Not Once in 75 Years | The Weekly Wrap – 4/5/2026
TL;DR
Michael Mauboussin presents data showing OpenAI's projected 108% CAGR has no precedent in 75 years of market history, while Harris Kupperman explains why extending investment time horizons to 2-3 years creates a sustainable edge over short-term focused markets.
📊 OpenAI's Unprecedented Growth Challenge 3 insights
108% CAGR is a 9.5-sigma event
OpenAI's forecast growth from $3.7B to $145B revenue by 2029 represents a statistical outlier never achieved by any company starting from a $2-5B revenue base in 18,900 firm-years of Compustat data since 1950.
Historical base rate is 7% growth
Companies of OpenAI's size historically averaged 7% annual growth with a 10.6% standard deviation, making the projected target a 9.5 standard deviation event that falls outside all historical precedent.
Base rates inform but don't limit
While no company has achieved this growth trajectory before, base rates are living distributions that can be broken, though such claims require exceptionally high evidentiary standards given the historical improbability.
🏢 The New Economics of Tech Scale 2 insights
Mag 7 captures 66% of economic profit
The top 10 companies represent one-third of market capitalization but generate two-thirds of all economic profit, suggesting their scale advantages from proprietary software create defensible moats previously unattainable by industrial-era firms.
Technology enables unprecedented scalability
Unlike traditional industrial companies that faced diseconomies of scale, modern tech giants leverage proprietary software and AI to grow faster as they get larger, potentially enabling future 'one-person billion-dollar companies' through massive leverage.
⏳ Time Horizon as Competitive Moat 2 insights
Short-term markets are efficiently priced
Wall Street's focus on 30-120 day timeframes creates hyper-competitive, efficient pricing where no edge exists, whereas 2-3 year outlooks offer probability-adjusted certainty away from noise.
Volatility is a buying opportunity
Investors should use short-term price drops from earnings misses or geopolitical tweets to add to high-conviction positions rather than panic selling, provided position sizing allows for such volatility.
Bottom Line
Calibrate ambitious growth forecasts against historical base rates to understand the magnitude of the challenge, but seek your investment edge in multi-year time horizons where short-term efficiency creates long-term opportunity.
More from Excess Returns
View all
We Asked Ben Carlson How an 86% Crash Still Led to 8% Annual Returns
Financial expert Ben Carlson explains that successful long-term investing requires accepting unavoidable risk, resisting action bias during volatility, and understanding that even an 86% crash in 1929 couldn't prevent 8% annual returns for patient investors who avoided panic selling.
He Predicted the AI Bubble in 2023 | Doug Clinton and Gene Munster on Why We're Still in 1996
Deep Water Asset Management's Doug Clinton and Gene Munster argue AI remains in its 1996-era infancy with years of growth ahead before any bubble peak, as enterprise adoption explodes with coding agents driving unprecedented demand while threatening severe short-term disruption to knowledge workers who fail to adapt.
He Invested Through Five Bubbles. He Wrote the Book on Them | What We Learned This Week
Legendary bubble investor Jeremy Grantham and macro strategist Andy Constant join the podcast to explain why mega-cap tech has defied historical mean reversion through regulatory tolerance of monopolies, while detailing the three-phase lifecycle of financial bubbles from innovation through escalation to peak.
We Asked Jeremy Grantham Why AI Won’t Boost Profits — and What It Will Do Instead
Legendary investor Jeremy Grantham argues that AI will ultimately become a commoditized cost of doing business rather than a driver of permanently higher corporate profits, while defending his reputation as a selective "perma bear" who believes mean reversion remains the dominant force in markets despite recent monopoly distortions.