The risk that really matters isn't on Wall Street

The market fell on news suggesting the AI bubble could be about to burst. If only people paid as much attention to news about climate risks.

U.S. stock market watchers clearly care about the tightly correlated, highly concentrated nature of the current bubble—seven tech giants (Microsoft, Apple, Nvidia, Alphabet, Amazon, Meta, and Tesla) account for more than a third of the entire market value of the S&P 500, with Nvidia alone accounting for 8 percent. Their earnings are growing enormously quickly, at a rate of about 14 percent, while the other 493 companies' earnings are growing at just 3 percent. And all this growth and value is tied to AI, and the idea, common since 2023, that companies that have to do with artificial intelligence are worth investing in.

Focus on the potential that the bubble could suddenly burst is so intense that three tiny pieces of recent news—Sam Altman saying there's a bubble, MIT researchers saying companies are not finding AI investments profitable, and an NYT piece saying Meta is thinking about reorganizing its AI staff—led to a significant selloff, sending the market value of the S&P 500 down $1 trillion. (For an excellent discussion of all this, listen to Ed Elson and Josh Brown here.)

If we were able to take tightly correlated, highly concentrated physical climate risks as seriously as the Magnificent Seven's market cap, two pieces of recent news would be grabbing attention. First, according to a recent study published in Nature, Antarctica is showing signs of rapid, abrupt change, and is fast approaching a threshold that would cause at least 9 feet of sea level rise, flooding coastal areas where hundreds of millions of people live.

Second, observed physical evidence—saltiness of water and changes in surface heat and freshwater—signals that the Atlantic Meridional Overturning Current (AMOC), the great conveyor belt of currents that keeps Europe temperate, is highly likely to start collapsing during this century assuming continued emissions. Even for a low-emissions scenario (which feels very unlikely), the AMOC will weaken between 4 and 46 percent by 2100. For high-emissions scenarios, the models informed by this physical evidence point to a risk of collapse as early as the 2020s through the 2070s—a much earlier window than previously thought.

These are both global catastrophic risks that will affect the lives and well-being of many, "causing suffering to at least one billion people or a total loss of 10% of global gross domestic product" as Johan Rockstrom, director of the Potsdam Institute for Climate Impact Research, said at the beginning of July.

These risks are correlated: As the AMOC weakens, it carries less heat to the north, cooling the North Atlantic and Europe and allowing more warm water to reach the Antarctic shelf, accelerating the melting of glaciers there. And as Antarctic ice melts, the ice cover at the South Pole shrinks, which means it reflects less sunlight, so the planet absorbs more heat, which means the cold fresh water sent into the AMOC increases, slowing down the AMOC (which relies on dense salty water sinking to the bottom of the ocean to drive its conveyor belt).

These risks are also concentrated: The more vulnerable western portion of Antarctica's ice (the West Antarctic Ice Sheet), with most of its base below sea level, could alone raise sea levels by 10 feet or so if it melted. Antarctica's total melt potential is about 190 feet; Greenland's total melt potential is about 23 feet. That's a lot of risk.

The Antarctic news stems from a careful examination of sea ice cover there. It has retreated dramatically over the last decade—about three times faster over those ten years than the decline in Arctic sea ice over almost 50 years. Here's a visualization from Zack Labe:

Scientists have been fixated on the Arctic in recent years, but it's time to turn everyone's attention to Antarctica. Sea ice anomalies there—departures from the norm—are extraordinary. Here's another Zack Labe image:

This chart lets us see when and how much recent Antarctic sea ice has diverged from the historical normal (a reference period from 1981 to 2010), with unusually low “red” values highlighting periods when sea ice coverage has been much less than average since 2010. The five lowest Antarctic summer minimum extents of sea ice have happened since 2017, and NSIDC scientists report that the region seems to be "wandering off into new territory." This is "overwhelming evidence of a regime shift," according to the researchers behind the Nature article.

The AMOC news comes from the group of scientists in Utrecht who lead the world in studying the physics of this crucial current, headed by René van Westen. Essentially, the team has been working on validating observable indicators taken from measurements of the ocean's surface that can help us forecast when the AMOC is likely to start collapsing. They believe that we'll cross a critical threshold quite soon—at 2.5 degrees Celsius above pre-industrial temperatures—meaning that the risk of collapse is larger and nearer than most people have thought.

"Part of the tragedy here in 2025 is that we're at 1.35 degrees now, and we'll be at 1.5 in 2030 at the current rate," Tim Lenton said at the Global Tipping Points Conference at the beginning of July. Based on current trends and policies, this 2.5 level of warming is likely in the second half of this century. Which is not very far away.

Lenton, Rockstrom, van Westen, Australian National University professor Nerilie Abram (the lead author of the Nature study on Antarctica, see Grist story here), and a host of other scientists are trying to tell us that we face abrupt and irreversible changes.

If their names were Microsoft, Amazon, Meta, and Tesla we'd be paying more attention.