After the midterms, AI stocks and the democracy discount

Global stock markets fell sharply in the days following the Democrats’ midterm triumph, with AI-related equities bearing a disproportionate share of the losses. Analysts cited “big worries over AI” as the driving factor, a phrase that repays unpacking: it is not the technology itself that markets are suddenly afraid of, but the regulatory environment that a Congress less controlled by AI-friendly Republicans might now construct. The BBC’s analysis connected the sell-off directly to the election outcome, noting that investors had spent two years pricing in a legislative environment sympathetic to rapid AI deployment with minimal guardrails. Meanwhile, even as the election results registered, the practical policy picture was more complicated: Democrats won big on immigration as an issue, yet ICE retained its appropriated funding, a reminder that winning an election and changing policy are two very different things.

The received wisdom

The standard progressive narrative treats the sell-off as a feature rather than a bug: the market is finally reckoning with the possibility that AI will be regulated, and if some speculative froth comes off AI valuations in the process, this is a price worth paying for democratic accountability. On this view, the Trump years enabled Silicon Valley to operate as a largely unregulated utility, reaping monopoly rents while externalising costs onto users, workers, and democratic institutions. A Democratic Congress, even a fractured one, will at minimum restore regulatory capacity at the FTC, the FCC, and perhaps create new AI-specific oversight mechanisms. The sell-off is the market repricing regulatory risk — which, on the progressive reading, is precisely what the market should be doing.

There is something to this. The degree to which AI companies had come to depend on a permissive regulatory environment — not as a temporary gift but as a structural assumption baked into their valuations — was itself a governance failure. When a company’s business model requires the state to look the other way, the company has become dependent on political conditions in a way that ultimately undermines both market efficiency and public legitimacy.

A different read

And yet the celebration should be tempered. The Democratic Party’s relationship with technology regulation is, to put it charitably, complicated. The Obama administration was broadly solicitous of Silicon Valley; the Biden-era FTC under Lina Khan was more aggressive, but faced constant litigation, jurisdictional fights, and the structural difficulty of applying century-old antitrust frameworks to markets that look nothing like the ones those frameworks were designed to regulate. There is no reason to assume that a Democratic House will produce coherent, functional AI regulation rather than a decade of performative hearings, industry-drafted compromise bills, and enforcement mechanisms that look robust on paper and crumble in court.

The deeper problem is structural. Markets are pricing “Democratic regulation” as a single variable — a dial that turns from permissive to restrictive. But the actual regulatory problem with AI is considerably more specific and technically demanding than anything a general-purpose Congress is well-equipped to address. The questions that matter — how do you audit a large language model for discriminatory outputs; what liability rules should apply when an AI system gives medical advice that harms someone; how do you prevent AI-generated content from polluting democratic information environments — require technical capacity, regulatory bandwidth, and institutional patience that neither party’s congressional caucus currently possesses.

The ICE funding story is instructive here. Democrats ran successfully on immigration as an issue — it was, according to post-election analysis, a significant factor in their margins in key suburban districts. But winning a congressional majority does not, by itself, redirect appropriations that are already spent, restructure bureaucracies that have been built over years, or change the operational priorities of career federal officers. The gap between electoral mandate and policy implementation is always larger than campaign rhetoric suggests.

The same will be true of AI. The market is pricing a regulatory step-change that will, in practice, be a slow, contested, lawyerly process producing incremental adjustments at the margins. This does not mean the election does not matter for AI policy — it clearly does. It means that the sell-off, like most market panics, is probably overdone. More seriously, it means that citizens who voted Democratic partly in the hope of meaningful tech accountability should hold their representatives to account, loudly and specifically, for the gap between promise and delivery that is already visible in other policy areas.

The democracy discount, as the markets are calling it, is real. But democracies that deliver should trade at a premium, not a discount. The question is whether this particular democracy will.

What to watch

Watch the composition of key congressional committees: who chairs the House Energy and Commerce Committee and the Senate Commerce Committee will matter enormously for what AI legislation actually reaches the floor. Watch whether the new Democratic leadership prioritises AI regulation or treats it as a second-tier issue behind healthcare, immigration, and the economy — the political incentives do not obviously point toward investing legislative capital here. Watch the FTC: with restored budget and personnel capacity, it will be the first mover on AI enforcement, and its early cases will set the template. And watch Nvidia’s share price as a forward indicator: if it stabilises quickly, markets have concluded the regulatory threat was overpriced; if it continues falling, the repricing has further to run.

— J