Listen to the audio version of this article (generated by artificial intelligence).
Albany pulls the plug… Ford’s ‘AI Boomerang’ in Jobs… Flashing Light on Margin Debt… Why We’re Not Worried About AI Trading Yet… And How to Trade Either Way
Yesterday, New York Governor Kathy Hochul signed an executive order prohibiting the construction of new large-scale data centers—those that consume 50 megawatts or more of power—for up to one year.
New York is now the first state in the country to impose such a ban, although at least 15 other states are considering or actively developing legislation to restrict, consider or temporarily prohibit the construction of new data centers.
Hochul framed it as a matter of survival for taxpayers:
We are in the midst of one of the most significant economic upheavals in generations… and perhaps ever.
Hyperscale AI data centers consume huge amounts of energy…
They raise costs on local taxpayers, and I refuse to allow those costs to be passed on to New Yorkers.
This isn’t just a New York issue — it represents a national anti-AI mood that is increasingly framed in political and moral terms.
Back in April, Sen. Bernie Sanders (D-Vermont) said:
AI oligarchs don’t just want to replace specific jobs.
They want to replace workers.
This is a highly provocative assertion that can easily be argued. But even if that were true, these “AI oligarchs” would still have to contend with something more powerful than themselves: the agent.
Laying off too many suitable employees leads to low quality – and the client retaliates long before any politician does.
About a year ago, Ford (F) CEO Jim Farley made a headline-grabbing prediction
He said:
Artificial intelligence will replace half of white-collar employees.
While Ford subsequently cut parts of its workforce and turned to AI during the broader restructuring process, it turns out that AI couldn’t finish the job alone.
Over the course of an ongoing three-year initiative to fix costly vehicle defects, Ford has rehired 350 veteran engineers — internally called “graybeards” — to fix quality issues that its AI-driven design systems couldn’t detect on their own.
The fix worked: Ford just topped J.D. Power’s preliminary quality survey for the first time in 16 years.
Ford is not alone in its return to human workers.
ibm (IBM), Starbucks (Sex), McDonald’s (MCD), Air Canada, and Commonwealth Bank of Australiaamong others, have all had similar setbacks recently after a drop in quality.
But the most dramatic example may be Klarna (clear)the Swedish buy-now-pay-later giant. It cut 22% of its workforce and boasted that its AI chatbots could do the work of 700 human agents – before quietly launching a recruitment drive to bring humans back into customer support after service quality collapsed.
Workforce firm Careerminds found that two-thirds of companies that made AI-driven cuts last year have already rehired, more than half of them within six months.
This pattern already has a nickname: “AI Boomerang” — cutting down AI, watching difficult cases that require a lot of judgment pile up, and then rehiring humans at a higher cost.
Nearly a third of companies that rehire end up spending more than they originally saved, because replacement roles in the age of AI now command a pay premium of 20% to 25% compared to the roles they replace.
However, the broader jobs data tells a complex story
The Challenger, Gray & Christmas report for June showed that total layoffs in the US fell sharply – down 53% from May.
But AI was still the top reason cited for job cuts that month, the fourth straight month it topped the list. AI-related cuts now represent nearly 23% of all announced this year.
Meanwhile, an MIT study using a tool called the Iceberg Index suggests that AI can already do work equivalent to 11.7% of the US workforce — about $1.2 trillion in wages.
So far, actual layoffs represent only a small portion of this exposure, but the potential for more is clearly there.
Put these two threads together, and you get the true shape of this story…
Fear of AI is currently progressing well ahead of AI layoffs.
Our technology expert, Luke Lango, has been tracking this broad anti-AI sentiment
Luke, editor Innovation investorhas argued for months that the thing most likely to end this AI bull market is not a technological stumble — but a policy one:
The force that could derail the AI boom is not technological failure, a collapse in demand, or even a recession.
It’s politics – specifically, the populist backlash against AI, that has already begun to build momentum, fueled by the growing economic pains hitting American families right now.
It is on track to reach full power in the 2028 presidential election period.
Yesterday’s New York data center moratorium is exactly the kind of early earthquake that Locke predicted. Rising energy bills, viral outrage, and politicians willing to act on it, are shaping up as fierce headwinds against AI.
But Luke isn’t sitting on the sidelines waiting for that day. He remains optimistic about AI today because he believes we are in the strongest stage of prosperity.
At the same time, he constantly monitors the political landscape, corporate spending, and market leadership for signs that the investment landscape is starting to change. This allows it to remain in place in companies that continue to benefit from the rapid expansion of AI today – while preparing subscribers for the eventual shift before it becomes apparent to Wall Street.
If you’d like to follow Luke’s research and see which AI companies he believes are best positioned at this point in the cycle, You can learn more about it Innovation investor here.
So where does all this leave investors?
We still take AI trades on a long basis – but we watch the calendar as closely as earnings.
But given the long trading period, AI trading is not looking very good these days, and there is a new reason for caution
Yesterday we learned that margin debt has just reached a level only seen at previous market tops.
Let’s go to CNBC:
According to data from Leuthold Group, margin debt has grown by more than 40% over the past 12 months, a threshold seen at previous market peaks in 2000, 2007 and 2021.
To make sure we’re all on the same page, margin debt is money you borrow from your broker, using your existing shares as collateral, to buy more shares. It harvests your gains on the way up – but magnifies your losses on the way down.
What is unusual this time is the relative pace of the market itself…
The S&P 500 has risen about 22% over the past year, including dividends — roughly half the growth rate of margin debt. Simply put, investors are borrowing money faster than stocks actually rise.
Scott Opsal, chief investment officer at Leuthold, was live with CNBC About the monuments:
When people start doubling down on their borrowed money, that’s a contrarian sign that’s really hard to overcome…
This is a very bearish look.
Opsal has a theory about where the borrowed money goes: artificial intelligence trading.
He points to the recent boom in leveraged ETFs — assets in those funds nearly doubled in just two months this spring.
This is important because concentrated bets amplify themselves on the way down as well. If one AI infrastructure or data center stock crashes, the margin calls that follow can ripple through every other investor relying on the same trade — forcing sales that have little to do with the underlying business and everything to do with the broker demanding more collateral.
To be clear, this does not mean that the AI business is over. But coupled with the political backlash we were tracking above, this is another sign that this bull market may be entering a more fragile and heavy phase.
As always, make sure you know what you own, why you own it, and what your exit strategy is if you’re not in it for the long term.
If this makes you panic, take a breath…
All of the above is a reason for caution, not unbridled fear.
A recent sharp pullback is not unusual – it is often just a toll booth on the way through a bull market, especially in the group leading that market.
This kind of decline can feel like the beginning of the end. But let’s be clear about where the damage is concentrated, and then put it in context…
the PHLX Semiconductor Sector Index (Sox) — a proxy for AI trading since half of the companies are AI masterminds — is down about 18% from its last peak in June.
It’s been shut down again today though Taiwan Semiconductor (TSM) It announced a record 77% increase in quarterly net profits that beat Wall Street expectations. Investors are concerned about a 15% increase in the 2026 capital spending plan, reinforcing sector-wide fears that excessive spending will hurt profit margins in the long term.
Now, this pullback is certainly painful, but it is also much lower than the 30%+ drawdowns that set previous sector-wide peaks in 2000, 2008, and 2018.
In addition, we must remember the almost vertical ascent that preceded it (trend lines added for perspective).
From that perspective, we’re just working to get rid of some of that excess.
Now the skeptic can say:
Okay, Jeff, but what’s to prevent the SOX’s current 18% decline from becoming a 30% crash next month, and then a 90% crash by Christmas?
Well, I can’t guarantee you that won’t be the outcome, but history argues against it.
Studies of S&P 500 volatility since 1929 show that only about 39% of corrections deepened into a bear market of more than 20% – and in the decades after World War II, that rate fell to roughly 25%.
Most corrections stop in the mid-teens and recover within a few months, which is almost exactly where SOX is today.
Bottom line: If you are headed for this correction, there is reason for optimism.
However, if you want to be more selective in how much exposure you have to today’s volatile market, we have an idea for you…
Let historical data guide your trading.
This morning, legendary investor Louis Navellier and TradeSmith CEO Keith Kaplan went live Breakthrough 2026 It happened.
In short, rather than predicting where the market or stock will go next, Keith’s team studied decades of price history across nearly 5,000 stocks, looking for historically favorable windows when those stocks would rise or fall with real consistency — in bull markets and bear markets alike.
Running an 18-year backtest, trading just within those windows produced overall growth of 857%, more than double the S&P 500 over the same stretch, and the strategy still outperformed in 2007, the worst year in the backtest.
With this approach, you can Pick and choose your pickskeeping as much money as you want on the margin. You only take advantage of trading opportunities when you decide – and for how long – and all according to historical data.
If you missed the broadcast We have a free replay available for you here.
We’ll keep you updated on all these stories here at digest.
I wish you a good evening,
Jeff Remsburg
