E-Paper
AI can’t make companies rich: MIT study finds 95% of firms fail to profit from their investments
AI is draining business budgets fast, not delivering promised results. Most companies are struggling to see value; few are turning projects into growth.
Artificial intelligence was supposed to be the technology that launched businesses into a new era of efficiency, streamlined operations, and untapped revenues. But as the dust settles and the bills roll in, a new MIT study is showing just how far reality has lagged behind the dream. Across the United States, companies have poured an estimated $35–40 billion into AI projects in just the last year. For most, though, the only visible outcome so far is an alarming drain on resources.
The expectation was straightforward: invest big in AI, reap even bigger returns. Instead, nearly all businesses are seeing little, if any, measurable results. According to the MIT report, “The GenAI Divide: State of AI in Business 2025,” 95 percent of firms surveyed have not managed to move the needle on profits or tangible business improvements thanks to their AI investments. The notion of AI as a gateway to future growth is fast turning into a lesson in patience, disappointment, and financial risk.
Where the hype fell flat
So, what’s going wrong? In short, most businesses jumped on the AI train without a map or destination. The study found that the majority of companies wasted time and cash applying AI to sales and marketing, betting these consumer-facing applications would bring instant wins. In reality, these areas still require a human touch and aren’t where automation shines brightest. Instead, the potential of AI remains locked up in back-office and repetitive tasks - the sorts of everyday admin duties that are less glamorous but far more suitable for automation. Companies that ignored these areas simply never got the payback they’d hoped for.
Add to that the mismatch between generic, off-the-shelf AI tools and the unique workflows at each company. Many relied on big-name platforms like ChatGPT, but instead of integrating seamlessly, these tools often feel awkward, stalling out and having minimal impact in real business contexts. Custom problems need custom solutions, something many forgot in their race to invest.
Warning for the workforce
Yet a handful of businesses have found the secret sauce. The study highlights young startups and nimble teams that took a different tack: target one business pain point, solve it with precision, and partner with third-party vendors who know the tools inside and out. For them, the results are silver-lining material, some saw their revenues climb from zero to $20 million in just a year by staying sharply focused and realistic. Notably, two-thirds of successful AI deployments came from third-party vendors, compared to only a third built in-house.
On the job front, layoffs due to AI aren’t happening yet, but companies are filling fewer support and admin roles as technology steps in. How long this quiet approach can last before real disruption remains a hot question. Experts warn that when (or if) AI masters true contextual work and autonomy, the workforce may not be so safe.
