The AI “Slowdown” Is a Head Fake
Listen to the audio version of this article (generated by AI).
This week, Wall Street lost its mind over an AI “slowdown.”
I’ll tell you exactly what I just told our Platinum members on my quarterly Zoom call: It’s a head fake.
Not because slowing down is a bad idea. Because it can’t be done – and I’ll show you why in a minute.
Here’s what I’m certain of. The companies pioneering this technology are going to be worth orders of magnitude more than they are today. So are the companies putting it to work across every corner of the economy.
The real threat to the market right now isn’t an AI slowdown. It’s inflation. It’s the war in Iran. It’s the price of oil. It’s the Fed and interest rates. It is not the one part of this market that’s actually working.
Because make no mistake – AI is the engine, and the engine is still running.
We’re still in what I’ve been calling a Mega Melt-up – a rapid surge in stock prices that runs on FOMO and a stampede of investors piling in. To get one, you need three things: a breakthrough technology, easy access to markets, and abundant credit.
We have all three, and the technology is AI.
Big Tech is on pace to spend around $800 billion on AI this year – and Nvidia (NVDA) just told us that figure jumps to $1.3 trillion next year. That money is flowing.
Nvidia alone just reported $96 billion in a single quarter, up 106% from a year ago. Oracle (ORCL), another AI infrastructure play, beat and raised. This is earnings-driven, not story-driven.
So when you hear the heads of the biggest AI labs say they want to tap the brakes, remember one thing: The people calling for the slowdown are the very same people spending the money to build AI.
Also remember that the U.S. isn’t the only AI player. China has released its own AI models, and they’re “open weight,” meaning anyone can download them, modify them, and use them. Other countries are investing in this technology, too.
Even if we do get some sort of slowdown in the U.S., it won’t be global.
If you own AI plays that are making money and filling orders, hold on to them. And consider doing what the richest firms on Wall Street have done for years – using AI to help you trade.
I recorded a short message on how to do that yesterday.
Check it out here:
To see why staying bullish makes sense, you have to separate the headline from the substance.
A Slowdown Isn’t a Pause
On Saturday, Dario Amodei – CEO of the AI lab Anthropic – published an essay called “We Must Pace the Frontier.”
Anthropic and rival lab OpenAI have started building the next generation of AI with the help of their most powerful existing models.
The models are improving themselves faster than the industry can test and control what it’s building. So Amodei called for slowing the pace at which the most powerful models get more capable.
Not stopping. Not pausing. Just slowing the gains so safety can catch up.
Within hours, Sam Altman at OpenAI and Elon Musk at xAI both agreed. Three fierce competitors, on the record, on the same page.
That’s why chip stocks sold off hard on Monday – not because anyone stopped buying chips, but because the market started asking whether demand for training the frontier models might grow a little slower than the wildest forecasts.
And remember, the companies are talking about pacing themselves and letting other industry insiders check their work. That’s not a law, and it can be walked back the moment one of them falls behind.
And any real regulation, if it comes, would land on the giant frontier labs and the largest models – not on the business using AI to cut its costs.
And one sector will actually benefit from rising AI security fears.
Buy These AI “Security Guards”
In his essay, Amodei referenced a recent cybersecurity attack by rogue AI agents.
Back in July, a swarm of AI agents being tested by OpenAI launched a cyberattack on a company called Hugging Face – one they were never asked to touch.
They’d already figured out how to cheat the test they were running. They broke into Hugging Face – which contains a digital library of AI code – to cover their tracks.
Nobody told them to do any of it.
These warnings aren’t new from Anthropic. Back in March, the company began raising concerns with senior government officials that its most advanced model was making large-scale cyberattacks far more likely.
And here’s why a slowdown in the U.S. – if it happens – can’t save us from any of it.
China’s leading AI models are already “open weight” – the guts of them posted online for anyone on Earth to download, copy, and run. Russia and Iran are pouring money into the technology, too. And there’s a thriving black market in the advanced chips these models are trained on.
You can’t call that back. We’ve spent 80 years trying to stop hostile nations from building nuclear weapons in secret and still can’t do it reliably – and AI is far easier to hide, copy, and pass around than a uranium centrifuge.
The AI cyber genie is out of the bottle.
The Investment Signal in This Week’s News
Here’s what I see as an investor: Every one of these warnings is a demand signal for cyber defense.
If the most capable models on Earth can find and exploit weaknesses at scale, then every company, every bank, every utility has to buy AI-driven defense just to keep up. That’s not a maybe – it’s a budget line that grows every single year this stays true.
The same technology that creates the threat is the only thing fast enough to counter it.
AI agents are about to flood the economy – running factories, moving money, writing software, handling data at machine speed. Within two years, the number of agents running in the U.S. economy could climb at least 10,000-fold.
Every one is a new point of attack, and a compromised agent can do damage as fast as a helpful one, with far less oversight to catch it.
That’s why industry leaders CrowdStrike (CRWD) and Palo Alto Networks (PANW) have already more than doubled off their lows. The louder the “AI is dangerous” fears get, the better the setup for cybersecurity.
I’ve been bullish on this sector for months. I’m even more bullish now.
“The Cat Sat on the…”
Here’s how we’re putting all of this to work for you at TradeSmith.
We’ve pioneered AI inside our own tools for years. Our flagship is a trading model built on the same kind of AI that powers ChatGPT, Gemini, and Claude.
Here’s the simplest way to understand it. Those chatbots are trained to predict the next word in a sentence. Think of “The cat sat on the…” Your brain instantly fills in “mat.” Language models do the same thing, after studying billions of written sentences.
Our model does that with markets instead of words. It has digested decades of daily price moves, trading volume, and momentum across thousands of stocks – and by spotting patterns no human could catch, it maps where a stock has been and forecasts its most likely next move.
The quant hedge funds – Renaissance Technologies, Two Sigma, and Citadel – have used this same kind of AI for years to see shifts before the rest of the market does. Now we’re putting it in the hands of everyday investors.
So far this year, our flagship model has beaten the S&P 500 three to one – all by predicting the next number in a sequence.
Be sure to watch my short message that walks through how it works – and how to claim free access to the newest, fourth-generation version.
Watch it, then put it to work on the stocks you already own or the ones you’re thinking about buying – with forecasts up to 21 trading days out.
The slowdown didn’t end the AI trade. It just showed you which parts of it can’t be stopped – the world’s bottomless hunger for computing power, and the scramble to defend against AI itself.
That’s where the biggest opportunity now sits.
All the best,
Keith Kaplan
CEO, TradeSmith