Here’s what’s on our plate today:

• 🧪 Who pays when the AI bubble bursts?

• 📰 Musk on the stand, big tech's AI clock, bankers build their escape.

• 💡 Roko's Pro Tip: audit your real AI exposure before the market does it for you.

• 🗳️ Poll: Who pays when the bubble pops?

Let’s dive in. No floaties needed…

The Laboratory

TL;DR

• Your 401(k) is already exposed: Seven tech giants make up ~36% of the S&P 500, and 62% of Americans own stocks. A dot-com-scale AI crash could erase more than $20T in household wealth, hitting people who never intentionally bet on AI.

• Jobs are vanishing mid-boom: Tech companies cut 130k+ jobs in 2025 while posting strong earnings, redirecting payroll savings into AI infrastructure. Entry-level roles are disappearing fastest.

• Capital is dangerously concentrated: AI startups absorbed nearly two-thirds of VC funding in early 2025, leaving fintech, biotech, and cybersecurity starved. Meanwhile, AI companies are financially intertwined through cross-ownership, mutual contracts, and stock-based payments.

• If it unravels, the fallout is broad: Retirement accounts, electricity bills, job markets, and the venture ecosystem are all wired to AI's success. Unlike past tech cycles, ordinary Americans have almost no way to opt out.

Who pays when the AI bubble bursts?

Whenever a new technology emerges, it brings a surge of excitement, as breakthroughs fuel ambitious promises, capital flows in quickly, and expectations begin to outpace reality, with companies scaling rapidly and profitability often taking a back seat.

However, that phase rarely lasts, because once the limitations become clear and funding tightens, many players fall away, leaving behind those that can build on cheaper infrastructure, clearer demand, and a more grounded understanding of what actually works.

So far, the story of artificial intelligence has followed a similar trajectory. The launch of ChatGPT in 2022 generated massive hype, and within a couple of years, billions of dollars in debt were incurred to build the underlying infrastructure that made it possible.

While some, often people heading companies riding the AI wave, have assured that this massive investment is necessary for the future of technology, not everyone is convinced. Even as early as 2023, investors were warning that hype and valuations were moving faster than business fundamentals.

However, 2025 became the year when even people building today’s AI systems became more candid about the AI bubble.

OpenAI CEO Sam Altman has openly said the market feels overexcited. At the same time, Google’s Sundar Pichai has acknowledged clear signs of irrationality and warned that no company would be immune if a bubble bursts.

These are not outside critics. They are insiders signaling that the enthusiasm around AI has run ahead of reality. The real question is not whether excess exists, but who pays the price when it fades.

And though businesses may be used to the cycle of bubbles and bursts, for most Americans, the most immediate risk of an AI bubble is not abstract. It sits quietly inside their retirement accounts.

Inside your retirement account

A small group of technology giants now dominates the stock market. Apple, Microsoft, Alphabet, Amazon, Meta, Tesla, and NVIDIA together make up about 36% of the S&P 500.

The top 20 companies account for more than half of the market’s total value, with AI-linked firms driving much of that concentration. Anyone with a 401(k) or an index fund is heavily exposed to AI valuations, whether they realize it or not.

That exposure is widespread. Gallup estimates that 62% of Americans own stocks, and more than half of people earning between $30k and $80k have investment accounts. Former IMF chief economist Gita Gopinath has warned that a dot-com-scale collapse in AI stocks could erase more than $20T in household wealth.

When a handful of companies drives market gains, losses tend to spread just as quickly.

Workers feel it first

Workers are already feeling the pressure, even without a crash. Tech companies cut more than 130k jobs in 2025 while still posting strong earnings. Microsoft laid off over 15k employees while experiencing double-digit revenue growth.

The pattern is increasingly clear. Companies are shrinking payrolls to redirect money toward AI infrastructure.

Some executives have said so outright. Klarna’s CEO said the company reduced its workforce by 40% in part because of AI. Salesforce eliminated thousands of customer support roles after claiming AI could handle much of the work.

Entry-level workers are taking the hardest hit. Job postings for new graduates have fallen sharply, even as employers flood job descriptions with AI requirements. Young workers are told they need AI skills to compete, while AI is often cited as the reason jobs are disappearing.

Whether AI is truly replacing these roles or serving as a convenient explanation for cost-cutting is still debated. Surveys suggest many layoffs stem from weak demand or skills mismatches. For workers who lose their jobs, the distinction offers little comfort.

What AI crowds out

The boom is also reshaping what gets built. AI startups absorbed nearly two-thirds of venture capital funding in the first half of 2025. Investment in fintech, biotech, and cybersecurity has fallen sharply.

This is not just about which companies get funded today. It shapes which ideas never make it off the ground. If AI delivers transformative returns, the shift may be justified. If it does not, the cost will be measured in abandoned projects and unsolved problems.

Even the venture ecosystem itself is shrinking. The number of funds raising capital has collapsed since 2022, despite soaring headlines about AI investment.

Another form of exposure shows up on electricity bills. Data centers already consume more than 4% of U.S. electricity, and demand is projected to surge.

Utilities are spending billions to upgrade grids based on forecasts that may never fully materialize. Consumer advocates warn that if those projections fall short, ordinary ratepayers could be left paying for infrastructure built on inflated expectations. In places like Virginia, data centers already account for more than a quarter of statewide electricity use.

The fragile web

What makes this cycle especially fragile is the tight financial interdependence among AI companies. Major players own stakes in one another, supply one another, and in some cases pay for services with stock rather than cash.

Economists warn that stress in one corner of this system could ripple quickly through the rest. At the same time, tech companies are expected to spend trillions on AI infrastructure, much of it financed through debt or complex arrangements designed to keep obligations off balance sheets.

History offers a mixed lesson. The dot-com crash destroyed enormous wealth but also left behind infrastructure that powers the modern internet. Today’s data centers and AI research may prove valuable even if many of the companies building them do not survive. As Jeff Bezos once noted, society often benefits from the inventions left behind after a bubble bursts.

The problem is that very few people get to choose whether they end up on the winning side. Unlike past tech cycles, AI’s concentration means millions of people who never intentionally invested in AI are already exposed through retirement accounts, job markets, and utility bills.

The unmasked truth of AI economics

According to Alex Heath of MIT Technology Review, what makes this moment feel strange is how openly everyone is discussing the risk. The same executives and investors pouring billions into AI will freely admit that it could all unravel.

As one investor put it, two things can be true at once. AI is likely to transform the economy, and at the same time, the market around it looks inflated, and many people are going to lose money. Those ideas are not in conflict. They are unfolding together.

The internet offers a familiar parallel. Webvan collapsed, but years later, Instacart succeeded with a similar model. Amazon rewarded patient shareholders, while investors in Webvan, a dot-com-era online grocery business, were wiped out. The technology endured, even as many of the early bets failed.

Wall Street is starting to sound more cautious, too. Goldman Sachs has compared today’s AI boom to the tech market of 1997, well before the dot-com crash, and has flagged familiar warning signs, such as heavy spending, rising debt, and tightening credit. Others, including Michael Burry, have drawn similar comparisons.

The AI bubble is real, whether we like it or not. And though it hasn’t reached peak-bubble territory yet, the pressure is clearly building.

As of now, the AI business is stuck in a strange middle ground. Everyone sees the risks, yet the money keeps flowing. As Sam Altman put it bluntly, someone is going to lose an extraordinary amount of money. The only uncertainty is who.

Roko Pro Tip

Bite-Sized Brains

• Musk on the stand: Elon Musk is testifying in court, and his own old tweets are being weaponized against him as evidence.

• Big tech's AI clock ticks: Pressure is mounting on big tech to prove AI can actually generate the returns needed to justify its trillion-dollar infrastructure spend.

• Bankers build their escape: A group of junior bankers tired of grunt work built a $2B AI tool to automate the very tasks they were hired to do.

Monday Poll

Meme Of The Day

The Toolkit

• Deepgram: Speech-to-text API built for scale, handling real-time transcription and voice intelligence for production apps.

• Descript: AI-powered audio and video editor that lets you edit recordings by editing the transcript like a doc.

• Drift.ai: AI-powered conversational marketing platform that turns website visitors into qualified pipeline through automated chat. 

Rate This Edition