When the Media Cries AI Bubble, Should You Sell?

The media is having a field day with the “AI bubble.” Headlines scream caution as if panicking were the hallmark of sophistication. But before you run screaming for the exits, consider this: history suggests that when the media cries bubble, it’s often crying wolf. Only the timing is wrong, and the consequences for jumpy investors can be spectacularly messy.

This isn’t the first time this bubble talk has echoed around AI. In my previous deep dive, The $2.9 Trillion Question: Who Will Survive the AI Bubble?, I explored the staggering scale of investment and the brutal math behind AI’s shaky returns. That analysis highlighted how real survivors will be those who move beyond hype, embedding AI where it actually drives value and productivity.

Now, as media warnings grow louder, it’s worth revisiting the old bubble playbook to see what history says about reacting to these alarms.

Bubble Warnings Through History: What the Media Got Right and Mostly Wrong

Market crashes and media coverage have a complicated relationship. The media rarely calls the peak; it mostly reports the panic. Here’s a bullet-point refresher of notable market tops and how media warnings (or the lack thereof) played out:

• 1929 Great Crash: Roger Babson warned of doom early, but mainstream media and economists like Irving Fisher dismissed him. Optimism ruled until the market lost 25% in two days, then 50%. Media pivoted from cheerleader to alarmist overnight.
• 1987 Black Monday: Some financial press flagged computerized trading risks weeks ahead, but the general public remained blissfully unaware. Media negativity swelled only days before and after the Dow’s historic 22% single-day fall.
• 2000 Dot-Com Bubble: Media fawned over the “New Economy” tech boom until a Barron’s cover story exposed cash crises weeks before the crash. Genuine warnings arrived late, coinciding closely with the precipitous Nasdaq plunge, then became the soundtrack for the carnage.
• 2008 Subprime Crisis: This time, early warnings bubbled up years in advance from The Economist calling it the “biggest bubble in history,” to Robert Shiller and niche bloggers sounding alarms. Yet mainstream media largely downplayed risks amidst the housing euphoria until the crash was already underway.

The media rarely calls the peak; it mostly reports the panic.

Media: Barometer, Amplifier, or Predictive Oracle?

• Media coverage is notoriously pro-cyclical: optimistic on the way up, only pessimistic once markets topple.
• Negative headlines often amplify sell-offs, but rarely cause crashes.
• Social media accelerates these swings, sometimes surfacing earlier signals. But the noise-to-signal ratio is high.
• Extreme media negativity often signals market bottoms rather than tops (the infamous “magazine cover indicator”).

The AI Bubble in 2025: Spectacular Burst or Slowburn?

AI is the shiny new disruptor occupying Wall Street’s pulse, and media makes no secret of the bubble talk. It’s not just hype; it’s investment mania echoing dot-com excess, but with a twist: massive cloud infrastructure giants are setting the stage for a different ending.

Think less dot-com implosion, more “sticky subscription” model.

Microsoft’s Grand Internalization:

• Microsoft has effectively internalized OpenAI by backing its transition into a Public Benefit Corporation (PBC), snagging a $135 billion stake (27%) and securing exclusive IP, API rights, and cloud exclusivity through 2032.
• More than just a financial bet, this cements Azure as the AI cloud platform of choice, swallowing OpenAI’s growing API footprint.
• Rather than letting OpenAI roam free, Microsoft’s move is like the wolf inviting the little AI startup into his den—where the cloud revenues flow generously beneath.

Amazon’s Expected Move:

• Whispered rumors suggest Amazon is eyeing Anthropic, another promising AI lab, planning to fold it snugly under AWS’s massive cloud umbrella.
• AWS supports startups building with models like Meta’s Llama via generous cloud credits and mentorship, betting the farm on a thriving developer ecosystem.
• This signals a clear pattern where hyperscalers don’t just compete for AI dominance. They absorb promising challengers, living off the cloud revenues those companies generate.

Meta’s Strategy: Partnerships Over Clouds:

• Meta’s not playing the cloud hosting game in the same league as Microsoft or Amazon. Instead, they’re all about strategic partnerships with AI startups—think Midjourney, Scale AI, and Reliance Industries.
• These collaborations help Meta move faster on product innovation and quietly spread AI magic across Instagram, WhatsApp, and their virtual reality gadgets.
• Internally, Meta’s splurged something like $72 billion by 2025 on AI. But they also lean heavily on open-source models like Llama to speed things up and make AI more accessible.
• Their partnerships go beyond just software. There’s also AI hardware work with Arm and data labeling with Scale AI, shoring up their ecosystem without needing to lead in cloud.

On Google’s Side:

• Google Cloud still plays second fiddle to Azure and AWS, not really in the big leagues yet.
• Instead of chasing cloud domination, Google focuses on building proprietary AI tools like Bard and weaving AI tightly into everything from search to Maps.
• They’re buying and partnering with startups too, feeding those innovations back into their growing, but still relatively small, cloud network.

What This Means for the Bubble:

The real danger is letting all that media panic make the decisions for you.

• These mega-corporations aren’t just competing. They’re absorbing innovation, locking promising startups into ecosystems ensuring steady cloud or platform revenue.
• The AI bubble might not burst spectacularly but rather slowly deflate as companies consolidate, secure recurring revenue streams, and refine market control.
• Think less dot-com implosion, more “sticky subscription” model: expensive, hard to quit, and quietly profitable over time. The money flows remain within the cloud monopolies’ ecosystems.
• It resembles less to a bursting bubble, but more to a slow season of your favorite soap opera: endlessly drawn out, mildly infuriating, yet oddly addictive.

So, Should You Sell?

The main thing to remember from history? Media warnings usually come late, usually when the crash is already unfolding. The real danger is letting all that media panic make the decisions for you. Jumping in and out on hysterical headlines usually means selling low and missing out on the rebound when things calm down.

Instead of chasing every media headline, think like a seasoned investor and consider:

• Market structure changes, like cloud consolidation in AI.
• Which companies actually endure, innovate, and monetize beyond hype. (Check out my previous article: The $2.9 Trillion Question: Who Will Survive the AI Bubble?)
• The timing challenge: nobody knows exactly when a bubble bursts.
• The value of measured, long-term perspective over reactive moves.

When the media cries AI bubble, the wise investor doesn’t slam the sell button. They sharpen their wits, stock their patience, and maybe sip a glass of dry wit with their portfolio.

Lynn Raebsamen

Zurich

Lynn runs EdTech operations with a CFA in her pocket and fresh powder on her mind. From her Swiss mountain base, she skewers AI myths one story at a time. Author of Artificial Stupelligence. Freeskier. Professional bubble-burster.