@jiahanjimliu on X

NBIS server sales hide a software money machine

NBIS's one-time server sales are the start of a repeating software and revenue-share business that the market is reading as a one-off.

Result

If held over the last 3 months (Aug 15, 2026 – Nov 14, 2026): −43.3%. Over the same days: S&P 500 −2.0%, Chip stocks −13.4%.

What it holds

  • NBIS itself, 55% of the money, −52.9% over 3 months: NBIS (Nebius). If the server sales turn into repeating platform revenue, NBIS is worth more than the market thinks.
  • Other GPU-cloud renters, 20% of the money, −45.4% over 3 months: IREN, CRWV (CoreWeave). If the asset-light model proves out, demand for GPU cloud capacity is strong across the whole group.
  • Chips inside the servers, 15% of the money, −25.2% over 3 months: NVDA (Nvidia). Every server NBIS sells or runs needs Nvidia GPUs.
  • Cash, 10%.

The plan behind it

The main bet is NBIS, since the take is that its revenue is better and more repeatable than the label says. NBIS designs the servers and writes the firmware. It sells them to customers like BoostRun, then may keep earning from its cloud software and from owning the customer bill. If that repeats, the stock should be worth more. A smaller slice goes to other GPU-cloud companies (IREN, CoreWeave), which gain if renting GPU power keeps booming. A small Nvidia slice covers the chips that go into these servers. The take is unproven, because NBIS may not be able to count the same server money twice, so some cash is held back.

$NBIS: Asset-Light Revenue Most of X is missing the mark on why the asset-lite revenue NBIS reported today is labeled no-recurring. I'll breakdown the explanation given in earnings call: "asset-light model. By that, we are bringing capacity online with the Nebius architecture. In this case, some revenues are recognized when we lock this capacity, the supply chain, and the architecture, and some of the revenues are recognized when the capacity actually comes online and being sold to customers. This is actually our first transaction, but as the transactions will grow, and they will become more regular and part of our business, then I assume that we will add them to our base model" (1). Some NBIS investors might have forgotten that NBIS designs it's own servers and writes it's own firmware and then sends the designs to ODMs like Foxconn to manufacture. Thus the initial revenue of asset-lite is selling the NBIS servers to the clients like BoostRun. This is why the term supply chain and architecture is used. Supply chain for sourcing the server components and architecture for the design and firmware. Then the recurring part is once the GPUs sold to the customers are commissioned, NBIS does some type of revenue or profit sharing for providing the Nebius Cloud Software Platform to the client. So the question becomes, if the non-recurring revenue is 325m, what is recurring? Well if GPU payback is 2 years, 325m initial cost means that the total revenue is 162.5m. So how much of the recurring counts as NBIS vs BoostRuns revenue. However, NBIS already charged for the cost of the servers and most of the initial revenue is the client paying back the cost of the servers plus financing so I don't know if NBIS can double count the the GPU servers payments as revenue again since NBIS sold them the revenue. But NBIS does own the customer billing and relationship so I could imagine some form of accounting where the count the 162.5m as revenue receive and then pass out whateve

What would prove it wrong: NBIS shows the server sales are one-offs with little or no repeating platform revenue, or the accounting rules stop it from counting that revenue.

FAQ

How do I bet that NBIS's one-time server sales are the start of a repeating software and revenue-share business that the market is reading as a one-off?

This tested idea holds NBIS (Nebius), IREN, CRWV (CoreWeave) and NVDA (Nvidia), with 10% in cash. Make a copy to run it on practice money first, at real prices.

What would prove it wrong?

NBIS shows the server sales are one-offs with little or no repeating platform revenue, or the accounting rules stop it from counting that revenue.

Can I make one from my own idea?

Yes. Write what you believe in plain words at https://f.xyz/create. fxyz asks a couple of questions, builds the portfolio and shows how it would have done over the last 3 months.

Can I turn a post on X into a theory?

Yes. Reply to the post and tag @fxyzai. It tests the view in the post and replies with the 3-month result and a link.

What can it trade?

Stocks (Tesla, Apple, Nvidia), crypto (Bitcoin, Ethereum, Solana), pre-IPO companies (OpenAI, Anthropic), indices (Nasdaq 100, S&P 500), commodities (gold, silver, oil, copper) and currencies (EUR/USD, USD/JPY). One theory can mix them, and bet for or against any of them.

How does it react to news?

On each check it reads the last two days of news on its biggest holdings and says whether the idea still holds. It can make one small change at a time: trim or drop one holding by up to 5 points. On autopilot the change waits 10 minutes so the owner can skip it; otherwise it asks first.

How do I run it with real money?

For real money, tap Make it mine, open your copy and choose Use real money, with an exchange connected: Hyperliquid, Lighter, Lighter (Robinhood), Backpack, Pacifica or Binance. The money stays in your own account there. fxyz can place trades, but it can never withdraw.

Can the AI get it wrong?

Yes. fxyz's AI turns the idea into holdings, and it can get that wrong: add a coin or stock that doesn't fit, or miss one that does. Check the holdings before you put money in. In your own copy, ask its chat to add or drop any of them.

What do the 3-month numbers mean?

What $100 would have done holding this mix over the last 3 months, at real prices. It is a what-if, not a record: past or simulated results are no promise of future ones, and nothing here is investment advice.

Test your own idea on fxyz · More theories