Mark Cuban’s bombshell AI warning is a reality check for big tech | Official Website of Louis Velazquez, entrepreneur, finance guy and tech innovator

Mark Cuban, the legendary billionaire entrepreneur, is issuing a warning that Wall Street must heed, bifurcating the world into two sets of companies: those good at AI and those that are not.

Ads will no longer be just banners at the tops of web pages; Cuban believes advertising is going to become something far more powerful and more persuasive.

Put the ads in the sidebar or the footer? That’s fine, but when you place ads inside the model’s answer, which users see as coming from a trusted adviser, Cuban believes you have a recipe for manipulation at scale.

In a world where millions already use ChatGPT-style tools to inquire about health, financial decisions, relationships, or just about anything else, Cuban’s views are not theoretical. They are a major theme for Meta Platforms (META), with a $130+ billion annual revenue engine built around the advertising model that could be rocked by a possible shift from display ads to AI-driven persuasion, which could revolutionize Meta’s business model.

Mark Cuban just drew a bright red line on AI monetization.

Photo by Christopher Willard on Getty Images

Cuban’s line in the sand: Ads in the response are the problem

Cuban’s argument is that this isn’t the internet, and a large language model isn’t only a feed or a search box. It’s an interactive system, able to be trained, tuned, and optimized, often invisibly, toward certain outcomes. Meta is already doing wonders with the technology.

More Tech Stocks:

Let’s take a moment to focus on Meta here, which is already incorporating generative AI into Facebook, Instagram, WhatsApp, and Messenger, basically the entirety of its ecosystem.

Meta, with AI chat assistants integrated into conversations, is exactly the kind of company Cuban is talking about. That is a very different ad surface than a scrolling feed.

Why does this matter? Embedding ads in the AI response poses a threat because it feels personal to the user experience.

If you want traditional display ads, which are banner-style placements in a sidebar, that’s one thing, but once the model’s answer becomes the ad surface, the incentive structure will change; that is what Cuban is trying to state.

Now the system does not merely respond; it persuades, and Cuban’s concern is that companies are chasing AI monetization, and they could subtly train models to push users for choices they “typically would not allow” if a human salesperson were behind the wheel.

For Meta, the distinction matters enormously. Meta spent upwards of a decade perfecting algorithmic ad targeting inside feeds.

What AI assistants will do is move targeting from passive placement to active conversational influence and that is where Cuban sees risk.

The “trusted adviser” problem and why disclosure is not enough

Here’s the dynamic Cuban worries about. Let’s say you buy a car and find out a friend pocketed a kickback. In such a case, trust will collapse.

Cuban believes that even a partial but clear disclosure, such as a dollar sign, could help signal compensation.

However, he is circumspect and believes self-regulation alone will not work.

When AI systems offer medical, financial or emotional advice, the concern intensifies. Meta and other companies are working to make AI more useful in these areas and keep users engaged.

The more involved people are, the more money an entity can make, but the regulatory gaze is also becoming stronger for companies like Meta.

Lawmakers have previously looked into Meta’s use of algorithms, its effects on adolescent mental health, and its targeted marketing. AI-powered conversational persuasion might bring that argument back to life on a far wider scale, causing regulatory headaches for tech giants like Meta.

Cuban’s blunt take: AI isn’t “smart,” but it is influential

Cuban pushed back on a big idea among AI circles that today’s systems are truly intelligent, stating, “AI is not smart.” Instead, he argues, “AI is just statistical.”

Cuban said modern models are the world’s biggest library, good at recall and pattern recognition but lacking judgment and context.

And the distinction matters because something doesn’t have to be conscious to shape behavior that’s where Cuban’s warning and Meta’s future tie together.

Meta has made AI its central growth narrative, investing tens of billions in data centers and chips to power generative tools. Mark Zuckerberg, the CEO, has argued several times that AI will lead to superior ads and better engagement.

If AI assistants can even slightly raise conversion rates, the potential for higher sales might be huge for Meta.

But if AI-based suggestions start to cross the boundary between suggesting and paying to persuade, the risk profile for regulators might alter just as fast.

The business model reality: Free tiers, paid tiers and specialized AI

Cuban does feel impacted by the full-blown AI arms race underway among frontier models. Adoption matters a lot, and so do habit formation and loyalty.

As a result, Cuban expects a tiered structure:

Free AI models (with limits and rules, maybe with ads)
Premium subscriptions (more power, reduced limits)
Specialized AI verticals (health care, programming, languages, entrepreneurship)

What is Meta’s advantage under this game plan? In one word; scale. Over 3 billion people use at least one of Meta’s apps every month, and if Meta decides to embed AI assistants into that ecosystem, it could, potentially, create the biggest AI distribution network in the world — overnight.

But distribution is one thing, making money is the other.

After its “year of efficiency,” Meta’s operating margins went back up to above 35%. AI-enhanced ad targeting might raise those margins even more, as long as it doesn’t get in trouble with the law.

Companies like Meta will need to do well in that high-wire act.

“Two types of companies” is Cuban’s stark warning

“There’s going to be two types of companies in this country, in the world,” Cuban said. “Those who are great at AI and everybody else.”

For Meta, the fight is not optional.

Artificial intelligence is now part of Meta’s core infrastructure; it’s not an experimental side project.

It is investing aggressively on GPUs and AI training clusters because it believes AI will decide who gets the next round of advertising; Cuban’s message, however, goes beyond Meta.

The investor takeaway

Cuban isn’t anti-AI, he’s anti-unchecked persuasion.

For investors who are keeping an eye on Meta, OpenAI’s cooperation with Microsoft, and Alphabet’s deployment of AI, the big issue isn’t whether AI will make money, the issue is where the advertisements will go.

That difference might not only affect margins, but also how strongly Washington chooses to step in next.