The blue tsunami and a tech rout: What do the midterms mean for AI?

The US midterm elections have seen the Democrats take control of Congress, winning both the Senate and the House of Representatives in a landslide victory.

Fresh talent within the Democratic party turned the midterms into a referendum on money in politics, the power of the tech-industrial complex, and the widening wealth gap. It paid off. James Talacrico turned Texas blue for the first time in two generations; Adam Hamilton became Kansas’ first Democratic senator in nearly a century; and Abdul El-Sayed became the first Muslim senator after a fierce campaign in Michigan.

So great was the blue tsunami that even President Donald Trump conceded defeat, acknowledging a raft of divisive foreign and domestic policies, from Iran to ICE. These results will hobble Trump’s ability to prosecute the MAGA agenda for the final two years of his presidency.

The Democrats’ win was in large part a rebuke of the tech-industrial complex built under Trump’s watch. In the last two years the White House has worked to loosen regulations on giants like Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia, Oracle, and Tesla in the hope of supercharging economic growth.

While the strategy did help the US weather several self-inflicted macroeconomic storms, it also concentrated wealth and political power, eroded corporate accountability, hampered the evolution of AI safety controls, and — by extension — accelerated environmental damage resulting from infrastructure projects and data centres. All such factors drove US voters to demand greater checks and balances on the tech industry at the ballot, specifically through tighter AI regulations.

In anticipation of this reckoning, a widespread sell-off of AI stocks was triggered two days after the midterm results, causing tech share prices to nosedive. In total, more than $3.2 trillion was bitten from the US stock market, a value greater than the annual output of Canda’s economy. Given the scale of concentration risk, private credit leverage and circular financing within the AI ecosystem, almost no S&P 500 or Nasdaq firm emerged from the rout unscathed. Nvidia, the world-leading AI chip manufacturer very much at the centre of the ecosystem, saw its shares fall by over 9% — a crisis that will be compounded by the honeytrap investigation into CEO, Jensen Huang’s relationship with an alleged Chinese spy.

Asia’s markets were hit hard too. With sharp plunges in the shares of memory chip makers Kioxia and Tokyo Electron, Japan’s Nikkei share index closed down 6% on 6 November. In South Korea, the Kospi index dropped over 8% at one point, triggering a circuit-breaker mechanism, which halted trading for 20 minutes to stifle panic selling. Though the US stock market has since staged a modest recovery, some commentators have cast it as a dead-cat bounce, arguing that the sell-off was a mere tremor before a bigger earthquake yet to come.

Given the prevailing political winds, it seems all but inevitable that AI will soon face stricter regulations. With a perfect storm of public demand, surging Democrat ambition to rebalance power between the state and the private sector — as well as the UK government’s signal that it is open to work with partners, especially the US, to develop robust AI controls — tech leaders are bracing for a sharp tug on their reins.

Politicians like UK foreign secretary Ed Miliband, arguably the most strident among Prime Minister Andy Burnham’s cabinet on the hazards of AI, will no doubt be looking at the UK’s upcoming G20 presidency as a chance to advocate for synchronised international action to safeguard critical digital systems and civilian infrastructures. A case for doing so has been thoroughly made by recent AI hacks at Google, Anthropic, OpenAI and Meta.

Though desperately warranted, an international pact on AI safety may prove elusive. World powers are after all notoriously shoddy at ratifying agreements on complex, long-term, and existential risks — as the United Nations Framework Convention on Climate Change (UNFCCC) would attest to.

The UK’s world-leading AI Security Institute (AISI) could offer a model for success, but AI superpowers, like the US, will doubtless have reservations about any measure that might squander their chance of winning the global race for recursive intelligence.

Whether the AI bubble can avoid popping hinges on a few factors. First, data centres must sidestep the not-in-my-backyard issue. Does this mean space is the next frontier? The option is not without its operational challenges, though a partial Elon Musk argues it can be done. Second, can AI’s hunger for energy be satiated sustainably? Without a greener energy mix, data centres will face more powerful public opposition yet. Third, can tech giants bring sizable enough returns to make good on their debts? Convincing investors that they can will go a long way to stabilising the market and delivering on the once-in-a-century productivity promise.

It needs to be said that, in spite of cratering enthusiasm for AI stocks, most indices remain up for the year. Anthropic is sticking with its planned IPO of $1.5 trillion, data centres are still being financed, and new frontier models continue to be built. Shareholder capitalism rides on to the horizon — brazen, bold, and blind — for better or worse.