There’s a strange tension running through San Francisco right now. On one side, you have founders and early employees at AI giants walking away with generational wealth. On the other, a growing number of engineers are quietly wondering if their careers just hit a dead end. The AI boom, it turns out, is creating two very different realities.
Menlo Ventures partner Deedy Das recently captured the mood in a post that ricocheted across the tech world. His “back of the envelope” calculation puts the number of people who’ve hit retirement-level wealth — north of $20 million — from the current AI cycle at roughly 10,000. That group includes early employees and founders at OpenAI, Anthropic, xAI, Nvidia, and Meta’s AI divisions. Meanwhile, everyone else is staring at a landscape where their hard-won skills may not be the safety net they once were.
The wealth concentration problem
This isn’t just San Francisco hand-wringing. The numbers reflect a structural shift in how value accrues in AI compared to previous tech waves. The social media era spread wealth across thousands of companies — ad networks, platforms, agencies, app developers. AI, by contrast, concentrates value in a handful of infrastructure plays and frontier model labs. If you weren’t at one of those half-dozen companies at the right moment, your upside is dramatically capped.
“The same technology is both the lottery ticket and the thing eating your fallback,” one X user noted in response to Das’s thread. It’s a succinct summary of a genuinely novel dynamic in tech history. Previous revolutions didn’t simultaneously create the new opportunity and undermine the old one with the same tool.
Layoffs and the skills question
Das also pointed out that “layoffs are in full swing” and that “many software engineers feel that their life’s skill is no longer useful.” This isn’t hyperbole. The rise of AI coding assistants, agentic workflows, and vibe-coding tools is reshaping what it means to be a developer. The bar for shipping software is dropping, which is excellent for startups but unsettling for engineers who built careers on skills that AI is increasingly commoditizing.
Some critics pushed back, arguing that even the “losers” in this scenario are still earning well above median incomes. Entrepreneur Deva Hazarika pointed out that most people in Das’s framing are “incredibly fortunate and can simply make a choice to be happy.” Fair point — but it misses the psychological whiplash of watching your stock options flatline while your former coworker’s moon.
What this means for startup founders
If you’re building a startup right now, this divide presents both a challenge and a signal. The challenge is talent: the people who can build frontier AI products know exactly what they’re worth, and competing for them against OpenAI-level comp packages is nearly impossible without massive funding.
The signal is more encouraging: there has never been a better time to build with AI rather than build AI itself. The infrastructure layer is largely settled. The real opportunity is in application-layer startups that leverage foundation models to solve specific, high-value problems. The 10,000 people who struck gold built the picks and shovels. The next wave belongs to the miners.
The takeaway
Yes, AI is concentrating wealth in ways that feel uncomfortable and unfamiliar. But for founders who can resist the FOMO of trying to compete with frontier labs, the application layer is wide open. Build something people need, use the best models as a service, and remember that the gold rush isn’t over — it’s just shifting from who finds the gold to who does the most with it.
Based on reporting by TechCrunch.
