The Third Rush: a thesis for where the value is in the Fourth Turning
Fran Strajnar · 11 min read
Originally published 13 June 2026 on LinkedIn and Brave New Coin. Revised for 4ts, November 2026.
Every rush ends the same way. A few people get rich, most of the diggers go home with less than they brought, and the money that lasts ends up with the people who sold them what they needed. It happened in California after 1848. It happened again in crypto after 2009. It is happening now with AI.
This essay sets out the thesis behind 4th Turning Strategy. It asks one question about the AI rush: where does value collect, and which seats can a small, fast firm actually take? The answer starts with a difference between this rush and the two before it.
The asset you cannot buy
In 1849 the prize was a metal. Anyone who reached the river could pan for it, and gold kept its value long after the rush ended. In crypto the prize was the asset itself. Anyone with a phone could buy bitcoin, and early holders were paid for patience more than for skill.
AI has no equivalent. There is no coin for intelligence. The leading model labs are private, and the nearest public proxies are chipmakers and cloud providers already priced as giants. Even the frontier companies that do list arrive enormous: SpaceX raised $75 billion in June 2026 at a valuation of about $1.75 trillion.1 For most people, owning this rush directly is not possible. That moves the opportunity toward serving it.
The six layers of a rush
Every rush sorts its participants into the same layers. Where the money settled in the first two tells you a good deal about where it will settle in the third.
Prospectors
Around 300,000 people came to California looking for gold, and by 1850 most of the easily reached gold had already been collected.2 Most diggers barely covered their costs. In crypto, a visible minority of traders got rich while most bought near the top. In AI, the crowded claims are thin products built on someone else's model: wrappers, prompt packs and courses. Each new model release can erase a year of their work.
Picks and shovels
Levi Strauss arrived in San Francisco in 1854 to sell dry goods to the boom town, and the company that carries his name is still trading.3 Crypto's equivalents were exchanges, mining hardware and wallets. In AI the suppliers are chipmakers, cloud providers and the tooling around them. Alphabet, Amazon, Meta and Microsoft were expected to spend about $650 billion on capital expenditure in 2026, much of it on AI infrastructure.4
Infrastructure
The rails everything runs on. Wells Fargo was founded in 1852 to provide express and banking services to gold-rush California, and by the end of the rush it was the dominant express and banking business in the West.5 In crypto, stablecoin issuers earn the interest on the dollars they hold. In AI the rails are data centres, power and fibre. The International Energy Agency projects data centre electricity use roughly doubling to about 945 terawatt hours by 2030.6 This layer suits giants and sovereign funds.
Merchants and services
Samuel Brannan bought up picks, shovels and pans, then ran through San Francisco announcing the strike. He reportedly made $36,000 in nine weeks selling them at a markup, and became California's first millionaire without panning for anything.7 Crypto's merchants were tax software, compliance, audit and media. In AI this layer is the integrators, advisers and specialist products for businesses that must adopt AI and cannot build it themselves. It is still forming.
Financiers
Bankers bought gold dust at a discount and lent to merchants at high rates. In crypto, the funds that did best mostly backed the right exchange rather than the right coin. In AI the cheques are enormous: PitchBook counted $192.7 billion of venture investment in AI companies by early October 2025.8 The minimum ticket is out of reach for almost everyone.
Grifters
Every rush attracts them. California had salted mines and fake claims. Crypto had rug pulls, vapour ICOs and FTX. AI has products described as AI-powered that are nothing of the kind, and agencies reselling prompts as platforms. Their presence tells you something useful: buyers want operators they can trust, and trust becomes a moat.
The pattern
In 1849 you could not buy gold exposure from home. In 2013 you could buy bitcoin from your sofa, which is why the easy trade worked. In 2026 you cannot buy AI directly. So Brannan's route is open again: stand beside the mine and sell what the miners need.
Every business adopting AI is a miner, and every miner needs outfitters.
Two archetypes that recur
| The toolsmithbuilds what miners cannot | The merchantowns the store, spreads the word | |
|---|---|---|
| 1849 | Blacksmith, assayer | Samuel Brannan's store |
| Crypto | Exchange engineer, wallet maker | Exchange founder |
| AI | Builder of integrations and verification | Dealmaker who gathers demand |
| Edge | Technical skill pointed at the diggers' problems | Turning excitement into contracts |
| Risk | Drifting into speculation, or fighting the giants | Selling vapour and burning trust |
The toolsmith builds what the miners cannot: the blacksmith and assayer in 1849, the exchange engineer in crypto, the builder of integrations and verification systems in AI. The edge is technical skill pointed at the diggers' problems, and verification is worth most in markets short of trust. The risk is drifting into speculative products, or competing head on with the giants.
The merchant owns the store and spreads the word: Brannan, the exchange founder, the dealmaker who gathers demand in AI. The edge is turning excitement into contracts, and distribution usually beats invention. The risk is selling vapour, which burns the trust the whole position depends on.
Where the playable seats are
If you are not a chipmaker, a cloud provider or a sovereign fund, the open layer is services and applications. We see three plays, ordered by how quickly they pay.
- Cash now: outfitting expeditions. Done-for-you builds, integrations and deployments, priced on outcomes rather than hours.
- Recurring: the assay office. Specialist, trust-grade products for regulated niches such as compliance, verification and audit trails. Small markets, few competitors, revenue that renews.
- Leverage: the miner's school. Training and enablement for an organisation's own people. It costs little to deliver, grows by reputation and feeds the other two plays.
The rush rule
Every boom has a winter. When the easy gold ran out, most of the diggers moved on, and the businesses that sold goods and moved money kept trading. The durable position is revenue that survives a pullback in the hype: recurring contracts tied to real operating needs.
Today there is a wide gap between the few people who use AI well in practice and the many who know they have to do something. That gap is the opportunity. The technology is already changing how work gets done, and the funding and experimentation climate feels like crypto in 2016, with one difference: far more of the people involved are professionals doing work that will last.
Since this essay first ran, our own work has moved toward the point where capital and judgement meet: preparing companies for AI diligence, structuring transactions and mapping frontier markets for investors. The archetypes have held.
Notes and sources
- NPR, SpaceX prices record $75 billion IPO, 11 June 2026
- Wikipedia, California gold rush
- Wikipedia, Levi Strauss
- Data Center Knowledge (Bloomberg), Big Tech to spend $650B this year, 6 February 2026
- Wikipedia, Wells Fargo
- DCD, IEA: data centre energy consumption set to double by 2030 to 945 TWh
- Wikipedia, Samuel Brannan
- Bloomberg Law, AI is dominating 2025 VC investing, pulling in $192.7 billion, 3 October 2025