THE UNCERTAIN ROIC OF AI 💰
THE NUMBER
$6 Trillion
Bain & Company Projects That The AI Economy Will Need To Generate $6 Trillion In Annual Revenue By 2031.
At $1.5 trillion in 2031 capex and a 25% capex-to-revenue ratio, the consulting firm calculates AI must generate $6 trillion in annual revenue to justify the build-out. Consumer subscriptions, advertising, and enterprise productivity gains will cover part of that. The remaining ~$4.5 trillion, more than Japan’s nominal GDP, must come from new sources of economic value that largely don’t exist yet.
THE VISUAL
Note: OpenAI is in talks right now to raise $30 billion at a $1.4 trillion valuation. Anthropic is planning for a November IPO at a $2 trillion valuation.
Source: The Blind Spot / Data: Publicly available financial and news data
The AI revolution has made Anthropic and OpenAI two of the world’s most valuable companies, and both expect to reach $1 trillion valuations by year-end. Rapid model improvements have driven their revenue growth, but heavy concentration among enterprise customers is a real risk. At those valuations, investors are betting on strong, sustained top-line growth, large profits and free cash flow, and durable customer retention, none of which is assured, though each would be clear evidence that AI is creating real economic value.
THE QUOTE
“[T]he economy just has so much inertia. People keep doing the same things they're doing. They keep buying from the same company. They keep sort of wanting to use their tools in the same way. […] I think it means we've all been too ambitious on timelines, even with this incredible technology. I think AI is one of the most incredible technologies humanity has ever invented. Society and the economy will adapt more slowly.”
The CEO of the company that sparked the AI boom is now urging patience on its economic impact. Few regular users doubt the technology is impressive and improving fast. Bulls, however, may be overestimating near-term progress and underestimating how long tangible results take to arrive.
THE MENTAL MODEL
Ecosystem
Applied To The Uncertain ROIC of AI
GenAI is applauded as the fastest-adopted general-purpose technology in history, outpacing PCs, the internet, and smartphones. While remarkable, its growth has been helped tremendously by an already-established supporting ecosystem that prior technologies lacked at first. Users face no friction because these tools are free, need no new hardware device, and work with natural language commands, so adoption metrics could be a poor proxy for AI’s return on invested capital.
THE QUESTION
In five years, what percentage of the U.S. adult population will knowingly and intentionally use AI daily? (vs. 23% today)
(My take italicized in bold)
Under 35% (Adoption stalls)
35–50% (Steady climb)
50–65% (The tipping point): Daily AI use becomes the majority behavior, mainly through workplace mandates and AI-first apps, helped by overbuilt infrastructure that speeds adoption. Lingering public distrust from the internet era slows voluntary adoption, while power, water, and grid constraints cap how fast compute can scale. The result is usage that more than doubles from today but stops short of smartphone-level saturation.
65–80% (Mainstream utility)
Over 80% (Near-ubiquity)
Reply with your answer, and why
Until next week,
Neil Patel
One person in your network needs to read this. Forward it to them.
