It seems like the “AI will replace you on your job” scare is slowing down. However, this is not because we have proven that humans are better than artificial intelligence (AI), but because companies are avoiding taking on a bigger challenge to implement it.
In a recent study published by the Massachusetts Institute of Technology (MIT) titled “The GenAI Divide: State of AI in Business 2025,” they found that almost 95% of companies and organizations that invested billions into Generative AI (GenAI) are getting no results. Only 5% succeed in getting their money’s worth and extracting millions of dollars in value.
As MIT calls it, the “GenAI Divide,” where they conducted interviews, surveys, and analysis of 300 public implementations, they found these patterns:
- Only 5% of custom enterprise AI tools ever reach production.
- Only two out of eight major sectors (tech and media) exhibit meaningful structural disruption.
- Big firms lead in pilot volume but lag during scaling up.
- Investment money favors sales and marketing pilots despite low impact.
- Having external partnerships almost doubles the success rate.
This is an interesting finding, as an estimated $30 to $40 billion in enterprise investment went into GenAI, only to produce almost “no measurable P&L impact.”
According to the MIT study, the main obstacle to scaling is not infrastructure, regulation, or talent. It is learning. Most GenAI systems do not retain feedback, adapt to context, or improve over time. The small group that is advancing in the GenAI field faces these limitations until they can build something that can continuously learn.