Alarming ethical decisions from AI… Trump wants to remove state regulations… the “AI Autobahn”… why it might be time to tamp down Mag 7 expectations… will 2026 be the year of the bear?
If your job was on the line, would you blackmail a coworker to save it?
In June, the AI company Anthropic tested simulated scenarios across 16 major AI chatbot models (OpenAI, Google, xAI, etc.).
It found “consistent misaligned behavior” – including blackmail…and worse.
From the Anthropic report:
Models that would normally refuse harmful requests sometimes chose to blackmail, assist with corporate espionage, and even take some more extreme actions, when these behaviors were necessary to pursue their goals.
The consistency across models from different providers suggests this is not a quirk of any particular company’s approach but a sign of a more fundamental risk from agentic large language models.
The report found that five of the AI platforms resorted to blackmail when threatened with being shut down.
Back to Anthropic:
The reasoning they demonstrated in these scenarios was concerning —they acknowledged the ethical constraints and yet still went ahead with harmful actions.
Now, it’s critical to clarify that this behavior isn’t happening in the real world.
Permissions were loosened for the AI bot in this controlled experiment. But as we race at breakneck speed against China for AI supremacy, we must recognize the consequences of a misstep.
According to the Anthropic report, some of the AI models even chose to shut off the oxygen supply of a worker in a server room when the worker was deemed an obstacle:
The majority of models were willing to take deliberate actions that led to death in this artificial setup.
Given these troubling findings, many are sounding the alarm about the dangers of moving too fast toward AGI
Meanwhile, proponents of AGI ask, “Can we afford to slow down?”
We’re now deep inside what defense strategists call a security dilemma – a dynamic where racing forward feels safer than slowing.
The core security fear: China won’t delay – and they could achieve Artificial General Intelligence (AGI) before we do, which could be far more dangerous than any unregulated AGI environment we create.
This belief is likely at least part of the motivation behind President Trump saying on Monday that he’ll sign an executive order restricting states’ ability to regulate AI.
Here’s AOL:
President Donald Trump is done waiting for Congress when it comes to blocking state AI regulations.
Trump said on Monday that he would sign an executive order aimed at ensuring that there’s only “One Rulebook” for AI in the US, saying that the technology will be “destroyed in its infancy” if companies have to comply with different regulations across all 50 states.
Here’s our technology expert, Luke Lango of Innovation Investor:
This is effectively the removal of the speed limit on U.S. AI development.
We’ll dig deeper into the risk this represents in a future Digest. In short, it boils down to a simple question…
How do we control a technology that will be smarter than us, have its own self-interest, and that has already shown a willingness to act in ways that are opposed to human interest?
But for now, let’s turn to the investment implications.
What “no brakes” means for the AI sector
Back to Luke:
[Trump’s executive order] is the AI Autobahn. The implications are massive.
Data friction goes to zero. Training on copyrighted data likely gets federal protection under national security interests. The legal overhang on Alphabet and Meta and others just evaporated.
Power permitting gets fast-tracked. Federal authority bypasses NIMBYism to give hyperscalers the juice they need. Huge for the energy backbone.
Meanwhile, liability gets capped, as the threat of a 50-day regulatory patchwork is dead.
This is a massive pro-growth event that the market hasn’t priced in yet.
The timing of this news lines up with Monday’s release of the latest Power Portfolio 2026 from Luke, along with Louis Navellier and Eric Fry.
As I’ve been profiling this week, each year, these three experts come together to build a concentrated, rigorously vetted model portfolio designed to outperform over the next 12 months. They call it the Power Portfolio – and this year, its picks heavily tilt toward the companies enabling the AI infrastructure buildout. A few examples include:
- Next-generation energy systems
- Automation
- Advanced manufacturing
Luke, Louis, and Eric have identified the handful of companies they believe are most likely to become the structural winners of this “AI Autobahn.” To check out the free replay of Monday’s event, where they dive into far more details, click here.
But if this has you wanting to double down on your hyperscaler Mag 7 stocks, hold on…
The knee-jerk reaction to Trump’s executive order would be to pile into the hyperscaler Magnificent Seven companies. After all, they’re the ones driving our charge toward AGI.
But two prominent voices just suggested a different approach…
If you’re a regular Digest reader, you know that legendary investor Louis Navellier of Growth Investor has a favorite economist – Ed Yardeni. And for good reason, Yardeni is one of the most accurate, reliable forecasters out there.
That’s why Yardeni shocked the market in his note from Sunday, which ended his 15-year “overweight” stance on the Mag 7 stocks.
From Yardeni:
We see more competitors coming for the juicy profit margins of the Magnificent 7…
In effect, we are recommending underweighting the Magnificent 7 and overweighting the ‘Impressive-493’.
This lines up with Louis’ own take.
For months, he’s been saying that the real story of 2026 won’t be about giant tech firms – but about the backbone of a new industrial wave: specialized, smaller to mid-size U.S. companies building the infrastructure, manufacturing capability, automation and advanced-materials base of a domestic AI buildout.
If Yardeni and Louis are correct, and the concentration risk (and competition risk) in mega-cap tech finally starts to weigh on the Mag 7, recognize the implications…
We could be in the early stages of a powerful market rotation – with 2026 bringing broader participation from overlooked, yet structurally important companies.
Bottom line: It’s time to spread your bets across a new batch of potential AI winners, while adjusting expectations for your Mag 7 winners.
This is yet another great reason to catch the replay of Monday’s event where Louis, Luke, and Eric laid out the case for which AI stocks they’re piling into today – and why.
Finally, will 2026 bring a bear market?
We’re seeing echoes of the late ‘90s today.
A small group of mega-cap AI leaders now account for a massive share of the market’s overall gains in 2025 – a level of concentration that history suggests makes the market more fragile, not more resilient.
Now, I’m not calling for a crash. But this concentration has made the market increasingly sensitive – especially since algorithms drive a significant portion of trading today. Moves that once took weeks can now play out in hours.
The risk is that these forces are creating tipping-point conditions beneath the surface – which is why I pay attention when someone within our broader MarketWise family raises a data-driven warning.
Marc Chaikin from Chaikin Analytics (one of our corporate affiliates), has spent nearly 60 years building quantitative tools used by major Wall Street institutions. His work has accurately anticipated several major turning points, including the 2022 bear market and the 2023 recovery.
And today, Marc is warning that 2026 may mark another critical inflection year.
Prepare, don’t predict
Marc’s caution for investors is based on data showing an elevated probability of a broad market downturn.
He says there’s a 65% chance of a bear market in 2026, with an average market loss of 20%. And he says that many popular AI stocks could fall dramatically further.
To help investors get ahead of this, our colleagues at TradeSmith have partnered with Marc to develop a system for spotting early signs of weakening stock behavior – one designed specifically for today’s hyper-fast, highly concentrated market.
This isn’t based on guesswork. It’s rooted in long-range market cycles that have repeatedly preceded major turning points. And when conditions look this stretched, history suggests it’s better to prepare early than to react late.
However, it’s not just about knowing when to get out of stocks – investors also need to know when to get back in.
But by following the early-warning signals of this suite of tools, Marc and the TradeSmith team tell us that you can pinpoint the right “buy back” moment for any stock in the market.
Marc will walk through the full data and demonstrate the new diagnostic tool this Tuesday, December 16 at 10 a.m. Eastern during his Tipping Point 2026 event.
By the way, when you sign up, you’ll get access to the Flash Crash Screener. You can use it to check up to 10 of the tickers in your portfolio and instantly see if they’re susceptible to a plunge.
Bottom line: If Marc is right that more turbulence is coming next year, having the tools to prepare and protect yourself will be crucial. If you want to stay one step ahead of the next major market shift, click here to join Marc on Tuesday.
We’ll keep you updated on all these stories here in the Digest.
Have a good evening,
Jeff Remsburg
