To The Moon

Hot off the press this morning.

CoreWeave just locked in another $21 billion deal with Meta, expanding their total commitment to $35 billion through December 2032. The capacity will span multiple data centres and include first commercial deployments of NVDA's Vera Rubin platform. CRWV shares jumped 3.5% on the news.

But here is the more interesting question.

Meta already signed a $27 billion deal with Nebius in March. And Nebius stock outperformed CRWV by nearly 4x over the past 12 months, up 400% versus CRWV's 109%. So why is everyone still talking about CoreWeave?

Let's break both down properly.

🏭 The AI Infrastructure Arms Race

The $21 billion Meta deal is not just news for CRWV. It is a statement about the entire neocloud sector.

Hyperscalers like Meta are spending $115 to $135 billion on AI infrastructure in 2026 alone. They are building their own data centres but still cannot keep up with demand. So they are signing long-term contracts, more like energy supply agreements than software deals, with specialized GPU cloud providers. That is the neocloud thesis in one sentence.

The global AI infrastructure market was $72 billion in 2025 and is heading toward $465 billion by 2034. Both CRWV and NBIS are positioned at the centre of that build-out. But they are very different bets.

⚡ CoreWeave: The Heavyweight With a Debt Wall

CRWV is the established leader. Trading around $92, up 130% since its IPO in March 2025.

The bull case is straightforward. Revenue hit $5.13 billion in 2025, up 168% year on year. Guidance for 2026 is $12 to $13 billion, another near-140% jump. The contracted backlog now sits at $66.8 billion, including $35 billion from Meta, $22.4 billion from OpenAI, and $6.3 billion from NVDA directly. By end of 2027, annualised run-rate revenue is projected above $30 billion.

NVDA owns more than 12% of the company. That is not just an investment. That is preferred GPU access in a world where GPU availability determines revenue.

But the bear case deserves respect.

CoreWeave is spending $2.60 in capex for every $1 of new revenue in 2026. Total capex this year: $30 to $35 billion. Long-term debt as of end of 2025: $21 billion, up from $8 billion a year prior. In Q4 2025 alone, interest expense hit $388 million, consuming nearly 25% of top-line revenue. Free cash flow is deeply negative. Net losses expanded 690% year on year. CRWV's 2031 bonds are yielding 11.5%. That is the market pricing in real risk, not just growth uncertainty.

And there is one more flag. CRWV insiders have set up 10b5-1 plans to sell. When the CEO and top strategist are preparing to exit via structured selling programs while publicly talking up the business, that deserves attention.

Wall St. consensus price target: $121. Implies roughly 30% upside from here.

🌍 Nebius: The Agile Challenger Nobody Expected

NBIS is the more interesting story right now.

Trading around $135, up nearly 400% in 12 months. Market cap $25.2 billion versus CRWV's $40.7 billion. Smaller, but closing fast.

The background matters. Nebius was originally Yandex, Russia's dominant tech company. After Russia's invasion of Ukraine, the company divested all Russian assets, renamed itself Nebius Group, and relisted on NASDAQ in October 2024. Led by Yandex co-founder Arkady Volozh. Clean slate, European-headquartered, with a Finnish data centre as its base.

The deals that followed were staggering for a company this young.

Meta signed a 5-year deal worth up to $27 billion in March 2026. Microsoft committed up to $19.4 billion for GPU compute from a New Jersey data centre. Together that is $46 billion in committed contracts. NVDA also made a $2 billion strategic investment, enough to deploy over 5 gigawatts of NVDA systems through 2030. Importantly, NVDA also holds 12% of CRWV. They are betting on both horses, which tells you everything about how hot this sector is.

Revenue in Q4 2025 came in at $228 million, up 547% year on year. The company is targeting $7 to $9 billion in ARR by end of 2026, and 3 gigawatts of contracted capacity. It had just 170 megawatts of live capacity at end of 2025 but is aggressively building toward 800 MW to 1 GW by year-end.

The key differentiator: Nebius claims 3x more compute power per megawatt than CRWV. If that holds at scale, the unit economics are structurally superior.

The risks are real though. Revenue of $228 million is a fraction of CRWV's $5.1 billion. The net loss was $173 million in Q4 alone. Capex of $4 billion in 2025 is set to accelerate sharply. The geopolitical origin story still creates perception risk for some institutional allocators, limiting the investor base compared to a clean-sheet US company. And at 62x trailing sales, the valuation leaves zero room for execution misses.

Wall St. consensus price target: $157. Implying around 16% upside from current levels.

🎯 Side by Side: The Numbers That Matter

CRWV

Revenue 2025: $5.13B - up 168%

Revenue guide 2026: $12 to $13B

Contracted backlog: $66.8B

Total debt: $21B

Capex 2026: $30 to $35B

Market cap: $40.7B

Interest expense Q4: $388M

Stock from IPO: +130%

NBIS

Revenue Q4 2025: $228M - up 547%

ARR target 2026: $7 to $9B

Contracted deals: $46B (Meta + Microsoft)

Net loss Q4: $173M

Capex 2025: $4B, accelerating

Market cap: $25.2B

Stock last 12 months: +400%

NVDA backing: $2B strategic stake

💬 Which One Do You Pick?

This is not an either-or. Both are valid plays on the same mega-trend.

CRWV is the safer neocloud bet. Larger, more established, proven execution, deeper NVDA relationship, and a backlog that stretches to 2030. The debt load is the only real question mark. If AI demand sustains and rates come down, the debt concern fades and the $121 target is conservative. If rates stay elevated and capex keeps burning cash faster than revenue accretes, the pressure builds.

NBIS is the higher-risk, higher-reward alternative. Faster growth rate, lighter balance sheet relative to scale, European presence as a genuine differentiator as EU data sovereignty rules tighten, and a valuation that still has room to re-rate if execution matches the contract flow.

The Meta $21 billion deal today confirms one thing above all.

The hyperscalers need these neoclouds. They are not a temporary fix. They are structural. And the race between CRWV and NBIS is just getting started.

I am not a financial advisor. Trade wisely, Comrades.