• Bull run at risk as kneejerk policy juices inflation fears
• Treasuries yield efforts doomed without fiscal consolidation
• AI capex faltering would help inflation, dampen spirits

It’s sometimes tough to believe that US Treasury Secretary Scott Bessent ever ran a macro fund.

From an M&A and financing perspective, it’s worrying just how kneejerk his interventions seem to have become. With equities priced for perfection amid AI hopes, and inflation sticky, reckless policy action could be just the catalyst to upend a bull run lacking widespread conviction.

If stock markets correct downward, that could create headaches for megadeals employing paper consideration – while yawning spreads and falling valuations would surely herald pain for sponsors with overleveraged portfolio companies, especially those stuck in software purgatory.

Last week, this column described the US Treasury’s attempts to support the Japanese yen as conjuring an image of frailty. Little did we know Bessent would follow that up on Wednesday (19 August) with a Treasury buyback plan nakedly designed to flatten – or at least temper the vertiginous slope of – the yield curve.

That has only worsened his Wall Street cred. The yield on 30-year Treasuries fell on Wednesday but has crept back to 5.25% and remains lodged close to the top of its 52-week range. Obeying Newton’s Third Law, Bessent’s financial engineering saw the dollar fall to compensate, with the US Dollar Index down 1.2% over five days.

This was inevitable. If you run a deficit heading towards 7% of GDP and sovereign yields rise, then buying long-dated paper may deliver temporary yield relief, but only at the cost of the market Whac-A-Mole popping up somewhere else.

Bessent seemed to have cottoned on to this yesterday. “Once the market realizes we are focusing on fiscal consolidation and that we are trying to bring the market back into equilibrium in a thinly traded market, I’m confident bonds will continue to climb,” he told reporters.

Although no one buys the market equilibrium argument, Bessent is right to focus on fiscal consolidation as the necessary counter to make any Treasuries buyback intervention a success.

Indeed, credible tightening on that score might obviate the necessity for buybacks at all.

But good luck with that ahead of the midterms.

Since Jimmy Carter’s presidency, Democratic presidents have cumulatively consolidated deficits by 16.7 percentage points of GDP while Republicans have hiked them by 18.9 percentage points, according to research by the Center for Economic and Policy Research. The GOP would argue tax cuts juice growth, but commensurate spending cuts seem tough to deliver.

Indeed, a US Government Accountability Office report earlier this month found supposed savings from the initially Elon Musk-led Department of Government Efficiency program – the totemic Trump 2.0 effort to cut expenditure – were “incorrect or lack supporting evidence”.

If Bessent’s market communication is found wanting, his approach to government borrowing considered panicky, and his commitment to fiscal restraint dubbed fanciful, his ability to support Fed Chair Kevin Warsh in taming inflation will be neutered.

All of which puts even more pressure on the Federal Open Market Committee to consider raising rates next month – and for Warsh to abandon his distaste for forward guidance. Inflation expectations will lose their moorings if both the Treasury Secretary and the Fed Chair are considered either incompetent or gnomic.

Could AI yet ride to the rescue? Yes – but there’s definite scope for monkey paw-style outcomes here.

The USD 765bn in AI capex expected this year is projected to balloon to USD 1.6tn by 2031, according to Goldman Sachs. This will undoubtedly be inflationary over the period, even if the underlying technology lives up to billing by turbocharging productivity.

A key question is how solid that capex projection turns out to be.

Cut-price open-source models from China, including DeepSeek, Qwen and Kimi, alongside Facebook parent Meta’s Muse and Llama Series, sit alongside Google opening a price war within the AI subscriptions space. All of this weighs on Anthropic’s touted USD 2tn IPO valuation.

And there’s a world where this is just the tip of the iceberg.

A report published by researchers at Stanford University earlier this month pointed to local language models – hosted directly on laptops – being able to deliver “serviceable query coverage” in 71.3% of cases, up from 23.2% in 2023. If this rate of progress is sustained, the vast majority of work and social applications for language model technology could conceivably be delivered entirely by the device on which you are currently reading this article.

Should this come to pass, “hyperscalers are wasting hundreds of billions of dollars in investments,” according to a note from Panmure Gordon analyst Joachim Klement.

Which is not to say frontier models won’t have their place: Moderna’s 134% share price run this week off the back of its astonishing melanoma vaccine breakthrough alongside Merck is at least partly thanks to AI analyzing tumor sequencing data. We can expect further such announcements in the years ahead. AI is far from a busted flush.

But it seems the hyperscalers are rushing to build hundreds of Ferrari factories when most customers will settle for a Nissan Versa.

Yet if cooling in the AI market does come to Bessent and Warsh’s rescue in tempering inflation, it could be at the cost of a fallout in the private credit markets financing data center rollout and a wider curtailment of animal spirits.

Despite occasional briefing that he wasn’t fully on board with President Trump’s tariffs agenda, Bessent’s MAGA-coded public persona across not only trade policy but also subjects most predecessors would have steered clear from – such as Iran and the Greenland imbroglio – are a far cry from the market orthodoxy displayed by the likes of Tim Geithner or James Baker.

At a time where bullish enthusiasm, tech uncertainty and geopolitical flux coexist, clarity and surefootedness from policymakers are essential. That seems lacking.

Corporates and sponsors considering their financing and footprint options would do well to ensure they’ve truly modeled adverse scenarios before putting capital to work. Or they risk looking as clumsy as Bessent does right now.

Continental Drift is a weekly column offering commentary on the macroeconomic, political, and policy forces shaping the M&A landscape across the US and Europe. The opinions expressed here are those of the writer only.