Listen to the audio version of this article (generated by AI).
The Lower-K consumer is falling behind… 5 trillion reasons why Wall Street hasn’t cared… what will spark the reckoning… watch growing anti-AI legislation… exactly when it all ends
Two weeks ago, the University of Michigan’s Consumer Sentiment Index sank to 48.2 – the lowest reading in the survey’s 74-year history.
That’s a lower consumer sentiment reading than during the Great Financial Crisis. Lower than the Dot-Com Crash. And lower than every recession in modern American history.
Meanwhile, last week, the AI semiconductor company Cerebras Systems went public at a $100 billion valuation after being 20 times oversubscribed.
Contrasting this massive IPO against the glum sentiment report, our technology expert Luke Lango, editor of Innovation Investor, put it plainly in his Daily Notes:
This [Cerebras IPO] is America Two. It has no idea America One exists.
He’s right.
But for AI investors, understanding why he’s right – and exactly when it’ll stop being true – is the most important portfolio question of the next three years.
America One: The diagnostics
Let’s start with the data…
Regular Digest readers are familiar with our K-shaped economy, where lower-income households face a fundamentally different financial reality than their asset-owning counterparts. That picture has gotten materially worse since we last covered it here.
According to the New York Fed’s quarterly household debt report, in Q4 2025, overall household debt delinquency rates hit 4.8% – the highest level since 2017. The stress is concentrated almost entirely in lower-income borrowers.
Meanwhile, the subprime auto market – one of the most reliable canaries in what we call the Lower-K economy – related to what Luke is calling “America One” – sounds a little quiet these days.
More than 6% of subprime auto loans are at least 60 days past due, the highest rate ever recorded in Fitch data going back to 1993. Vehicle repossessions hit 1.73 million last year, the most since 2009.
Meanwhile, a PYMNTS Intelligence survey from early 2026 found that need-driven paycheck-to-paycheck living has overtaken choice-driven for the first time – meaning financial pressure, not lifestyle, is now the dominant explanation.
This is not a soft patch – it’s a structural squeeze.
America Two: Why Wall Street is largely unbothered
Here’s where most coverage of the K-shaped economy goes wrong…
The typical framing is that Wall Street is ignoring the Lower-K data. That’s not quite right.
The more accurate read is that Wall Street is accurately pricing an economy in which the Lower-K consumer has become structurally less relevant to corporate earnings. It’s less “callousness” and more “arithmetic.”
The AI multi-billion-dollar capex loop that I highlight regularly – the engine powering the Upper-K/“America Two” economy – doesn’t run through a fast-food customer or a subprime auto borrower. It runs through enterprise contracts, sovereign AI deals, and hyperscaler balance sheets.
So, what’s the status of that tsunami of money flow?
Bloomberg Intelligence projects cumulative AI capex of $5 trillion over five years. For context, that’s more than the entire GDP of the United Kingdom.
Back to Luke:
The AI economy has achieved a complete structural decoupling from the consumer economy. The Iran War didn’t slow it. The all-time low consumer sentiment doesn’t slow it.
The only economic variable that affects the AI buildout is whether Nvidia can manufacture enough chips and whether the grid can supply enough power.
Everything else is noise.
So, for now, if your portfolio is positioned in the right layer of the AI trade, this flagging Lower-K/ “America One” is largely just background noise.
But the thing about structural decouplings is that they don’t last forever. History suggests that economic equilibria that become this far out of balance don’t stay that way.
And what brings that back into alignment isn’t always markets…
Sometimes it’s politics.
What could end the AI trade
So far, the much-predicted wave of AI-driven mass unemployment hasn’t materialized.
That may yet change – my hunch is that it will. But it doesn’t have to – at least not to generate the political backlash that eventually threatens the AI trade. It just requires enough people feeling enough economic pain and then connecting that pain to AI.
Enter your power bill…
According to the nonprofit PowerLines, electric and piped natural gas bills were among the largest drivers of inflation last year, rising 7% and 11%, respectively. Utilities requested a record $31 billion in rate hikes in 2025 – more than twice the amount requested in 2024.
Here’s Charles Hua, executive director of PowerLines:
There are millions of Americans who are paying 10% to 20% of their incomes just on their utilities, which would be unfathomable for the vast majority of Americans.
This is turning data centers into a kitchen-table political issue.
Take Democratic Pennsylvania Gov. Josh Shapiro – a 2028 presidential hopeful. He initially embraced the data center boom in his state. Then the public pushback mounted.
In his February state budget address, he reversed course:
We need to be selective about the projects that get built here.
I know Pennsylvanians have real concerns about these data centers and the impact they could have on our communities, our utility bills, and our environment.
And so do I.
It won’t stop with Shapiro in Pennsylvania.
Here’s Hua, speaking to Fortune:
You could argue utility bills will play the most prominent role in a national election this year that perhaps at any other election in American history.
The AI boom is being partially subsidized, on a monthly billing cycle, by the same Lower-K households already squeezed by gas prices, negative real wages, and rising delinquencies.
That generates a specific, personal grievance. And specific, personal grievances often become votes.
The backlash is moving from grumbling to legislation
This morning brought the following headline from The Wall Street Journal:
And here’s the subhead:
Back in February, Axios surveyed the landscape and found that only 7% of Americans believe AI will increase jobs – statistically unchanged from the prior fall, meaning the boom’s rising visibility has done nothing to ease displacement fears.
But sentiment – while critically important – is no longer the main story. It’s what the sentiment becomes when crystalized…
Legislation.
In the first six weeks of 2026, over 300 data center bills were filed across 30+ states – a clear shift from incentive-focused policies to regulatory oversight.
No state has yet enacted a statewide data center moratorium, though Maine came closest, passing one through both chambers before the governor vetoed it in April.
But the movement is finding its footing at the local level. Seattle announced a 365-day emergency moratorium last week, resulting in this headline from the Seattle Times:
Denver and Minneapolis vote on their own moratoriums this week. Camden County, Georgia, passed a nine-month ban on May 5.
And according to Good Jobs First, grassroots opposition has already blocked or delayed $156 billion in data center projects across 40 states in just over a year.
The direction of travel is clear.
Since we’ve mentioned AI and “jobs,” let’s address that too…
The California Labor Federation has pledged to support more than two dozen AI-related worker protection bills this year.
California’s SB 951 would amend the state’s WARN Act to cover AI-driven displacement, requiring 90 days’ advance notice before automation-related layoffs plus disclosure of the specific AI system used.
And California’s AB 2027 would prohibit employers from using worker data to train AI systems designed to replace those same workers.
This last bill is, in legislative form, a direct response to the “train your replacement” dynamic I highlighted in our May 5 Digest – the deliberate workflow documentation programs that preceded Oracle’s 30,000 layoffs and Meta’s 8,000 cuts this spring.
Put it all together, and here’s Luke with where this backlash takes us:
The force that will derail the AI Boom is not a technological failure, demand collapse, or even a recession.
It is politics – specifically, a populist backlash against AI that is already building momentum, fueled by the growing economic pain hitting American households right now.
So, when does all this hit your portfolio?
Luke projects right around the 2028 presidential election cycle.
His case rests on three compounding pressure points: rising energy costs from data center construction landing directly on residential electricity bills… accelerating AI-attributed layoffs across major employers… and widening wealth inequality that is visible, measurable and personal to the households experiencing it.
Luke says that by 2027, anti-AI messaging will have become a dominant political narrative. That will result in AI-curbing legislation – taxes, restrictions on data center construction, and labor displacement provisions.
And that, according to Luke, is when the curtain falls:
That is the scenario that ends the AI Boom. And it is not a remote tail risk.
Make your money now.
The window for transformational wealth creation in this AI cycle is the next two to three years.
This isn’t a bear call. It’s the opposite – an urgent bull call with a specific expiration date. As Luke put it:
This trade will not last forever. Like everything, it has an expiration date.
For how Luke is playing the AI trade while the window is open – including his latest research on what he believes could be Elon Musk’s most ambitious project yet (it has nothing to do with Tesla or SpaceX) – click here for his full presentation.
As we begin to wrap up, two things to consider
The first could accelerate this trade, while the second could slow it considerably.
The visibility of Luke’s political clock – the fact that sophisticated investors can now see the 2028 timeline coming – may actually pull capital and returns forward.
If the window closes in two to three years, the rational response is to accelerate into it now, not retreat from it. That’s part of what you’re seeing in the Cerebras oversubscription, the SpaceX IPO queue, and the latest Tesla-related opportunity Luke has found: capital racing to get positioned before the friction arrives.
The clock’s visibility doesn’t slow the trade. It intensifies it – right up until it doesn’t.
Meanwhile, the other issue is a wildcard – or perhaps we’ll call a “Trump” card…
In March 2026, the White House published a National AI Legislative Framework calling on Congress to preempt state AI laws that “impose undue burdens” – a direct attempt to neutralize the California bills, the data center moratoriums, and the entire state-level wave before it reaches critical mass.
If that federal preemption push succeeds, Luke’s political clock extends considerably.
We’ll keep tracking both.
So, what does all this mean for your portfolio today?
The AI trade – taking a breather now – is working. The capex loop is intact and growing. What Luke has called “the Summer of AI” appears to be underway.
But this isn’t a forever trade…
The Lower-K’s deteriorating financial health isn’t a risk to the AI trade today. But it is the kindling for tomorrow.
Meanwhile, rising electricity bills, stagnant wages, record delinquencies, and all-time low sentiment – none of that threatens Anthropic’s Google Cloud deal or Cerebras’ IPO. But it creates the conditions for when the spark of political anger strikes and ignites that kindling.
Luke’s bet: that happens right around 2028.
Two Americas. One trade. And a clock that’s ticking.
Have a good evening,
Jeff Remsburg
P.S. While the political clock ticks on the AI trade, Louis Navellier is focused on what’s happening right now…
And he believes a rare window is opening in a corner of the market most investors are overlooking entirely thanks to the Fed and new chair Kevin Warsh.
In his latest research presentation, Louis walks through exactly what his system is seeing and why the timing matters, and he even shares a free stock pick tied to this opportunity.
We’re taking Louis’ research video offline tomorrow, so if you want to catch it before it comes down here’s your chance.
