From Clients to Providers, AI Struggles to Prove Its Worth

95% of Firms Adopting AI Report Little to No Results
Even OpenAI Expands Its Business to Boost Profitability
What Strategies Are Global Tech Giants Using to Monetize AI?

A recent study from the Massachusetts Institute of Technology (MIT) revealed that 95% of companies running pilot programs with generative AI saw little to no improvement in profitability. In most cases, the adoption of AI models did not translate into measurable business growth. The findings come as AI firms themselves, including OpenAI, continue to struggle with monetization—highlighting how both providers and their clients are finding it difficult to turn AI into sustainable profits.

AI Brings Little Improvement to Corporate Performance

According to reports from major outlets on August 22, MIT released a study on August 18 titled “The GenAI Divide: State of AI in Business 2025.” The report, compiled by the MIT NANDA (Networked Agents and Decentralized AI) Initiative, drew on 52 interviews with corporate leaders, analysis of more than 300 AI plans and announcements, and a survey of 153 business executives.

The findings show that companies are pouring between $30 billion and $40 billion into generative AI adoption. More than 80% of firms have tested or experimented with tools like OpenAI’s ChatGPT and Microsoft’s Copilot, while about 40% are actively deploying them.

Yet, 95% of companies using generative AI reported no meaningful improvement in profitability. Only about 5% of early adopters generated millions of dollars in measurable value, while the vast majority saw little to no bottom-line impact. The report noted, “These tools are primarily enhancing individual productivity rather than boosting overall corporate performance,” adding that fragile workflows, lack of contextual awareness, and poor integration into day-to-day operations have left most AI initiatives falling short of delivering real financial gains.

Even OpenAI Struggles to Secure Profitability

The profitability challenge extends to the providers of generative AI services themselves. OpenAI, the industry’s frontrunner, doubled its revenue in 2024 to $3.7 billion, yet still posted a $5 billion loss over the same period. This year, its annual recurring revenue (ARR) is projected to surpass $20 billion, but analysts say the company will remain mired in losses.

To close the gap, OpenAI is exploring new business models. In a recent interview reported by Bloomberg on August 21, CFO Sarah Friar noted, “We’ve built significant expertise in designing and constructing data centers optimized for AI workloads. While we’re currently focused on meeting our own computing needs, we are also looking at opportunities to commercialize this capability in the long term.” The idea mirrors the trajectory of Amazon Web Services (AWS), which grew by leasing out surplus cloud computing capacity.

OpenAI’s urgency is driven by massive upcoming investments in AI infrastructure. On August 15, CEO Sam Altman told reporters, “We’re going to spend trillions of dollars building data centers in the not-so-distant future. Economists may say it’s crazy or reckless, but our answer will be: ‘We’ll take care of it.’” If realized, such a plan would reduce OpenAI’s reliance on partners like Microsoft and Oracle, which currently support its infrastructure. But without a sustainable revenue model, the company risks undermining its long-term viability.

AI Firms’ Survival Strategies

Global tech giants are also intensifying efforts to monetize AI, with premium subscription tiers emerging as a leading tactic. Google has rolled out “Google AI Ultra” at $249.99 per month for its Gemini model, while Elon Musk’s X offers “SuperGrok4 Heavy” at $300 per month. With many AI companies now unveiling more advanced agent features, even higher-priced tiers are expected to follow.

Another strategy is weaving AI more tightly into advertising and commerce. Google has begun integrating ads into its Gemini-powered AI Overview search results. OpenAI earlier this year added a “shopping” feature to ChatGPT to showcase products and reviews, and is now developing a checkout system that would allow users to search and pay without leaving the platform. Merchants processing orders through the system would pay fees to OpenAI, effectively turning ChatGPT into a closed-loop commerce hub.

Meta is likewise leveraging generative AI to raise ad prices and improve efficiency, boosting revenue that is then funneled back into further AI investment. CEO Mark Zuckerberg underscored this shift, saying, “Now is the time to invest aggressively in AI for future growth. AI-driven advertising is already delivering strong results.”