— Opinion column by manufacturing expert Buckley Brinkman
AI has been on an incredible roll — and we’ve all benefited. The technology clears mundane tasks from our desks, creates opportunities for our manufacturers, and adds one to two percentage points to our economic growth. Investors have backed these ventures with hundreds of billions of dollars. It’s a terrific time to be engaged with AI.
Three recent data points suggest that this roll is hitting a wall. The leading AI companies appear to be reaching the limits of their ability to raise capital and grow demand as fast as they’d expected. None of them expect to turn a profit for years, so they depend on outside investment just to keep operating. At the same time, they need healthy margins to improve cash flow — and cheaper Chinese models are grabbing market share and squeezing prices.
OpenAI’s Financial Limits
OpenAI introduced millions of us to accessible AI. Its impact is undeniable: the fastest product ever to reach a million users, unprecedented leaps in capability, and a firm grip on the world’s attention. OpenAI has been a terrific innovator and a catalyst for change.
Still, OpenAI is not a profitable business. Recent financial statements show the company on pace to lose $14 billion on revenue of about $25 billion this year. The real picture is worse: massive capital expenditures push negative cash flow to roughly $27 billion. Analysts expect the cumulative cash drain to reach $115 billion before OpenAI turns a profit in 2029 — and hitting that target requires revenue to climb to $125 billion.
These financial pressures cast a new light on OpenAI’s recent proposal to give the U.S. government a 5% stake in the company, ostensibly to “share the upside” with the American people. Call me skeptical, but I always look for the catch when someone offers a $43 billion gift. Two issues come to mind. First, OpenAI — and its rivals — would benefit from friendly regulation: federal agencies could restrict competition or clear away obstacles to their market activity. Second, private investors may be nearing their limits on funding new capital, and government ownership would give federal
agencies a built-in incentive to keep the money flowing. Either move would change the competitive game.
Meta Can’t Create AI Demand
Meta made news this week too, announcing a new business that leases AI capacity to other firms — capacity built with more than $180 billion committed to AI infrastructure. The change matters because Meta’s own justification for that spending shifted. The company first framed the investment as necessary to keep up with its own AI demand, then as capacity that would meet internal needs. Now, Meta admits it can’t use all of the created capacity. That’s a telling admission: one of the best-funded players in AI just conceded it can’t generate enough organic demand to fill the infrastructure it created.
China’s Cheaper Models Invade
China keeps developing its own AI models, including DeepSeek, Alibaba’s Qwen, and Moonshot’s Kimi. These models take a different approach to large language models — one that works around technical constraints and plays to China’s strengths, particularly cheap energy. The results rival Western incumbents at a fraction of the cost. Chinese models jumped from 1.2% market share to nearly 30% in a single year. Individual model shares will keep swinging, but the trend is unmistakable: this competition caps pricing power at exactly the moment U.S. companies need margins and cash flow most.
Why Should I Care?
AI’s role in our lives and businesses keeps growing, and demand will keep climbing as we find new uses for these tools. That growth demands billions in cash — to run the service and to build better models — and none of the key players expect to turn cash-flow positive anytime soon. By their
own estimates, they need a fivefold jump in revenue to get there. If investors balk at supplying that cash, these companies will have to slow down or scramble for new capital.
These business models depend on substantial growth — yet Meta just admitted it can’t generate enough organic growth to use the capacity it built. If growth slows further, the cash math gets ugly fast.
Meanwhile, Chinese competition is squeezing both expansion and margins at precisely the moment growth and revenue matter most — making success harder for the domestic incumbents.
AI companies need years of strong growth to deliver returns and justify the infrastructure investments driving that growth. If these three data points mark the start of a trend, they would disrupt that growth and jeopardize future investment. Capital would slow or get more expensive, stalling the business model. That stall would shrink the biggest engine of economic growth and stock market value creation we have — and drag the broader economy down with it.
I’m not an economist – or a technology forecaster – and this isn’t a prediction; but OpenAI’s cash crunch, Meta’s admission it can’t create enough demand, and China’s pricing pressure are three signals worth watching. Together, they mark the line between AI’s next chapter of growth and the end of its roll. Keep an eye on all three — the companies fueling today’s economic momentum may be the first to tell us when it’s over.
