Who keeps the value
after the sale?

1. Start with the transaction, then the technology

Two suppliers can serve the same industry while receiving payment for very different work. One sells a machine; another operates equipment on behalf of the customer. A third makes components to a specification. Calling all three an automation platform conceals the question that matters for commercial analysis: what must happen before the next invoice can be earned?

Our framework has four fields: the buyer, the unit purchased, the acceptance condition and the party carrying the asset. Each creates a research task. A buyer category is not a named customer. A published product specification is not proof of an accepted installation. A contract award is not cash collected. Keeping those records separate prevents a strong technical story from becoming an unsupported financial conclusion.

2. The installed base can matter as much as new shipments

ASML reported FY2025 total net sales of €32.667 billion and installed-base management sales of €8.193 billion. Its definition includes net service and field options. [1] Dividing 8.193 by 32.667 gives approximately 25.1%. This is our calculation of the disclosed revenue mix, not an estimate of recurring subscription revenue.

The distinction changes the thesis. An installed base can generate additional customer spending after initial installation, but a field upgrade and routine maintenance do not necessarily have the same cadence. ASML also describes upgrades affecting the installed-base revenue category in its annual report. [2] Our interpretation is to track the content of that category before treating its growth as an annuity.

Counterargument: a large service stream could still depend on exceptional upgrades, customer utilization or concentrated spending. The current mix alone does not establish its future resilience. A useful follow-up is a multiyear reconciliation of maintenance, upgrades, utilization and gross profit where disclosed.

3. Revenue composition changes the relevant competitor

Rocket Lab’s FY2025 filing records total revenue of $601.799 million, including $402.8 million from space systems and $199.0 million from launch services. [3] Our calculation puts space systems at about 66.9% of total revenue; rounded segment figures explain the small reconciliation difference. A launch-only comparison would leave most reported revenue outside the frame.

Our interpretation is that diligence should run along two tracks: mission delivery for launch and manufacturing or programme execution for systems. The filing also identifies a negative revenue adjustment associated with revised estimated completion costs on a contract. [3] That is a concrete reason to inspect estimation and delivery obligations alongside growth.

Counterargument: segment diversity can spread exposure but also creates coordination demands. Revenue share is neither profit share nor cash contribution. We would change the commercial view if delivery costs, concentration or cash conversion deteriorated despite higher reported sales.

4. Selling a service changes the allocation of work

Formic presents factory automation with continuing service under a monthly commercial proposition. [4] The reviewed website does not establish company revenue or gross margin. That absence is material: the model can be attractive to a buyer without revealing the supplier’s return on deployed equipment.

For procurement, ask for the boundary of the promise. Does it include site preparation, tooling changes, consumables, extra shifts and downtime remedies? For supplier analysis, ask how installation and support scale per deployment. These are diligence questions, not allegations about this company’s performance.

Our interpretation is that a recurring bill deserves a deployment-cost model. The buyer may avoid an initial purchase while the provider or financing partner absorbs capital and operating obligations. A credible comparison must price the same output and service scope on both sides.

The implication is operational: negotiate and measure the variables most capable of consuming the apparent spread. A contract that limits uncompensated changes is economically different from one that leaves configuration and service demands open. Before applying any model to a company, replace every assumption with a contract term or clearly labelled estimate.

6. The strongest opposing case

This framework may underweight strategic option value. A young supplier could intentionally accept weak early deployment economics to learn, standardize installations or gain reference customers. An initial loss therefore need not disprove the model. But the explanation creates its own test: later installations should become faster or cheaper, support effort should decline, and retention should improve. Without a cohort record, learning remains a hypothesis.

It can also understate customer switching costs and integration value. Those benefits may support pricing power that is invisible in an early cost model. The right response is to request renewal, expansion and realized pricing evidence rather than assign a premium merely because switching sounds difficult.

7. What would change our view?

1. More favorable: disclosed contribution improves across successive deployment cohorts without shrinking the service promise.
2. Less favorable: installed-base revenue grows while support obligations or customer credits consume the incremental revenue.
3. Unresolved: growth appears only in bookings, announcements or pipeline rather than completed work and recognized revenue.
4. Scope correction: newly disclosed segment economics show the original company comparison grouped fundamentally different activities.

8. Research handoff

Use the company database to capture buyer, model and evidence. Then assemble three documents: the original commercial promise, the delivery record and the financial recognition policy. Add a dated list of contradictions. A convincing company thesis should survive the comparison of all three; a compelling product demonstration alone cannot do that work.