The AI Bubble: Crypto Degens Have Been Here Before
António Madeira7 min read·Just now--
Last week, Allbirds, a company known in tech circles for its wool sneakers and eco-friendly angle, did something that felt too familiar. After selling off its actual footwear business and brand to American Exchange Group for about $39 million (pretty much a clearance sale for what was once a much-hyped company), the remaining public shell announced a sharp pivot into Artificial Intelligence.
They lined up $50 million in financing from an unnamed investor to buy GPUs and rebrand as NewBird AI. According to their press release, the long-term plan is to become a “fully integrated GPU-as-a-Service and AI-native cloud solutions provider,” and they’re even asking shareholders to drop the old environmental “public benefit” charter.
Shares surged over 600% in a day, hitting peaks around $21–23 before cooling off a bit. It was trading under $3 right before the news.
What hit me hardest was the irony: A brand built on sustainability jumping ship to an industry that is known to devour electricity and water for data center cooling. Researchers at Cornell University have pointed out that AI-driven data centers could emit 24 to 44 million metric tons of CO₂ annually by 2030, which is like adding roughly 5 to 10 million more cars to U.S. roads.
Feels Like Deja Vu
Something feels off about this whole move: the mystery funding source, the vague wording in the press release, and the whole opportunistic vibe. If you’re curious for a quick take, Mo Bitar dropped a short YouTube video that sums up the weirdness nicely.
The video compares Allbirds’ recent move to the blockchain era’s classic: Long Island Iced Tea Corp. slapping “Long Blockchain” on its name and watching the stock jump 500%, even though they had zero real connection to the tech.
The pattern feels all too familiar if you’ve been around crypto and tech long enough. A genuinely powerful technology arrives, capital rushes in faster than actual revenue can justify, valuations drift into fantasy, hype takes over as the main selling point, and eventually reality pulls everything back down, and life goes on. The people who tend to lose the most, however, are the regular investors who show up late and are left holding the bags.
The current AI boom is following that script almost beat for beat. Back in 2021, I was writing for Cointelegraph, asking whether NFTs were “a bubble waiting to pop or a revolution waiting to explode.” Covering that mania and its messy aftermath makes the current situation feel eerily familiar. I’ve seen how easy it is for people to get caught up in the hype and how often FOMO ends up costing everyday people their savings.
The Template: The Dot-Com Bubble
To get where we are now, let’s rewind to the original tech bubble. The late ’90s dot-com era wasn’t just about the hype. Those folks were building the future, that’s obvious now. The problem was human greed taking over. The scenario was pretty similar to what we’re seeing now. Amazing technology that feels like magic, and valuations that felt too good to be true… Because they were.
From 1995 to its March 2000 peak, the Nasdaq shot up about 600%. Price-to-earnings ratios hit 200. By 1999, 39% of all venture capital investments in the United States were going to internet companies. The average internet IPO in 1999 saw stocks rise 266% above their offering price. Companies without revenue, without profit, without even coherent business models, were receiving billion-dollar valuations based on little more than a domain name ending in “.com.”
The stories are legendary now: Pets.com raised $82.5 million, blew $27 million on a Super Bowl ad, and went bankrupt nine months later with almost no sales. WebVan raised $375 million in its 1999 IPO to deliver groceries in 30 minutes, peaked at an $8 billion market cap, and filed for bankruptcy in 2001 after losing over $800 million. I could go on.
When the music stopped, a lot of people didn’t have a chair to sit on. By late 2002 the Nasdaq had dropped 78%, wiping out over $5 trillion. Silicon Valley lost something like 200,000 jobs.
Insiders saw it coming. They cashed out $43 billion between late 1999 and mid-2000 while retail investors kept pouring money in. The builders of the bubble walked away rich. Everyone else held the bag.
And yet, the internet didn’t die. Amazon, Google, and eBay became part of the fabric of modern life. The tech won in the end. The timing and the valuations? Not so much.
The NFT Chapter: When Hype Killed the Tech
A couple of decades later, the same dynamic played out in fast-forward with NFTs. I got a front row ticket, and it was a wild ride.
The core idea, cryptographically verifiable unique digital assets on blockchain, had solid real-world promise beyond cartoon apes. Think fraud-resistant ticketing with built-in royalties, supply chain tracking for authenticity, fractional ownership of real estate, portable digital credentials, and player-owned assets and economies in games. Pilots and early projects were showing it could work.
Then speculation took over and drowned the technological side. Everyone and their mother was jumping headfirst into NFTs, including companies like Nike and Prada. The market swelled to $41 billion in 2021. Wild sales grabbed headlines, volumes exploded, then crashed over 97% in a matter of months. By mid-2023, over 95% of NFT projects had effectively become worthless, with little or no trading activity.
The real damage went beyond the money: the hype made “NFT” a toxic word in serious industries. Legitimate efforts in ticketing or supply chains got lumped in with get-rich-quick schemes. The technology itself survived, but useful adoption got delayed for years. These cycles keep repeating. Now it’s AI’s turn in the spotlight.
The AI Bubble: Bigger Scale, Same Warning Signs
The numbers this time dwarf what we saw in both the dot-com era and the NFT craze. We’re looking at hundreds of billions in infrastructure spending from the big tech players. Amazon, Google, Meta, and Microsoft are set to collectively spend around $400 billion on AI this year, mostly on data centers, with some companies devoting about 50% of their current cash flow to construction.
Some estimates suggest we’ll need around $2 trillion in annual AI revenue by 2030 just to justify the buildout. That is more than the combined 2024 revenue of Amazon, Apple, Alphabet, Microsoft, Meta, and Nvidia. In the meantime, current consumer spending on AI services remains relatively small. OpenAI generated approximately $20 billion in revenue in 2025 while committing to $1.4 trillion in infrastructure spending over eight years.
We’ve already seen several hyped AI startups like Builder.ai, Humane, Robin AI, and others follow the familiar path: big valuations, heavy burn, then shutdown or fire sales. It’s reminiscent of the NFT projects that collapsed after the speculative wave passed, or the dot-com darlings that burned through cash with no sustainable model.
Who Gets Hurt: The IPO Pipeline and Retail Investors
The IPO pipeline looks set to be the final handoff, shifting exposure from sophisticated private capital to public retail investors chasing the AI story. This is where the parallels to both past bubbles feel most pointed.
In the dot-com days, companies were pumped up in private markets and taken public near the peak so insiders could cash out while retail poured in. With NFTs, we saw the same rush, massive hype-driven valuations followed by a brutal collapse that left late buyers holding worthless digital assets.
Now with AI, we’re seeing a similar mechanism being assembled. The risk is that retail ends up absorbing the losses once the easy money and narrative momentum fade, just like before.
“But This Time Is Different”
The AI hype feels like a mix of the dot-com infrastructure race, where everyone rushed to build out the internet backbone with little immediate payoff, and the NFT frenzy, where hype and FOMO drove massive short-term volumes and spiked prices to insane levels before reality hit.
AI delivers real capabilities today to retail users, and not just a niche subset of the internet, as NFTs did. Much of the heavy investment comes from established giants like Microsoft, Google, Amazon, and Meta, companies with actual profits, cash reserves, and diversified businesses.
Still, even if the big platforms weather a correction (as Amazon and Google did after the dot-com), that doesn’t protect the layers of smaller players or the retail investors who end up holding shares at peak valuations through ETFs or direct buys.
The Bubble Won’t Kill AI But It Will Leave Marks
I don’t think Artificial Intelligence is going anywhere. The models are already useful, and they’ll keep advancing, especially as we become increasingly dependent on them.
These bubbles don’t damage the technology itself. They damage the practical rollout and the trust needed to deploy it widely, and NFTs were a prime example of that. They burn capital, sour public perception, and shift the heaviest losses onto later arrivals.
As someone who watched the NFT surge and bust, the warning signs here look awfully familiar, and they remind me of when things went terribly wrong: Companies that have nothing to do with technology jumping on the bandwagon for a quick cash grab.
Allbirds isn’t the only company doing this. Shortly after the sneaker company, a social media outfit called Myseum jumped on the bandwagon with its own AI-tinged rebrand and saw shares pop 150–300% in the frenzy. We’re firmly in “add AI to anything and watch the stock rise” territory.
When all a company needs to do is mention AI for their stock to go parabolic, then you should probably stay away. Sam Altman put it plainly not long ago: someone is going to lose a phenomenal amount of money, and we already know who usually ends up with the heaviest bag.
