UBS: don't bet on new US AI rules after the midterms — AI spend is now nearly 2% of GDP

• Who: UBS, in a report dated October 22, 2026.
• What: major new US regulations on AI are unlikely to emerge, even with AI a central issue in the midterm elections.
• Politics: UBS says AI is "unlikely to sway the ultimate outcome" — Democrats retaking the House and possibly the Senate; "notable new regulations are unlikely to materialise because of the midterms."
• Money: hyperscaler capex projected at nearly 2% of US GDP in 2026.
• Growth: AI investment plus the equity wealth effect accounted for around 100 basis points of H1 2026 growth, when the US economy grew 2.4%.
• Counterpoint: OpenAI and Anthropic have called for stronger safety rules; the Trump administration's October 2026 voluntary pact with six companies carries no legally enforceable requirements.

UBS published a report on Thursday, October 22 arguing that major new US regulations on artificial intelligence are unlikely to emerge — even though AI has become, in the bank's words, "one of the central issues of the mid-term elections." The reason is economic: "much is riding on AI providing a long-term boost to the US economy via higher disinflationary growth," the report says, so a light-touch regulatory approach is likely to persist.

The report expects the midterms to change little. AI "is unlikely to sway the ultimate outcome, which should result in the Democrats retaking the House and possibly the Senate," UBS wrote — and "notable new regulations are unlikely to materialise because of the midterms."

The economic math behind the call

UBS leans on the numbers. Capital expenditure by major cloud computing companies — the hyperscalers — is projected to reach nearly 2 per cent of US gross domestic product in 2026. And the bank estimates that AI investment, together with the positive wealth effect on consumption from rising equity prices, accounted for around 100 basis points of US economic growth in the first half of 2026, when the economy grew 2.4 per cent.

The report even points to bond markets: some evidence, it says, that developments in AI models correlate with lower interest rates — a "positive sign," in its framing, for AI's economic implications.

Washington's current playbook is voluntary

That light-touch stance matches the Trump administration's approach. In October 2026, Trump announced a voluntary agreement with six major technology companies, including OpenAI and Anthropic, aimed at addressing AI safety concerns — internal safety controls, monitoring systems and independent assessments, but no legally enforceable regulatory requirements.

The debate is not one-sided. Leading developers, including OpenAI and Anthropic, have called for stronger safety measures and oversight of advanced AI systems — even as the administration prioritizes reducing regulatory barriers to keep American leadership in AI.

What it means

For the industry, UBS's read is a green light to keep building under the existing rules: no new federal rulebook waiting on the other side of the midterms, with the economic argument — AI as a disinflationary growth engine — doing the political work of holding regulation back. Whether that bet survives contact with voters, or with the next AI incident, is the part no report can price in.