By James Aspinwall
The central finding of the Global Technology Compounders research is simple: a great company is not automatically a great investment at any price.
The project examined whether seven global compounders could deliver annual shareholder returns of 12% to 18% over the next 15 to 20 years. It connected revenue growth, operating margins, reinvestment, cash generation, share count, capital allocation, and terminal valuation instead of extrapolating historical stock returns.
At market prices on September 2, 2026, none of the seven reached the 12% minimum in the base scenario.
The Result
| Company | Bear | Base | Bull | Research decision |
|---|---|---|---|---|
| Microsoft | 6.8% | 10.7% | 13.8% | Accumulate cautiously |
| Meta Platforms | 5.1% | 10.5% | 14.6% | Wait for price |
| Alphabet | 5.3% | 10.3% | 15.0% | Watch the thesis |
| Amazon | 3.9% | 8.7% | 13.0% | Wait for price |
| Mastercard | 4.7% | 8.4% | 10.5% | Pass at the current price |
| NVIDIA | 1.0% | 8.0% | 15.0% | Wait for price |
| Visa | 3.6% | 7.1% | 9.4% | Pass at the current price |
Microsoft is the nearest to the hurdle and has the most defensible business mechanism in the group. Enterprise identity, workflow integration, and switching costs make its position unusually durable. Even so, its estimated base return is 10.7%, not 12%.
Meta and Alphabet offer similar modeled returns but carry different risks. Meta’s core advertising engine remains exceptionally profitable, while infrastructure spending and Reality Labs consume increasing capital. Alphabet combines a strong cloud business with uncertainty over AI search economics and restrictions on exclusive search distribution.
Amazon improves substantially when AWS gains economic weight, but capital intensity still holds the base return below the target. NVIDIA has the greatest range of outcomes: its current platform moat is well supported, but demand durability over 10 to 15 years is not.
Visa and Mastercard are the clearest passes at today’s prices. Their networks remain excellent businesses, but neither reaches the return hurdle even in the modeled bull scenario.
What the Evidence Supports
The evidence supports a narrower conclusion than “technology stocks are too expensive.”
It supports the claim that these seven companies do not presently meet this project’s unusually demanding return objective under its stated assumptions. It does not establish that they will fall, underperform every index, or produce poor absolute returns.
That distinction matters. A 10.7% return would fail a 12% hurdle while still being an excellent long-term result. The decision depends on the investor’s opportunity set, risk tolerance, taxes, time horizon, and required return.
The conclusion is most robust for Visa and Mastercard, whose bull scenarios remain below 12%. It is less robust for Microsoft, Meta, Alphabet, Amazon, and NVIDIA because defensible bull scenarios exceed the hurdle.
The Main Uncertainties
The research is grounded in filings, reconciled financial data, explicit scenario assumptions, and defined thesis breakers. That gives it considerably more substance than a price target built from management guidance or market enthusiasm.
It still cannot remove several important uncertainties:
- Fifteen-year forecasts depend heavily on terminal multiples and the durability of growth.
- AI investment is visible, while its eventual return is often undisclosed or unproven.
- Scenario coverage is not the same as probability. A base case is a structured assumption, not a measured forecast distribution.
- The universe consists of successful companies selected for compounder characteristics, creating survivorship and selection bias.
- Only seven of the original 15 companies received completed qualitative and scenario analysis. Intuit, Adobe, Apple, Uber, Shopify, MercadoLibre, Booking Holdings, and Airbnb remain unfinished, not rejected.
- The model evaluates individual stocks, not taxes, currency exposure, correlations, liquidity needs, or whole-portfolio construction.
The most tempting reasoning error is precision theater: treating an entry price calculated to the cent as if the inputs were equally precise. These prices are decision thresholds under stated assumptions, not objective facts about intrinsic value.
A Portfolio Response
The research should guide a limited stock-selection sleeve, not dictate the entire portfolio.
A sensible response is to keep the core broadly diversified across markets and sectors, count the technology exposure already embedded in index funds, and reserve only a limited portion for individual compounders. Capital waiting for specific entries can remain in short-duration government securities rather than sitting idle.
For investors who insist on a 12% expected-return hurdle, the current answer is no new full-sized position in any of the seven. The modeled entry levels that change that answer are approximately:
- Microsoft: $426
- Meta: $478
- Alphabet: $268
- Amazon: $163
- Mastercard: $365
- NVIDIA: $127
- Visa: $200
These thresholds must move with the underlying businesses. If earnings and owner cash flow grow as expected, a company can become investable without its stock price falling. Conversely, a lower price is not a bargain when the thesis has deteriorated.
Existing holdings require a different decision from new money. Cost basis is economically sunk, but taxes are not. A position should be evaluated against its forward return, tax cost, portfolio weight, and available alternatives. Concentrated holdings can be reduced gradually or diluted with new contributions rather than sold indiscriminately.
Microsoft is the only reasonable candidate for a small starter position if the investor willingly accepts a return below the original hurdle. Calling it an accumulation candidate while maintaining a strict 12% minimum would otherwise be internally inconsistent.
What Would Change the Conclusion
The research defines observable conditions that would invalidate its assumptions:
- Microsoft’s free cash flow falls below 35% of operating income for four quarters.
- Meta’s core operating margin falls below 44%.
- Alphabet suffers sustained deterioration in search share or traffic-acquisition economics.
- AWS operating margin remains below 28%.
- NVIDIA’s purchase commitments keep rising while orders stop accelerating.
- Visa’s revenue captured per dollar of payment volume deteriorates persistently.
- Mastercard resumes gross-revenue disclosure with incentives above 40%.
Crossing one of these thresholds would not create a buying opportunity merely because the share price declined. It would require rebuilding the scenario and recalculating the entry price.
Final Assessment
The probability that the seven fail the project’s 12% base-case hurdle at current prices is likely, with medium confidence. Confidence is limited by the long horizon and the sensitivity of exact returns to assumptions that cannot be observed today.
The conclusion remained unchanged after the model was rebuilt by business segment and its cash balance, working capital, and repurchases were reconciled. The ranking changed, but the central answer did not.
The practical lesson is not to avoid exceptional businesses. It is to separate admiration for the business from the return implied by the price, preserve flexibility, and wait without pretending that waiting is free.