The Mike Moritz Corollary to Moore's Law
Price your market by how cheap compute gets, not by how big it is today
- Difficulty
- Moderate
- Time to result
- ~ongoing to results
- Steps
- 5
- Confidence
- 88%
Mike Moritz inherited Sequoia after funds containing Cisco, Oracle and Apple and asked the obvious question: how could anyone ever top those returns? His answer became a market-sizing rule. As long as Moore's Law holds and computing gets exponentially cheaper, the markets technology can attack keep getting bigger, so each generation of technology company should be able to reach a larger population than the last. The mechanism is a chain: compute price falls 10x, the price of an end device falls with it, the addressable population multiplies, and the revenue ceiling for anyone serving that population rises. The hosts illustrate it with hard numbers. In 1990 a 486-based PC cost $2,000 and about 42% of America used a computer at all. Today a smartphone with roughly a million times the computing power costs $200 and over six billion people carry one. Size the future market from the curve, not from the present.
Origin
Named on Acquired after Sequoia's Mike Moritz, who with Doug Leone took over the firm from Don Valentine and worked out that as long as computing kept getting exponentially cheaper, the markets technology could attack would keep getting bigger. The hosts credit him explicitly rather than claiming the insight.
Core principles
- 01Compute gets roughly 10x cheaper every five to seven years, and that has held for decades.
- 02Cheaper compute does not just improve products, it enlarges the set of markets technology can attack.
- 03Today's total addressable market is a snapshot of a cost curve, not a fixed ceiling.
- 04Exponential curves look flat in linear terms, so the biggest opportunities feel small right up until they are obvious.
- 05Macro drawdowns move the line but do not change the slope.
How to run it
- 1
Name the cost curve underneath the market
Work out which exponentially declining input actually gates adoption in your market: transistors per chip, cost per device, bandwidth, or cost of distribution. If no such curve exists, this framework does not apply.
Pro tip Plot the curve in log scale. In linear scale every exponential looks like nothing happened and then everything happened at once.
- 2
Project the curve, not the market
Take the historical rate of improvement and extend it 5-10 years. The hosts use roughly 10x per seven years for computing power, which has held across the 386, Pentium 4, Core 2 Duo, the iPhone A7 and the Apple M1.
Watch out You may need to adjust the definition of the curve slightly to keep it honest; the hosts concede this for Moore's Law and point at GPUs as the current carrier.
- 3
Convert cheaper inputs into new customers
Ask who enters the market at one tenth the price. Cost falling by 10x rarely means 10x more revenue from existing users; it usually means an order of magnitude more users who could not previously participate at all.
Pro tip Express the answer as a population, not a percentage. Forty-two percent of one country became six billion people.
- 4
Re-size the opportunity at the new population
Recompute the market with the expanded user base. This is the number the investment should be underwritten against, because it is the number the business will eventually operate in.
Watch out Do not confuse a bigger market with a guaranteed share of it; the curve creates the opportunity, execution captures it.
- 5
Hold the thesis through drawdowns
Market caps swing violently while the cost curve grinds on. Treat a crash as a change in price, not a change in the underlying trend, and keep deploying against the curve.
Pro tip Re-verify the curve annually. The day it genuinely breaks, the framework retires with it.
Watch out A curve that has stopped improving invalidates every downstream projection you made from it.
In the wild
When Moritz and Leone took over Sequoia from Don Valentine, they looked back at funds holding Cisco, Oracle and Apple and asked how they could ever top the greatest returns in the history of the asset class. Rather than conclude the best was behind them, Moritz reasoned that as long as computing power kept getting exponentially cheaper, the markets technology could attack would keep getting bigger. The firm went on to back Google, WhatsApp, Airbnb, Meituan and ByteDance, all serving markets an order of magnitude larger than the ones Cisco and Oracle sold into.
→ A generational franchise renewed instead of a firm managing its own decline.
In 1990 a PC with a 486 processor cost about $2,000 and roughly 42% of America used a computer at all. Three decades later a smartphone with about a million times the computing power costs around $200, one tenth the price, and more than six billion people own one. Nothing about the software business model changed; the population that could be sold to did.
→ The addressable market grew from a fraction of one country to most of the planet.
Common mistakes
Sizing the market at today's prices
Underwriting an opportunity against current adoption ignores the entire mechanism. The whole point is that the customer base is a function of a price that is still falling.
Reading the linear chart
On a linear axis every exponential looks like nothing happened and then everything happened, which makes people believe they have arrived at the peak. Look at the log chart before concluding the run is over.
Assuming the curve holds forever without checking
The corollary is entirely dependent on the underlying improvement continuing. Confirm which technology is currently carrying the curve before extending it another decade.
Is it for you?
Best for
Anyone sizing a technology market whose adoption is gated by the cost of compute or distribution.
Not ideal for
Businesses whose unit economics are set by physical inputs or regulation rather than a falling technology cost curve.
From the transcript
“but he realized as long as moore's law continues to hold and computing power continues to get exponentially cheaper the markets that technology can attack…”
“so today a smartphone with you know literally a million times of the computing power cost two hundred dollars which is one tenth the cost…”
From the episode
The Playbook: Lessons from 200+ Company Stories