR could prove ahead of true economic demand, which is where the bear case would reassert itself.

Key Decisions for Investors

• Long NBIS into the next earnings window only via call spreads or limited-risk structures; risk/reward favors upside convexity because the stock is trading on execution momentum, but outright equity is exposed to financing and buildout slippage.
• Pair long NBIS / short a basket of non-contracted AI infra proxies over 1-3 months; the trade expresses that pre-sold capacity and backlog conversion deserve a premium over narrative-only neocloud peers.
• Overweight NVDA on any pullback tied to cloud capex headlines; NBIS reinforces that compute demand is still outstripping supply, supporting GPU pricing power and utilization through 2026.
• Add MSFT and META selectively on weakness rather than chase NBIS strength; if their outsourced AI infrastructure spend is expanding, the market may be underappreciating that capex intensity is a competitive necessity, not a discretionary expense.
• Avoid shorting NBIS outright before the next print; the squeeze risk is high because any proof of on-time capacity delivery can re-rate the name sharply over days, while the bearish thesis needs months of operational evidence.