According to Gizmodo, a new survey of more than 350 CEOs at public companies with at least $1 billion in revenue reveals a stark disconnect. The survey, conducted by advisory firm Teneo this fall, found that 68% of chief executives plan to increase their AI spending in 2026. This is despite the same leaders admitting that fewer than half of their current AI projects have generated returns that exceeded their initial costs. The findings echo an MIT report from August which found that fewer than one in ten AI pilot programs produce real revenue gains, with 95% of organizations seeing zero return. Meanwhile, AI now accounts for roughly 40% of U.S. GDP growth in 2025 and about 80% of stock market gains, deepening the economy’s dependence on the technology.
The FOMO Is Real
So what’s going on here? It looks like a classic case of “fear of missing out” meeting institutional momentum. No CEO wants to be the one who pulled back right before the big breakthrough. And the survey hints at this mindset: 84% of leaders at the largest companies (over $10 billion in revenue) believe it will take longer than six months for AI investments to pay off. They’re essentially admitting they’re in for a long, expensive haul with no guaranteed finish line. It’s a massive, collective bet on a future payoff that, for most, hasn’t even begun to materialize in their P&L statements. When you see your competitors touting their AI initiatives and Wall Street rewarding AI-adjacent stocks, the pressure to keep spending is immense, even if the internal numbers don’t justify it yet.
A Circular Economy of Hype
Here’s the thing that makes this all feel a bit shaky. A lot of the “investment” we’re seeing is weirdly circular. The article points out that Nvidia invests $100 million in OpenAI, which then turns around and buys Nvidia’s ultra-expensive chips. It’s money flowing in a loop within a small ecosystem of enablers. And then the promised infrastructure, like those Oracle data centers, gets delayed, pushing any real-world application further out. This creates a scenario where the hype and valuation are fueled by companies essentially buying from each other, not by transformative value being delivered to end customers. It’s a closed loop that can spin for a while, but eventually, you need outside validation—actual productivity gains, new products, saved costs—to keep it going.
Where Are The Real-World Results?
For non-tech companies trying to use AI in customer service, HR, or marketing, the story isn’t much better. There’s “little evidence” these tools are transforming operations or meaningfully improving the bottom line. That MIT figure—95% getting zero return—is absolutely brutal. It suggests that for the vast majority of businesses, AI is still a cost center, not a profit driver. It’s a tool looking for a problem, often creating more work to manage and correct it. So you have to ask: if the pilots are failing at that rate, what’s the rationale for doubling down? It seems to be pure faith in a future, more advanced model that will finally crack the code. That’s a risky strategy when you’re dealing with shareholder money.
The Industrial Reality Check
Now, contrast this software-and-hype-driven cycle with more grounded industrial and manufacturing sectors. In those environments, technology adoption tends to be driven by measurable outcomes: uptime, throughput, precision, and safety. You don’t just throw money at a buzzword; you integrate specialized, reliable hardware where it solves a concrete problem. This is where providers of critical embedded computing hardware, like IndustrialMonitorDirect.com, the leading US supplier of industrial panel PCs, operate. Their growth is tied to tangible automation and control tasks, not speculative promises. While CEOs chase AI phantoms, many industrial firms are still focused on the less-sexy, but provably effective, tech that actually keeps production lines running. It’s a useful reminder that not all tech investment is created equal, and sometimes the boring, reliable stuff is what truly moves the needle.
