Sequoia’s Konstantine Buhler Bets AI Will Claim 99.9% of Human Thinking – Startup Fortune
Sequoia Capital partner Konstantine Buhler argues that artificial intelligence is on track to perform 99.9% of the world’s cognitive work—an echo of how machines captured nearly all physical labor since the 1800s. His thesis, first shared at Sequoia’s AI Ascent event and expanded in a widely read essay, is quickly becoming the intellectual scaffolding for one of Sequoia’s largest strategic moves to date.
The 99.9% Thesis
Buhler’s core claim is a blunt historical parallel: just as industrial machinery gradually took over physical work across two centuries, AI will assume nearly all cognitive tasks—only faster. He values this shift as unlocking a services market on the order of trillions of dollars, territory that traditional software has never fully penetrated. While Sequoia has floated a wide range for the total addressable market, the throughline isn’t precision—it’s direction.
He leans on economic history to caution that disruption comes with a time lag. Drawing on research into the “Engels’ pause” during early industrialization—when productivity surged long before wages caught up—Buhler suggests a similar pattern may repeat in knowledge work, compressed from decades into a handful of years.
A $10 Billion Bet on AI and Reindustrialization
Sequoia has reportedly committed about $10 billion to AI and what it calls “reindustrialization”—a strategy that backs not only model developers but also the physical backbone needed to operate them at scale. Under the leadership of Alfred Lin and Pat Grady, the firm is channeling capital into:
- Specialized chips and compute
- Data centers and power generation
- Defense technology and critical minerals
- Key model labs and AI ecosystems
The logic aligns with Buhler’s thesis: if AI will automate most cognitive work, the enduring value won’t reside solely in software. It will also accrue to the companies that control the factories, power, and supply chains that make large-scale AI possible—a hedge against frothy software valuations and a way to own the “where bits meet atoms” layer of the stack.
Deal Heat: Etched’s Rapid Rise
One emblem of this moment is Etched, a startup building transformer-specific AI chips. Founded in 2022, it has drawn intense investor interest within weeks of emerging from stealth, lining up significant capital, contracts reportedly topping a billion dollars, and valuation discussions spanning the tens of billions. The frenzy underscores investor conviction that specialized silicon will be a central choke point—and profit pool—of the AI era.
The Skeptics’ Case
For now, the financials don’t fully support the most ambitious projections. Even Sequoia’s own analysis suggests a gap between infrastructure investment and realized revenue: estimates peg the revenue required to justify current AI spending at around $600 billion annually, while actual AI revenue is closer to $100 billion. Meanwhile, a recent study found that the vast majority of firms have yet to record measurable productivity gains from AI, despite forecasting future improvements.
Labor market signals are similarly mixed. Recent data shows U.S. software development job postings rising meaningfully since the launch of cutting-edge AI coding tools, even as overall postings declined. But much of that growth is concentrated in senior roles and AI-titled positions, and total postings remain well below pre-2020 levels. That looks less like a workforce already liberated by automation and more like a targeted hiring rebound layered atop a still-normalizing job market.
Early, Not Necessarily Wrong
None of this outright disproves Buhler’s thesis—it simply puts it on a longer, bumpier timeline. Venture capital is about positioning ahead of the curve, and Sequoia appears to be setting up to win across multiple outcomes. If AI does seize most cognitive work, infrastructure owners could hold outsized leverage. If software multiples compress, tangible assets like chips, energy, and data centers can provide ballast.
The precise dollar value of the opportunity—whether it’s $5 trillion or $50 trillion—remains unknowable. What seems clear is the direction of travel: ever more computation, ever larger models, and an expanding lattice of physical systems to power and deploy them. Sequoia’s $10 billion bet is a wager that the foundations of that future, as much as the algorithms themselves, will define the next era of value creation.
The Takeaway
“99.9% of cognitive work” is a provocation, not a forecast to the decimal point. But it captures a simple idea: the scope of what AI can do is widening rapidly, and the winners may be those who build and control the essential rails—silicon, power, and supply chains—required to make it real. Whether the payoff arrives in five years or fifteen, Sequoia is betting that owning the backbone of intelligence will be as consequential as owning intelligence itself.
