PRICED-IN GROWTH
How much growth is in the price, and how rare is it?
A share price is itself a forecast. Taking today's enterprise value (market value + debt − cash) as right, we work back to how fast profits must grow each year for the next ten years, then ask how many US companies starting from similar sales have ever kept that up for a decade. How rare that required growth is gives a way to measure a bubble.
In three lines
- NVIDIA's price assumes profit grows at least 13% a year for ten years. Of the 24 US companies that started a decade with sales in its size group ($100bn+), the ones that kept that up: Amazon, Apple.
- 11 of the 26 companies here are priced for growth that at most 1 in 100 companies of their size achieved. The median US company worth $2bn or more assumes 14% a year, a pace reached by the top 6% of its size group.
- At the 2000 peak Cisco was priced for about 41% a year, something no company of its size had done. Close to that today: Palantir. At the other end, Micron assumes -1%, a pace 77% of its peers beat.
What each price assumes
If today's price is right, profit must grow at least this fast every year for ten years.
price assumesmediantop 10%top 1%of same-size US firms’ 10-year sales growth; scale 0–50% a year
| Company | Sales (4Q) | EV ÷ op. profit | Price assumes (per year, ≥) | Sales growth, last 3 yrs | Same-size firms that did it | vs the past 10 years of same-size firms |
|---|---|---|---|---|---|---|
| Cisco (March 2000 peak)Cisco at the March 2000 peak is an approximation that treats its P/E (about 200) as a multiple of after-tax operating profit. Sales are for the year to July 2000. | $19bn | ≈201x P/E | +41% | — | none | |
| PalantirPLTR · Services and applications | $7.3bn | 119x | +37%+34% to +39% at 8%–10% | +51% | top 0.3%Did it: NVIDIA, Meta, Tesla… | |
| AMDAMD · Chips and networking gear | $45bn | 103x | +35%+32% to +37% at 8%–10% | +28% | none | |
| ShopifySHOP · Users of cheap AI | $14bn | 93x | +33%+31% to +35% at 8%–10% | — | none | |
| GE VernovaGEV · Power, grid and buildings | $44bn | 73x | +30%+27% to +32% at 8%–10% | — | none | |
| ServiceNowNOW · Services and applications | $15bn | 72x | +29%+27% to +32% at 8%–10% | +22% | top 0.1%Did it: Centene | |
| EquinixEQIX · Power, grid and buildings | $10bn | 55x | +26%+23% to +28% at 8%–10% | +7% | top 0.6%Did it: Centene, Meta, Amazon… | |
| EatonETN · Power, grid and buildings | $32bn | 49x | +24%+22% to +26% at 8%–10% | +12% | top 0.5%Did it: Amazon | |
| DuolingoDUOL · Users of cheap AI | $1.2bn | 44x | +23%+20% to +25% at 8%–10% | +35% | top 1.4%Did it: Tesla, Meta, Builders Firstsource… | |
| Arista NetworksANET · Chips and networking gear | $12bn | 42x | +22%+20% to +24% at 8%–10% | +29% | top 0.9%Did it: Centene, Meta, Amazon… | |
| BroadcomAVGO · Chips and networking gear | $89bn | 36x | +20%+18% to +22% at 8%–10% | +36% | top 1.2%Did it: Amazon | |
| AppleAAPL · AI on the device | $479bn | 32x | +18%+16% to +20% at 8%–10% | +8% | top 0.9%Did it: Amazon | |
| VertivVRT · Physical infrastructure (data centres, power) | $13bn | 29x | +17%+15% to +19% at 8%–10% | +25% | top 1.5%Did it: Centene, Meta, Amazon… | |
| AmazonAMZN · Cloud and compute landlords | $797bn | 28x | +16%+14% to +18% at 8%–10% | +13% | top 0.9%Did it: Amazon | |
| Constellation EnergyCEG · Physical infrastructure (data centres, power) | $31bn | 25x | +15%+13% to +17% at 8%–10% | +6% | top 2.5%Did it: Amazon, Apple, Alphabet | |
| MicrosoftMSFT · Cloud and compute landlords | $346bn | 25x | +14%+12% to +16% at 8%–10% | +17% | top 0.9%Did it: Amazon | |
| SalesforceCRM · Services and applications | $44bn | 23x | +14%+12% to +16% at 8%–10% | +10% | top 2.7%Did it: Amazon, Apple, Alphabet… | |
| AppLovinAPP · Users of cheap AI | $7.6bn | 23x | +13%+11% to +15% at 8%–10% | +36% | top 7.6% | |
| NVIDIANVDA · Chips and networking gear | $303bn | 22x | +13%+11% to +15% at 8%–10% | +127% | top 1.8%Did it: Amazon, Apple | |
| QualcommQCOM · AI on the device | $43bn | 22x | +13%+11% to +15% at 8%–10% | +6% | top 2.7%Did it: Amazon, Apple, Alphabet… | |
| AlphabetGOOGL · Cloud and compute landlords | $471bn | 22x | +13%+11% to +14% at 8%–10% | — | top 1.8%Did it: Amazon, Apple | |
| Meta PlatformsMETA · Services and applications | $242bn | 21x | +12%+10% to +14% at 8%–10% | +24% | top 1.8%Did it: Amazon, Apple | |
| OracleORCL · Cloud and compute landlords | $72bn | 20x | +11%+9% to +13% at 8%–10% | +12% | top 3.9%Did it: Amazon, Apple, Alphabet… | |
| IntuitINTU · Users of cheap AI | $21bn | 15x | +8%+6% to +10% at 8%–10% | +41% | top 12% | |
| VistraVST · Physical infrastructure (data centres, power) | $21bn | 15x | +8%+6% to +9% at 8%–10% | +10% | top 13% | |
| AdobeADBE · Services and applications | $26bn | 10x | +2%-0% to +3% at 8%–10% | +12% | top 53% | |
| MicronMU · Chips and networking gear | $133bn | 8x | -1%-3% to +0% at 8%–10% | +105% | top 77% | |
| DatadogDDOG | $4.3bn | — | Operating margin under 5%; today's profit is too thin to work back from (the price also assumes margins rise) | |||
| CoreWeaveCRWV | $7.6bn | — | Operating loss; cannot be solved | |||
Sorted from the highest requirement. ‘Same-size firms that did it’: among US companies whose sales at the start were in the same size group as the company today, the share that grew sales at least that fast for ten years. The companies are those on the Survivors and Scenarios pages.
The yardstick: how fast do companies of each size grow in ten years?
The bigger the company, the harder it is to keep growing fast. The ‘15%+ a year’ column thins out sharply in the larger groups.
| Sales at the start | Companies | Median | Top 10% | Top 1% | 10%+/yr | 15%+/yr | 20%+/yr | 30%+/yr | Fastest growers |
|---|---|---|---|---|---|---|---|---|---|
| under $0.1bn | 1,313 | +6.6% | +39% | +102% | 41.4% | 31.4% | 24.4% | 15.1% | — |
| $0.1–0.3bn | 669 | +6.8% | +21% | +45% | 36.6% | 21.3% | 11.1% | 3.6% | — |
| $0.3–1bn | 806 | +5.7% | +16% | +30% | 27.2% | 11.4% | 5.0% | 0.9% | — |
| $1–3bn | 697 | +4.6% | +13% | +25% | 20.2% | 6.3% | 2.2% | 0.2% | Tesla +47% (from 2013), Meta +46% (from 2010), Builders Firstsource +36% (from 2012) |
| $3–10bn | 497 | +4.2% | +12% | +24% | 15.6% | 5.7% | 2.2% | 0.6% | NVIDIA +46% (from 2015), Meta +41% (from 2011), Tesla +41% (from 2014) |
| $10–30bn | 228 | +1.9% | +8% | +21% | 5.9% | 1.9% | 1.0% | 0.1% | Centene +30% (from 2013), Meta +29% (from 2014), Amazon +28% (from 2009) |
| $30–100bn | 85 | +2.4% | +9% | +22% | 6.4% | 2.5% | 1.2% | 0.0% | Amazon +27% (from 2010), Apple +20% (from 2009), Alphabet +19% (from 2013) |
| $100bn+ | 24 | +2.3% | +10% | +13% | 10.7% | 0.9% | 0.9% | 0.0% | Amazon +21% (from 2015), Apple +13% (from 2011), UnitedHealth Group +12% (from 2014) |
Annual sales of companies reporting to the SEC under US GAAP (XBRL), from each start year 2009–2015 to ten years later (last 2025). Grouped by sales at the start; the 'share that did it' is the share, within the company's current size group, that grew at least as fast. Each start year counts once, so a company can appear up to seven times. Growth by acquisition counts.
How it is calculated
- The calculation
- Enterprise value = the next ten years of after-tax operating profit, discounted, + the value after year ten. After-tax operating profit is the last four quarters' operating profit × (1 − 21%). We solve for the yearly growth rate that makes the two equal.
- Assumptions
- Discount rate 9% (a 10-year Treasury yield in the 4s plus about 5 points for equity risk); the table also shows 8%–10%. Margins stay where they are, so sales grow at the same rate. After year ten, growth adds no value (the company earns just its cost of capital; value after year ten = profit in year ten ÷ discount rate).
- Why 'at least'
- We deduct nothing for the investment needed to grow and assume all profit can be paid out. In reality growth costs money, so the true requirement is higher. The more capital-hungry the company (data centres, power), the more this figure flatters it.
- The historical yardstick
- Annual sales of companies reporting to the SEC under US GAAP (XBRL), from each start year 2009–2015 to ten years later (last 2025). Grouped by sales at the start; the 'share that did it' is the share, within the company's current size group, that grew at least as fast. Each start year counts once, so a company can appear up to seven times. Growth by acquisition counts.
- Limits
- Companies with no figures ten years on (taken over, bankrupt or delisted) are left out, so past success rates are flattered. Only one period is available, starting in 2009, and it was a good one for growth (low rates, the spread of the internet and smartphones). The $100bn+ group has few companies. No inflation adjustment (the calculation is nominal too). If margins are expected to rise, sales need to grow less.
This is not investment advice. A high requirement does not mean the price will fall; it means the price needs something that has rarely happened.
Survivors: which companies could weather a bust →
The same check on cheap stocks: how fast must their profits shrink? →
Updated 10 Nov 2026 06:20 JST · prices to 9 Nov 2026