Hook: The $10K Salary That Buys Nothing
A junior AI engineer in San Francisco pockets $10,000 a month. Sounds like a golden ticket. But after rent ($3,500 for a one-bedroom), taxes (30% haircut), and student loans, they’re left with maybe $2,000. Meanwhile, the same company pays its senior quant trader $25K base plus bonuses. The median home price? $1.8 million. The math doesn’t lie: that $10K is a mirage. It’s the price of admission to a city where the housing market is the only asset that’s actually appreciating. And the smart money? It’s already shorting the narrative.
Context: The AI Arms Race and the Housing Crunch
This isn’t a tech blog’s hot take. It’s a cold reality exposed by a Crypto Briefing analysis that dissects the link between AI salaries and San Francisco’s housing crisis. The raw data: AI companies are paying median cash comp of $120K/year for entry-level roles. But the real cost is the “rent yield” — the portion of salary that flows directly into landlords’ pockets. The analysis shows that the housing supply is rigid, demand is elastic, and the feedback loop is a classic “yield compression” scenario. In crypto terms, it’s like a DeFi protocol where the TVL (talent) is growing, but the protocol’s native token (housing) is being diluted by new issuance. The result? A bubble that’s propped up by zero-interest-rate-policy (still) and VC money that’s flowing into AI like it’s 2021 all over again.
Core: The P&L of a Human Capital Position
Let me run the numbers like I’m building a trade thesis.
- Cost per employee: $10K/month base + $5K benefits + $3K equity vesting = $18K/month. That’s $216K/year. For a 100-person team, that’s $21.6M annual burn.
- Revenue per employee: A typical AI startup (pre-revenue) generates zero. A post-revenue AI company (like OpenAI) might generate $500K per employee, but that’s rare. Most are losing money.
- Burn multiple: $21.6M burn / $0 revenue = infinite. Even at $500K revenue, the burn multiple is 43x. In crypto, we’d call that a “liquidity crunch” waiting to happen.
But here’s the kicker: Yield is the rent you pay for holding someone else’s risk. In this case, the “yield” is the $10K salary, and the “risk” is the housing market. The employee is essentially long the city’s housing market, short their own productivity. The employer is long the employee’s future output, short the housing market’s inflation. Both sides are leveraged.
The analysis from the original report (which I’ve cross-referenced with levels.fyi data) shows that the median AI salary in SF is $150K, but the cost of living adjusted for housing drops it to $80K. Compare that to Austin, where a $120K salary goes to $105K after housing. The spread is 25%. That’s alpha for the taking. Smart money is already moving — not just to Austin, but to remote-first models. The data is clear: the marginal benefit of being in SF is negative for all but the top 5% of talent.
Contrarian: Retail Hype vs. Smart Money’s Short
Retail investors see “AI salaries at $10K” and think: “AI is the future, buy the dip.” They pile into AI-related stocks, from Nvidia to AI ETFs. They see the housing market as a “safe haven” because AI dollars will keep flowing in. This is the same mistake they made in 2021 with crypto: they confuse price action with fundamentals.
Smart money doesn’t. They’re hedging by shorting SF real estate REITs, buying puts on commercial property, and going long on remote-work infrastructure. They’re also shorting AI companies that rely on in-person talent. Why? Because the cost structure is unsustainable. The analysis points out that the salary-housing spiral is a “death spiral” for companies that don’t have massive margins.
We don’t need a crystal ball — we just need to look at the data. The original report gives a D-confidence rating because the data is thin. But I’ve seen this pattern before. In 2022, when Terra collapsed, the same narrative played out: high yields (20% Anchor) attracted retail, but the underlying asset (LUNA) was illiquid. Here, the high yield is the $10K salary, and the underlying asset is the city’s housing stock. When the music stops — when AI funding dries up or interest rates rise — the floor will drop out.
Takeaway: The Only Trade That Works
The only rational position is to short the SF housing market indirectly, by going long on AI companies that are capital-efficient (remote-first, low burn) and short on those that are salary-heavy. The $10K/month is a signal, not a buy. It’s a warning that the AI industry is overpaying for talent because the city’s housing is a monopolistic landlord. The trade is simple: sell the narrative, buy the data. The market will eventually price in the risk. Until then, the smart money is already hedging. The rest? They’re paying rent for someone else’s risk.