Government Shutdown? Maybe, but we'd still look at tech

The three things investors need to know this week:

  1. The AI boom continues strong, with a significant pick up in investment deals during the past three weeks.

  2. US inflation keeps creeping up, but not nearly at an alarming pace – yet.

  3. The US government is likely to shut down this week. Most shutdowns are not usually consequential. This one could well be.

By George Lagarias


---------------------

Summary

A potential U.S. government shutdown looms by midnight September 30, 2025, risking a 0.1-0.2% GDP loss per week, with the 2018-19 shutdown’s $3bn hit as a precedent. However, its economic impact may be overstated, at least compared to the AI boom’s momentum, with $125bn in recent deals (NVIDIA, Oracle) fuelling Magnificent Seven outperformance. Investors should prioritise AI consolidation’s long-term growth over transient shutdown noise, as inflation remains manageable despite slight upticks. Focus on AI-driven opportunities and monitor inflation, not temporary government gridlock, for strategic decisions.

--------------

Two themes have defined equity markets since Chat GPT first spoke in human language, back in December 2022: The outperformance of US tech companies versus everyone else (just ten US stocks are responsible for 55% of the rise of all US large caps since 2022), and the lower-dollar/higher inflation which has favoured US large caps equities, while sending jitters in bond markets. 

Last week saw another outperformance from the Magnificent Seven stocks in the US, as NVIDIA pledged to invest $100bn in OpenAI, which has now moved away from the non-profit model. This comes a week after Nvidia pledged to invest $5 5 in rival Intel and Oracle’s $20bn investment in OpenAI. The recent deals, as well as regulatory protectionism on behalf of the White House in the past few weeks, have once again propelled Mag 7 valuations above their near-term average (and very near their 3y average) premium versus the rest of the stock market. This suggests that the AI boom is nearing its mature phase, but it's not quite there yet. A correction is likely, of course, but this theme is a very real and ongoing one.

Meanwhile, the inflation scare in the US is somewhat dissipating. Despite a small rise in core inflation (Core Consumption Expenditure) in August, for the time being, inflation remains well in check for the world’s biggest consumer market. 

Having said that, we should always be mindful of this graph from Goldman Sachs. Research at the beginning of the summer suggested that by Q4, consumers would bear 2/3 of the tariff onus (which is another way of saying that inflation will eventually rise more sharply-the question is how temporary it will be). Inflation risks, however, are on the upside (higher inflation than anticipated).

At the same time, markets are by and large dismissing the effects of what looks like a very likely US government shutdown. If Congress fails to agree on a continuing resolution by midnight tomorrow (Tuesday, 30th September), then the US government will partially shut down. The effects of such a shutdown are not negligible, but they are usually considered transitory. On average, economists calculate a 0.1-0.2 per cent of GDP loss for every week of shutdown. However, employees who have not been paid are usually refunded for the whole time they were furloughed, so growth rebounds next quarter. Since 1991, there have been 14 government shutdowns, most of which lasted for a few days, with a relatively anodyne overall impact. 

Having said that, the more relevant is the December 2018-early 19 shutdown, which occurred during the previous Trump administration. This was the least anodyne. It lasted a record 35 days and it is estimated to have caused the total loss of $3bn from the ($29.2tn economy), in lost government contracts and R&D spending. Additionally, it contributed to a repo crisis in September 2019, a money market freeze which forced the Fed to intervene. The Treasury had initially depleted 20% of its operating account with the Fed ($100bn out of $500), and when the government was restored, it sought aggressively to build a $750bn buffer, draining liquidity from money market funds. The Fed intervened very quickly, but the market remained jittery. In the end, the pandemic forced a much bigger intervention by the central bank, so the consequences of this (and all manner of economic sin) were ultimately overshadowed by the Covid-19 crisis.

So shutdowns are not usually consequential, but they can be.

Why is 2018-19 more relevant? 

  1. Because the President is the same and has pursued an even more aggressive domestic agenda.

  2. Because, like then, neither side shows any willingness to negotiate or compromise.

  3. Because the White House has indicated that it may take the opportunity to dismiss, rather than furlough, key government employees.

  4. Last year, when Senate minority leader Chuck Schumer cooperated to get a funding resolution passed, he was widely lambasted by his own party. That could make Democrats more militant this time around.

However, there is one key difference. In 2018, the Democrats had just won a midterm election in the House (Republicans still controlled the Senate) and were empowered to shut down the government. The House controls the budget, and it was Democratic. This time around, they are a minority in both chambers with only the power of the filibuster (a procedural obstacle in the Senate), which may not legitimise such radical action in the eyes of voters, and set an even worse precedent for lack of bipartisan cooperation. So, while many signs point to a consequential shutdown, we would remain more reserved as to how far the Democratic Party is willing to drag this out. 

What businesses and investors need to know.

A potential US government shutdown will likely make a lot of noise and take the focus off the recent AI-related dealmaking. Even if the shutdown turns out to be consequential, it’s not the only economic game in town. In fact, it’s really not the biggest. It could, of course, in extremis, contribute to a sharper-than-expected economic downturn, but it’s hardly the stuff of recessions. Instead, we would focus on the economic and financial impact of the hundreds of billions of dollars of AI consolidation in the past few weeks, or the generally less pessimistic than anticipated signals from the US economy.