AI looks more likely to ruin the world than rescue it

In early August, I wrote a column asking, "Is an 'AI Bubble' Masking Deeper Economic Problems?" I had planned to follow up with a piece on the impact that AI data centers are having on the energy sector, where individual consumers are increasingly being forced to subsidize the industry’s alarming electricity demand and its potential environmental impacts. I’ll get to that, but first, we need to talk about that bubble again.

It is impossible to understate how much of our economy is based on the promised potential of AI. A whopping 80% of this year's stock gains and 40%  of GDP growth are being driven by AI-related investments. You would have to go back to either the 1929 stock market crash or the 1999 dot-com bubble to find similarly alarming trends. It’s also worth noting that all kinds of economic warning lights are flashing bright red, even as the runaway stock prices of AI-related companies mask the reality of the stock market and the economy as a whole.

Consider this recent spate of “investments” by AI-related companies:

OpenAI has recently announced deals worth $600 billion with NVIDIA, AMD, and Oracle. The company will buy $200 billion in graphics cards from NVIDIA, which will then "invest" $100 billion in OpenAI. It will get this money from Oracle, which OpenAI is giving $300 million for the use of its data centers. Oracle then uses that money to buy graphics cards from NVIDIA, which uses the profits to invest in OpenAI. OpenAI also announced it would pay AMD about $100 billion for five million graphics cards. AMD then announced it would give $100 billion in stock to OpenAI. If this sounds like one big circle jerk, that’s because it is.

The price-to-earnings (P/E) ratio is a valuation metric that compares a company's current share price to its earnings per share (EPS), indicating how much investors are willing to pay for each dollar of earnings. These companies all have P/E ratios that reflect a market bet on future growth rather than current profitability. The average PE ratio for the S&P 500 has historically been 13-15. Right now, at least 65 of the top 100 stocks have a P/E ratio over 20. NVIDIA's P/E ratio is over 51, and Oracle's is just under 60.

OpenAI is a private company, so it does not have a traditional price-to-earnings ratio. However, its valuation reached $500 billion as of this month, up from $157 billion a year ago, even though it is expected to lose $5 billion while revenue grows. The company's financial metrics are often compared using price-to-sales ratios, which were estimated at nearly 40x sales in 2024, far exceeding those of major tech firms like Microsoft and Oracle during the dotcom bubble.

OpenAI is projecting a $9 billion loss for 2025, after suffering a $5 billion loss in 2024. The company's revenue is expected to grow to $12.7 billion in 2025, a 243% increase from $3.7 billion in 2024, but it's still not close to breaking even.

Other indicators that the economy is not doing well include the share of subprime auto loans that were 60 days or more past due, which reached nearly 6.5 percent in January and has remained near that number, according to Fitch Ratings. This is important because it is seen as one of the very last payments someone in financial duress is likely to forgo. If people are missing car payments and risking repossession of the thing that moves them through day-to-day life, it’s a bad sign, to say the least.

We have also seen a near doubling in the number of Americans registering to take the Law School Admissions Test, mirroring the jump during the Great Recession. For a class of people who can afford to do so, that’s an indicator that they see the current prospects of the economy and job market as being so bad that they are willing to load up on debt in the hopes of adding a valuable credential to their resumes, while the economy hopefully improves. Given what AGI means for the legal industry, however, that’s probably a much worse bet than it was back then.

It is also worth considering that it has become impossible to identify the amount of credit in the system. Whether we look at the 1929 crash, the dot-com bubble, or the subprime crisis, the main culprit in each instance was an economy that was leveraged with far too much credit (debt). Whereas that was once a knowable variable, existing on the balance sheets of banking institutions, most commercial debt today is serviced by investment products offered by private equity groups, which prevents us from knowing the full extent to which the economy is leveraged. 

Data Centers

Data centers are enormous buildings that house computer hardware to store and process digital information, and the excitement around massive investments in AI has driven a boom in their construction. AI requires a previously unthinkable amount of processing power, and companies are building them as large as over a million square feet to feed the rush. The construction of such data centers is also contributing to the same potential bubble.

A typical AI data center uses as much electricity as 100,000 households, and some of the newest ones use many times more. However, the energy that powers them has to come from somewhere, and that is increasingly forcing retail consumers to compete with data centers for access to energy.

For cooling purposes, data centers also require billions of gallons of water, which has to be potable because of the sensitivity of the equipment, meaning that residents who live near AI data centers have to compete for this resource as well, similar to oil and gas fracking or phosphate operations. For all of these reasons, citizens are increasingly pushing back on the development of data centers in their communities.

Energy Cutbacks

President Trump’s war on renewable energy has compounded the problem by drastically slowing the growth of renewables, which had been the fastest-growing energy producers in our economy. Despite cozying up to AI tech companies and their CEOs, Trump has proved unwilling to take an all-hands-on-deck approach to facilitate AI growth, or at least ease the burden on the American people who will suffer in the competition for natural resources.

According to Senator Martin Heinrich (D-NM), 218 of the 233 projects, ranging from battery storage to green hydrogen, were located in states with Democratic governors, causing a 10% spike in energy costs. This, despite warnings from stock market analysts that the U.S. cannot meet AI power demand without clean energy, battery storage, and grid upgrades.

What is the promise of AI?

As the Financial Times recently wrote, America's economy is now one big bet on AI. So, what does that mean? Increasingly, it seems as though one of the outcomes will be a massive, quite painful economic correction, as it’s worth pointing out that the tech industry is one of the few sectors that has remained relatively unscathed by Trump’s otherwise disastrous tariff policies. Some analysts say the AI bubble is masking the real impact tariffs are having on trade.

OpenAI founder Sam Altman recently tweeted the following:

“We made ChatGPT pretty restrictive to make sure we were being careful with mental health issues. We realize this made it less useful/enjoyable to many users who had no mental health problems, but given the seriousness of the issue we wanted to get this right.

Now that we have been able to mitigate the serious mental health issues and have new tools, we are going to be able to safely relax the restrictions in most cases.

In a few weeks, we plan to put out a new version of ChatGPT that allows people to have a personality that behaves more like what people liked about 4o (we hope it will be better!). If you want your ChatGPT to respond in a very human-like way, or use a ton of emoji, or act like a friend, ChatGPT should do it (but only if you want it, not because we are usage-maxxing).

In December, as we roll out age-gating more fully and as part of our “treat adult users like adults” principle, we will allow even more, like erotica for verified adults.”

Great. We’ll all get a digital friend and personalized porn to keep us busy once AI takes our jobs.

We already know the answers

Proponents of moving full speed ahead with unrestricted AI seem to answer every question about its costs, energy intensity, and social implications by arguing that AGI will be smart enough to solve more problems than it causes. What’s the big deal if we start cooking the earth at a greater clip if AGI figures out how to stop climate change?

But what happens when the billionaires who stand to gain the most from AI companies decide they don’t like its answers? As a society, we have failed to address even the simplest of challenges whenever they impede profit, greed, or even convenience. You don’t need AGI to know that if everyone who could manage their lives in a Prius instead of a Suburban did so, emissions would be cut drastically, or that if grasses that require watering and fertilizer were changed to native coverings, we would have less runoff.  You don't need it to know that we shouldn't be selling water and soft drinks in plastic bottles that pollute both our environment and our bodies.

It also doesn’t take superhuman intelligence to know whether it is a sound idea to keep minting billionaires who buy seven massive homes, fly around on private jets, and sail on megayachts, while more and more people slip into poverty and homelessness. Did you know that universal basic income was a Silicon Valley answer to what to do when AGI replaces the need for the non-wealthy to work? However, someone asked what people would do to find meaning and make productive use of their time if they didn’t have to work, and now all the tech bros seem much more open to keeping it themselves while parroting this supposed problem.

My point is, our world needs more equity, less conspicuous consumption, more balanced opportunity, and less financial inequality. We need to get rid of the ruinous private equity firms that now literally outnumber McDonald’s franchises in the U.S. and not only fail to serve the public interest but operate on a model of chewing up and spitting out our best companies, loading them with debt to pay fat returns before selling them off for parts. We need to make homeownership affordable again. We need to use antitrust laws to rein in companies like Amazon and Google in favor of small mom-and-pop businesses. We need to put outcomes above the bottom line in healthcare. We know how to do these things, but the wealthy elites who profit from the status quo simply don’t want to. AGI won’t change that.

Dennis "Mitch" Maley is an editor and columnist for The Bradenton Times and the host of our weekly podcast. With over two decades of experience as a journalist, he has covered Manatee County government since 2010. He is a graduate of Shippensburg University and later served as a Captain in the U.S. Army. Click here for his bio. Mitch is also the author of three novels and a short story collection available here. He can be reached at editor@thebradentontimes.com.