By Stephen Butler
Discussions around an “AI bubble” are everywhere right now and not just in the financial press. The tech markets are flashing familiar signs: soaring valuations for companies like Nvidia, Meta, Google, and Microsoft; aggressive investment in AI infrastructure; and earnings inflated by generous depreciation schedules. People are questioning whether the excitement is justified or not.
Our industry is feeling this pressure acutely. At this year’s MBA Annual, Newrez’s Brian Woodring stated plainly, “I don’t think you could look at the industry and not think we’re in the middle of an AI bubble.”
But the real cause of the bubbles rarely gets the attention it deserves.
The attention is almost always around the loss of financial wealth and that narrative. What triggers the correction isn’t the hype itself; it’s the moment the market realizes that many companies have over-promised and under-delivered. Stock prices fuel the headlines, but the true root is the widening gap between promised value and value actually delivered to customers. This point gets lost in all the noise when the market goes through a bubble bursting crash.
In the 90s and early 2000s we saw a Telco bubble and a Dotcom bubble powered by inflated earnings, futuristic projections, questionable accounting, and vague use cases. Eventually, the correction came. The weaker firms disappeared, and the companies that survived were those that could prove their technology worked and delivered sustained customer value.
Industry analysts have started asking the same questions of mortgage AI vendors:
- Are they generating real operational lift or just demos?
- Are they improving defect rates, turn times, and cost-to-produce?
- Or are they leaning on slick UI and broad promises?
The Mortgage Scoop recently had some coverage on lenders becoming indifferent about flashy AI demos. Instead, they’re asking whether vendors can actually reduce turn times, improve manufacturing consistency, and deliver automation that works at scale. In other words, the industry is shifting from AI theater to AI that actually moves the economics of lending.
The truth is: the AI market is full of companies whose value hasn’t yet matched their valuation. Earnings and expectations are up, but real results are mixed and, in some cases, underwhelming. A meaningful correction in the next 12–24 months wouldn’t be surprising.
Yet, as in past cycles, not all AI companies are built the same.
Some vendors have been operating long before the hype cycle, with mature technologies, proven deployments, and real customer outcomes. They focused on results, not rhetoric. These are the companies that will weather the storm because their value is grounded in evidence, not expectation.
What separates the durable vendors from the fragile ones?
- Transparent, explainable systems
- Proven production results
- Consistent accuracy
- Real ROI experienced by lenders—not promised
- Strong data governance and auditability
- Clear understanding of how models behave in regulated environments
TRUE is one of those companies. For nearly a decade, our AI has powered real production workflows across top mortgage lenders. The results speak for themselves: meaningful reductions in manufacturing costs, faster cycle times, and higher levels of automation that eliminate hours of manual review. In adjacent verticals, partners have standardized on TRUE because automation that once took hours now happens in minutes.
There may be an AI bubble forming. Some people in our industry certainly think so. But bubbles don’t kill good technology. They expose the weak players and elevate the strong ones.
The AI systems that already prove their value, reduce cost and risk, and that support the deepest operational layers of mortgage manufacturing—those will survive the hype wave. And they will power the next decade of lending innovation long after today’s excitement fades.
Join the Conversation!
Subscribe to our biweekly newsletter for a deep dive into where AI technology is going for mortgage lenders, specific use cases, and discuss a “smarter, not harder” approach to innovation.
