Actionable insights — Neocloud Economics

1. Split the loss into an operating loss and a financing loss before judging the business

1. For any leveraged, asset-heavy grower, never react to the net loss. Write down three numbers in order: operating profit/loss, interest expense, and net loss — the gap between the first and third is the capital structure, not the business.
2. Identify what the debt is secured on. Equipment-collateralized borrowing (GPUs, aircraft, rigs, ships) carries a higher rate and a residual-value risk that a corporate bond does not: if the collateral de-rates, the facility re-prices or shrinks.
3. Annualize the interest and compare it to the disclosed margin improvement. A business promising "5 to 10 points of contribution margin" on new contracts must first cover an interest line that is growing with every new tranche of capacity.
4. Then ask the deciding question: is interest growing faster or slower than gross profit? That, not revenue growth, is the direction of travel for equity value.

• Each new financing's disclosed rate and collateral coverage; whether interest expense grows faster than gross profit quarter over quarter; and any sign the lender is marking the collateral (GPU resale prices, shortened depreciation lives, tighter advance rates) — the point at which a funding model, not a demand problem, ends the growth.

2. Price contracted capacity per unit of the binding constraint — then read its term structure

1. Find the physical constraint that actually caps the business (here: megawatts of energized power, not chips, not floor space) and convert every contract disclosure into revenue per unit of that constraint.
2. Track that price series across quarters. In a shortage it rises; a flattening or falling series is the earliest evidence the shortage is ending — well before revenue growth slows.
3. Compare long-duration versus short-duration pricing. Normal markets pay a premium for commitment; a scarcity market inverts it and pays a premium for immediacy. The size of that inversion is a scarcity gauge.
4. Cross-check with any list-price action: an outright price increase mid-year, accepted by customers, confirms the pricing power the per-unit series implies.

• The first quarter new deals price below the prior cohort, or the short-term premium compresses toward the long-term rate — both say capacity is catching up with demand; also whether a price increase sticks (renewal rates) rather than merely being announced.

3. Convert a frightening capex number into a payback period before calling it reckless

1. Stop quoting capex as a multiple of revenue in isolation — for a pre-revenue buildout it is meaningless by construction. Instead compute or find the payback period: months for a cohort of new capacity to repay its own capex plus its operating costs.
2. Set that payback against the contract length and the asset's useful life. Payback well inside contracted duration converts "burn" into a timing mismatch; payback beyond it is a genuine bet on renewal or residual value.
3. Track the payback trend across cohorts. Shortening payback with rising capex is compounding; lengthening payback with rising capex is the signal to leave.
4. Only then judge the capex-to-revenue ratio, and read it as a measure of how far ahead of revenue the company is building — not as leverage or profligacy.

• Payback guidance quietly lengthening (higher power/land costs, weaker pricing, longer time-to-energize), and any change in assumed utilization inside it — payback math is only as good as the assumed load factor. Also watch depreciation lives: extending them flatters reported profit without changing the cash payback.

4. Identify who is funding the gap — debt, equity, or the customer

1. For every capital-hungry grower, name the funding source of the build explicitly. There are only three: lenders (interest and covenants), shareholders (dilution), or customers (prepayments, deposits, take-or-pay).
2. Compute the share of capex covered by customer cash. Prepayments are the cheapest capital in existence — no interest, no dilution, and they double as the hardest possible demand signal, since a customer who wires money in advance has revealed more than any letter of intent.
3. Check the balance-sheet counterpart: rising deferred revenue / contract liabilities should track the prepayment claim. If the claim is in the narrative but not on the balance sheet, discount it.
4. Rank peers by funding mix, not by growth. Two companies growing identically with debt versus prepayments are not the same investment.

• Prepayment coverage falling as capex scales (the customer base cannot fund an exponential build forever), any move from customer funding to debt or equity, and concentration inside the prepayments — a large prepaying customer is also a large single point of failure.

5. Treat backlog as a conversion problem, not a demand proof

1. Whenever backlog / RPO is the headline, immediately compute backlog ÷ current annual revenue run rate. A ratio in the high single digits or more is not a bull signal — it is evidence the constraint has moved to the supply side.
2. Re-anchor the analysis on the conversion rate: how much contracted work can physically be delivered per quarter, measured in the constraint's own unit (megawatts energized, ships delivered, wells completed).
3. Model revenue forward from the conversion rate, not from the backlog. Backlog growth beyond the conversion rate adds duration, not near-term revenue.
4. Test the counterparties. A backlog composed of a few names is a customer-concentration disclosure wearing a growth costume.

• Megawatts energized per quarter versus the plan (the real forecast input), any slippage in power, grid interconnect or shell delivery, and contract cancellations or restructurings — a backlog is a promise from a counterparty whose own funding may be less certain than the headline implies.

6. Reconcile GAAP to "core" on a newly public company before drawing any trend

1. In the first few quarters after an IPO, expect the GAAP income statement to be dominated by one-time stock-based compensation as founding grants vest — read the company's core/adjusted reconciliation and check what it excludes (stock comp, customer warrants, pass-through items).
2. Accept the adjustment only where the item is genuinely non-recurring and non-cash; flag anything that will recur (ongoing equity grants, warrants issued to win each new contract) and keep it as a real cost.
3. Build the trend on the core series alone, and require the same basis on both sides of any Y/Y or Q/Q comparison — mixing reported and core margin is the most common error with recent IPOs.
4. Separately isolate any self-inflicted, temporary cost of scaling and ask when management says it reverses. A dated margin trough is a falsifiable claim you can hold them to.

• Whether the Q3 margin trough actually arrives and reverses; whether stock comp declines as promised or is replaced by fresh grants; and whether customer-warrant charges recur with each new large contract — a warrant booked as contra-revenue makes headline growth understate, and future comparisons flatter, the underlying business.

7. Ask whether your supplier's customer is building the supply itself

1. In any shortage-driven business, list the largest customers and ask what each is doing about the shortage internally. In-house capacity programmes (custom silicon, owned plants, private fleets) are the exit path from your revenue line.
2. Distinguish a structural partner from a relief valve: partners are used because the customer never wants to own the capability; relief valves are used only while the customer's own capacity is short.
3. Look for a second supply source appearing from outside the industry — an adjacent giant monetizing spare capacity is proof that the barrier to entry is capital and power, not know-how.
4. Weight contract duration accordingly. Long, committed contracts with tough termination terms are what convert a relief valve into an annuity; short contracts at premium prices are the shortage's peak, not its plateau.

• Hyperscaler capex guidance and custom-silicon deployment milestones, any large customer declining to renew or shortening duration, new entrants selling spare capacity into the merchant market, and the short-term price per megawatt — the first place an easing shortage will show.