There are few things as jarring as being fully committed to an episode of, say, 60 Minutes or Landman when the TV audio suddenly cuts out and is replaced by that shrill, growling tone that announces a test of the Emergency Broadcast System (EBS). Whoever picked that sound was really good at their job, because even after decades and decades of hearing it on a semi-regular basis, that sound never fails to make the hair on my arm to stand up (and, as a child of the Cold War, I still suffer a split-second of terror as my little animal brain tries to convince the rest of my body that 'maybe this time it's for real...').

Of course, that's exactly the effect the chooser-of-the-tone was aiming for, because in an actual emergency, the authorities are counting on that horrible noise to shock us out of our stupor and pay attention to whatever they have to say next. It's a tidy little system, even at the ripe old age of 62 (the first broadcast of the EBS was evidently made in 1962).

As I write this on Monday morning, markets are having their own little test of the EBS. But instead of testing the public's willingness to duck and cover against an onslaught of Russian missiles beneath the plywood and plastic of a schoolroom desk of suspect integrity, what's instead being tested here is the market's ability to shelter in place against a barrage of cheap and effective Chinese AI models, beneath market multiples of perhaps equally suspect integrity.

Here's what happened: Over the weekend, somebody noticed that an app from a Chinese AI start-up known as Deep Seek had been downloaded millions of times through Saturday, suddenly making it the most-downloaded app in the U.S., U.K., Canada, China, Singapore, and Australia.¹ Why does this matter? Because Deep Seek claims not only to have developed functionality that rivals AI standard-bearers like ChatGPT, but also to have accomplished it without all the massively expensive, time- (and resource-) consuming effort that other AI incarnations needed to get to the same point in their development. Most notably, Deep Seek seems to have arrived at an impressive level of functionality without the help of next-generation computer chips that have heretofore been assumed to be absolutely necessary for the effective training of new AI models.

If that proves to be true (and I'm not nearly smart enough to even venture a guess as to whether it is), it threatens to undermine a key assumption about the AI supply chain that has kept a big chunk of equity markets humming for months and months: Namely, that to take full advantage of the transformative promise of AI, the global economy must first invest massively to build the data, computing, and energy infrastructure needed to support its development. Even if it keeps the economy-boosting elements of the AI story writ large fully intact, the sudden emergence of a competitive AI model that claims to have side-stepped all that places at risk nothing less than the huge and ongoing capital spending boom that has helped convince certain portions of the market that it's okay to be expensive as long as the payoff is real.

Such a test of the key underpinnings of the AI wave was probably always inevitable. No innovation worth the paper it's financed on ever proceeds in a straight line, and challenges to its underlying assumptions are part of the price we all pay to play. And that's interesting in and of itself. But to me, a far more interesting — and practical — test is whether or not the broader economic narrative will be taken down as a result of this shaking of the AI faith. Said a little differently, if the market (and, by extension, the US economy) is currently a one-trick pony built entirely on the electric dreams of an AI investment super-cycle, then the sudden emergence of a cheap young upstart like Deep Seek might well be capable of taking it down. If, on the other hand, the economy is more durable and diverse than that, then the expansion should prevail.

I'm of course hoping for the latter. But make no mistake, this morning's test of the market's resolve in the face of what at first glance appears to be a legitimate challenge to the AI super-cycle thesis is an important one. On one hand, if this morning's sell-off becomes broad-based and long-lived, it might show that those who have long been concerned that recent gains have been too concentrated among a small handful of AI-related plays were right to be concerned. On the other hand, if the damage can be mostly contained to a handful of AI pure plays, it will stand as compelling evidence that the economy's recent strength is more durable and broad-based, and the economy could be said to have passed its test. So stay tuned, because it will be important to see how companies, industries, and sectors far removed from AI's glow respond to Deep Seek's challenge.

Moving on... Yyes, for now, the economy still seems to be humming along mostly just fine. Fourth-quarter earnings results haven't been without a few hiccups (last week's guidance from the likes of Texas Instruments and American Airlines and a miss from rail operator CSX ran counter to a generally strong tone, for example), but results so far have mostly delivered on investors' high expectations. With a bulk of earnings reports due in the next several weeks, maintaining that positive momentum will be crucial to the market narrative.

Last week's scheduled economic releases were broadly similar (namely, strong but with a few hiccups.) Wednesday's update of the Index of Leading Economic Indicators from the Conference Board dropped marginally, but the study's authors went out of their way to say that they see "fewer headwinds to US economic activity ahead," and that their model is no longer signaling a risk of recession.² (Still unresolved though, is the yawning gap between leading- and coincident indicators, but I digress...).

Meanwhile, weekly jobless claims remain comfortably low, but continuing claims rose to their highest level in three years.³ That squares well with the observation that while the pace of layoffs remains subdued, hiring activity isn't exactly robust. That's enough to remain watchful for a sudden turn in the labor market that would undoubtedly catch the market's (and the Federal Reserve's [Fed]) attention if layoffs should suddenly spike. (As an aside, California's data was estimated rather than physically tallied last week and showed very little impact from the wildfires. That, too, will bear watching in future releases.)

On the productive side of the economy, regional Feds like Kansas City's Survey on Manufacturing and the Philly Fed's similar read on services providers within its territory were a little on the weak side, but not massively so.^4,5 Ditto for last week's flash Purchasing Manager's indices, which held in fine, but nonetheless executed a classic "pincer move," with manufacturing activity climbing its way back to neutral while services - and although still growing, tipped noticeably in the other direction. That's not great news, because the services sector remains the bulkiest portion of U.S. economic activity by far. Notably present in nearly all these releases, too, were mentions of price pressures that frankly remain too persistent for comfort.

Which brings us to our last stop on the hit parade this week: Friday's consumer sentiment release from the University of Michigan (UofM). Consumers are feeling slightly less upbeat about the direction of the economy than they were even just a few weeks ago, and interestingly, that view was driven by a less-sanguine view of the labor market (didn't see that coming...). But that's not even the important part: Instead, what probably matters more is that respondents to the UofM survey didn't back down from the 3.3% estimate of year-ahead inflation level that caught everyone's attention a few weeks ago. This week's report even hinted at panic-buying — a vicious cycle in which consumers rush to by things before prices ratchet higher, thereby accomplishing the very thing they were trying to avoid in the process — inflation.⁶

And that ultimately could prove to be just as important as whether or not Deep Seek was really able to pull off the impossible: A viable AI model built on the cheap, without the help of the market's heavily-spent-for AI All Stars.

What to watch this week

Fears that "panic buying" might be finding its way into the psyche of consumers, together with quiet but persistent inflationary overtones in other economic releases, will certainly grab the Fed's attention when it meets this week to consider the path of interest rates. As usual, the Fed's rate-setting committee will convene on Tuesday and issue their decision at around midday on Wednesday. Consensus is calling for a hold, and I see little reason to disagree. More important will be the tone of the post-decision statement and how well Chairman Powell navigates the Q&A session that follows. Watch specifically for any mention that committee members are overly concerned about the possibility of government policy creating an inflationary feedback loop that might cause them to re-think their future willingness to ease rates.

A close second to the Fed will be a number of important earnings releases, including Microsoft and Meta (both Wednesday), Apple (Thursday), and several companies in and around the semiconductor ecosphere. Commentary and forward guidance from these and other firms suddenly became a little more interesting given this morning's cost-quake in the AI space (for example, will Meta's Mark Zuckerberg — who last week hinted that his firm would spend $65 billion to fund AI research this year — change his tone at all in response to Deep Seek's professed minor miracle?). The future of the AI narrative may hang in the balance... .

Other results expected this week are perhaps less sexy but no less important. Starbucks, General Motors, and Southwest Airlines will provide more insight into the minds of consumers, while payments firms Visa and Mastercard — both expected Thursday — might show how well consumers' finances are holding up. A perennial favorite of mine, Caterpillar, will show how well global demand for things like commodities is holding up when it reports on Thursday, while Friday's trio of oil majors will do the same for oil (watch, too, for any hints that optimism related to the political environment has colored their future view.)

From a scheduled economic release perspective, Friday's income and outlays report is probably the most interesting thing to watch. As discussed above, UofM's consumer sentiment release specifically called out the idea of panic-buying as a tangible risk to inflation. If consumers are indeed rushing to buy stuff before prices run away from them, it might show up here, in Friday's release. If it does (that is, if consumer spending suddenly and unexpectedly inflects higher), then listen for the chorus of economists warning of a more reluctant Fed. If not, breathe easier — last week's UofM might prove to just be another example of consumers telling survey-takers one thing, while acting differently with their wallets.

Speaking of consumers, we'll also get a separate view into consumer attitudes on Tuesday, when the Conference Board releases its survey of consumer confidence. In addition to any insight into how consumers are feeling about inflation, watch the Conference Board's data on jobs ("plentiful" versus "hard to get") for confirmation of the UofM's suddenly less-upbeat view on employment.

We'll also get new data on the state of the housing market, beginning with new home sales on Monday, two reads on home prices on Tuesday, and pending home sales on Thursday. Last week's existing home sales data was surprisingly strong, and commentary from the National Association of Realtors' economist, Lawrence Yun, was upbeat.⁷ For housing to regain traction, however, you'll need to see employment trends hold up, wages continue to grow, and affordability improve. Even the most dedicated optimist would have to agree that it could be challenging to get all three pieces to align in the immediate future.

Finally, this week will bring several reads on the overall health of the productive economy. On tap this week are regional manufacturing surveys from both the Dallas- and Richmond Fed, as well as one of the oldest and most well-respected purchasing managers' surveys in the form of the Chicago Purchasing Managers Index, due Friday. For trends across a wider range of economic activity than just manufacturing, Monday's National Activity Index from the Chicago Fed - a broader read than last week's Leading Economic Index - is worth a glance. But for those who prefer to get their data on growth straight from the horse's mouth, the Bureau of Economic Analysis' first estimate of fourth-quarter Gross Domestic Product is expected Thursday. Expect it to show the U.S. economy has continued to expand at a reasonably robust rate.