The AI infrastructure boom is still accelerating, but Wall Street is starting to ask a much tougher question:
Who’s actually making money from building the machines that power AI?
Thursday’s market reaction was a warning. $NBIS(Nebius Group)$ plunged 16.84%, while $CRWV(CoreWeave)$ dropped 6.86%.
And the interesting part? Nebius actually BEAT revenue expectations.
📉 The problem isn’t just revenue. It’s the cost of growth.
Nebius reported $980.8M in revenue, ahead of expectations of approximately $930M. That’s encouraging.
But investors weren’t looking only at the headline number.
Three things raised concerns:
1️⃣ Guidance pressure: Full-year revenue guidance is now at the lower end of its $3.0B–$3.4B range.
2️⃣ Delayed ambitions: Its December ARR target has shifted to February, raising questions about the pace of capacity deployment and customer ramp-up.
3️⃣ Cash burn: Operating cash outflow reached $273M, while capital expenditure and intangible purchases totalled $7.30B.
That last figure deserves attention. Building AI infrastructure requires enormous upfront investment, and revenue growth doesn’t automatically translate into free cash flow.
💰 CoreWeave has a different problem
$CRWV(CoreWeave)$ has an impressive $132.5B backlog.
That’s a huge number, but a backlog isn’t cash in the bank.
The company also faces substantial financing costs, with approximately $846M in quarterly interest expense.
The critical question is whether those contracts can generate enough cash to cover infrastructure investment, interest payments and future expansion.
🔍 What would change my mind?
I’m watching four indicators:
* Cash conversion: Are reported revenues turning into operating cash flow?
* Capacity utilisation: Are newly built data centres filling up quickly?
* Customer commitments: Are contracts translating into deployed, revenue-generating capacity?
* Financing costs: Can these businesses fund expansion without continually increasing their debt burden or diluting shareholders?
⚡ My investment takeaway
I’m not bearish on the long-term AI infrastructure opportunity.
Demand for computing power could remain enormous for years. But there’s a difference between investing in a growing industry and paying any price for companies racing to build capacity.
The winners won’t necessarily be the companies spending the most.
They’ll be the ones that convert investment into recurring revenue, sustainable margins and eventually substantial free cash flow.
The next phase of the AI trade may be less about who can build the biggest data centre and more about who can actually afford to operate it.
👇 My question to investors:
If you had to choose today, would you rather own Nebius, with its growth potential but enormous capital requirements, or CoreWeave, with its massive backlog but significant financing obligations?
Or is the smarter move to wait until both demonstrate that AI demand is translating into real cash?
@Tiger_comments @TigerStars @TigerClub
Comments