The Unicorn Graveyard Nobody Is Talking About

Over 220 companies that once hit billion-dollar valuations are now dead or dying. PitchBook calls them "fallen unicorns." The list includes Glossier, Savage X Fenty, AG1, and The Farmer's Dog — brands that were venture darlings just a few years ago.

According to PitchBook data cited by CNBC, there are 857 U.S. startups currently valued at $1 billion or more. Nearly half of them haven't raised fresh funding in three years. That means their valuations are stale — numbers on a spreadsheet that no investor is willing to confirm with actual cash.

Startups that last raised in 2021 are now worth 68% less on average. Those that last raised in 2022 have seen valuations drop 52%, according to PitchBook's own estimates.

ChatGPT Didn't Just Change the Game — It Ended It for Hundreds of Companies

Samir Kaul, a partner at Khosla Ventures and an early OpenAI backer, put it plainly to CNBC: "Now you're seeing 50 engineers do what it would've taken 500 engineers to do five years ago."

That math is brutal for any startup built on the old model. If your entire valuation was based on headcount, proprietary software, or a workflow that GPT-4 can now replicate for $20 a month, you don't have a business. You have a liability.

More than $250 billion has poured into OpenAI and Anthropic alone, according to CNBC. That gravitational pull sucked funding away from everything else. Why bet on a scheduling software startup when you can bet on the technology eating it alive?

The Winners Are Obvious — And Unsurprising

The AI-Driven Enterprise Institute (AIDE) released a new study ranking S&P 500 companies on AI adoption across four categories: literacy, advocacy, orientation, and implementation. The top scorers won't shock anyone.

Nvidia, Amazon, Meta, and SLB (formerly Schlumberger) each hit perfect scores in their respective categories, according to CNBC. Walmart ranked fifth. Utilities AES and NextEra Energy also scored high — a sign that energy infrastructure is taking AI seriously, probably because they're the ones keeping the lights on in the data centers.

The AIDE index drew from publicly available data: earnings call transcripts, job postings, and patent filings.

What Nobody Is Saying Loud Enough

AI isn't free. Every single token generated by ChatGPT, Claude, Gemini, or anything else requires electricity. Electricity requires power plants, cooling systems, real estate, and capital-intensive infrastructure.

Chris MacIntosh, writing via InternationalMan.com and published by ZeroHedge, makes this point bluntly: the gap between what AI actually costs to run and what users pay for it is currently being covered by Wall Street, venture capital, pension funds, and hyperscaler balance sheets. It's losses rebranded as "growth."

Right now, you can use AI tools at prices that don't reflect their true cost. That subsidy exists because the big players — Microsoft, Google, Amazon, Meta — are in a land-grab phase. They're paying for your usage to lock you in. That strategy has a shelf life.

MacIntosh lays out the likely endgame: when financial discipline arrives (and it always does), low-value AI usage disappears first. The junk — AI-generated LinkedIn posts, spam content, automated slop — dies because nobody will pay real money for it. Serious users who derive actual productivity gains remain.

Growth slows. Valuations adjust.

The Media Is Missing the Irony

CNBC's coverage on the unicorn collapse is solid reporting. The data is specific and sourced. But the framing treats the AI boom as the unstoppable force that rightfully destroyed inferior startups — almost celebratory.

ZeroHedge's take, by contrast, asks the uncomfortable follow-up question: is the thing doing the destroying itself built on the same inflated logic? The dot-com boom also had real technology. The internet was genuinely transformative. It also produced a historic crash that wiped out hundreds of billions in paper wealth before the real value emerged.

What This Means for Regular People

If you're a founder or employee at a pre-2022 startup that hasn't raised in two-plus years, the clock is ticking. Your valuation is a fiction and everyone in the room knows it.

If you're an investor in a pension fund or 401(k) with exposure to private markets, understand that "unicorn" no longer means what it did. It may mean you're holding a stake in a company worth half what the paperwork says.

And if you're banking on cheap AI tools staying cheap forever — because a multi-trillion-dollar industry is absorbing the real cost on your behalf — plan for the day that math stops working.

It always does.