I've been in technology long enough to recognize what happens when excitement gets ahead of economics. I lived through the dot com boom, watched companies raise extraordinary amounts of capital on the promise of what the internet would become, and then watched many of those same companies disappear when the market demanded something they didn't have. Revenue, sustainable economics and a business that could actually support its valuation.
The internet wasn't the problem. The internet changed everything. The problem was that too many companies confused being part of a technological revolution with having a sustainable business.
I see some of those same characteristics developing in artificial intelligence today.
Don't misunderstand me. I believe AI represents one of the most significant technological advances of my lifetime. I've spent years building it into our own platforms because I understand what it can accomplish. But believing in the technology doesn't mean you have to believe every company, every valuation or every infrastructure investment supporting it.
And it certainly doesn't mean you should build your entire business around one AI model.
We've Been Managing Concentration Risk for Decades
In factoring and specialty finance, concentration risk is something we understand instinctively.
You don't want one debtor representing 80 percent of your receivables. You don't want one client responsible for your entire portfolio. You establish limits because you understand that even a strong relationship can become a liability when too much of your business depends on it.
Now consider what happens when a software company builds its entire AI strategy around one model from one provider.
What happens when that provider changes its pricing? What happens when the model is discontinued, its performance deteriorates, or its infrastructure experiences an outage? What happens when a competitor introduces something materially better?
If your entire intelligence architecture depends on that one model, you've created the technological equivalent of a portfolio with one debtor.
That's not a strategy I would accept in finance, and it's certainly not one I'm willing to accept in technology.
The GPU Gold Rush Looks Familiar
Look at the capital flowing into AI infrastructure.
Technology companies and data center operators are committing enormous amounts of money to GPUs, power, facilities and computing capacity. Some of those investments are justified by existing demand. Others depend heavily on demand that hasn't materialized yet.
That distinction matters.
A GPU isn't a piece of real estate that necessarily appreciates over time. It's technology. And technology has a habit of becoming obsolete faster than the financial models used to justify purchasing it.
Think about what happens when a company finances thousands of GPUs based on projected demand three or four years into the future. By the time that demand arrives, newer processors may offer substantially better performance, lower operating costs and greater efficiency.
The company still has the financing obligation. It still has the depreciation. It still has the infrastructure costs. But the market has moved on.
I've seen this movie before.
During the dot com era, companies invested heavily in infrastructure because everyone believed demand would eventually catch up. In many cases, it did. The internet grew beyond what most people imagined. That didn't save every company that financed the infrastructure.
You can be absolutely right about the future of a technology and completely wrong about the economics of owning it.
I'm not predicting that every AI data center will fail or that GPU demand will disappear. I'm saying that the amount of capital being committed, the pace of hardware advancement and the uncertainty around future utilization create risks that shouldn't be ignored.
Some providers will become enormously successful. Others will consolidate, restructure or disappear. That's what competitive markets do.
The mistake is assuming that today's leader will automatically be tomorrow's winner.
Why We Built FactorFox and FactorEvo to Be Model Agnostic
When we designed the intelligence architecture behind FactorFox and FactorEvo, we made a deliberate decision.
We would not build our future around one AI company, one model or one generation of technology.
We would use the best model for the job.
Today, that might be Claude. Tomorrow, it might be a model from OpenAI, Google, Microsoft or another provider that hasn't even entered the market yet.
I don't care which company wins the latest benchmark or generates the most headlines. I care about which model produces the most accurate, reliable and economically sensible result for the work we're asking it to perform.
Reading a commercial finance agreement requires different capabilities than analyzing a debtor's payment history. Extracting information from a bill of lading is different from evaluating a borrowing base. Generating a financial report is different from identifying an exception in a funding transaction.
Why would I assume one model is the best at all of those things?
I wouldn't.
That's why our architecture is model agnostic. We can evaluate different models for different responsibilities rather than forcing every operation through the same provider.
The intelligence belongs to our platforms. The models are components we use to deliver it.
That's an important distinction.
We're not building a business that exists because one AI company allows us to access its technology. We're building financial operating systems that can take advantage of whichever AI technology delivers the best results.
And when the technology changes, we intend to change with it.
The Real Asset Isn't the Model
There's another part of this conversation that I think the industry is overlooking.
AI models are becoming increasingly capable, and competition is accelerating. What appears extraordinary today may become a standard capability tomorrow.
If your entire competitive advantage is access to a particular model, what happens when everyone else has access to the same thing?
Where's your advantage?
At FactorFox, we've spent more than two decades working with factoring and specialty finance companies. We understand receivables, funding, collections, debtor behavior, reserves, credit policies, borrowing bases and the operational decisions that determine whether a finance company makes money or loses it.
That institutional understanding is what we've built into our platforms.
The model helps interpret information, reason through a problem and communicate an answer. But the underlying transactions, operating policies, permissions, financial controls and supporting evidence remain part of the financial operating system.
That's where the value lives.
A model can tell you that a debtor represents a concentration risk. Your operating system needs to know the debtor's actual exposure, the applicable limits, the supporting invoices, the client's agreements and whether a transaction should proceed.
Those are two very different responsibilities.
We've designed FactorFox and FactorEvo around that distinction.
Intelligence Should Be Accountable
I also believe we're going to see a major shift in how financial institutions evaluate AI.
Right now, much of the conversation is about what AI can generate. How quickly can it answer a question? How well can it summarize a document? Can it produce a report in seconds?
Those are useful capabilities, but they're only the beginning.
In specialty finance, an answer without supporting evidence has limited value.
If AI tells a factor that a debtor's payment behavior has deteriorated, I want to know what transactions support that conclusion. If it identifies a covenant exception, I want to see the agreement and the provision involved. If it recommends holding a purchase, I want to understand why.
And I want that information inside the operating system, subject to the same permissions and controls governing the rest of the business.
That's what we're building.
AI shouldn't become another disconnected application that employees have to manage, reconcile and secure. It should operate within the business, understand the business and remain accountable to the business.
Being model agnostic makes that architecture more durable. It allows the underlying intelligence to evolve without surrendering the operating system to any single model provider.
Ask Your Software Provider One Simple Question
If you're evaluating an AI platform, ask the provider what happens when its preferred model is no longer the best available option.
Can it switch to another provider? Can it select different models for different tasks? Can it accommodate changes in cost, availability and performance without forcing customers to rebuild their workflows?
And most importantly, does the software company own and control its intelligence architecture, or is its entire AI offering dependent on someone else's product?
Those questions matter far more than whether a vendor has integrated the latest model receiving attention on social media.
I've watched technology companies come and go for more than three decades. I've seen dominant platforms lose their advantage, promising companies disappear and technologies that once seemed indispensable become irrelevant.
The lesson I've taken from all of it is simple.
Never confuse a technology provider with your technology strategy.
First to the Cloud. Now AI Native.
We brought FactorFox to the cloud in 2002, when much of the industry was still running traditional installed software. We didn't make that decision because cloud computing was fashionable. We made it because we believed it was a better way to deliver and maintain financial software.
In 2026, we're bringing that same thinking to AI Native financial operating systems through FactorFox and FactorEvo.
But this time, we're making another important decision. We're not tying our future, or our customers' future, to the fortunes of one AI model.
I believe artificial intelligence will continue advancing at an extraordinary pace. I also believe some of today's most celebrated models and infrastructure providers won't be the leaders five years from now.
That's not a criticism of those companies. That's the nature of technology.
Our responsibility is to make sure our customers benefit from that competition rather than become dependent on its outcome.
We don't need to predict which AI model will win. We need to make sure our customers win regardless.
That's why FactorFox and FactorEvo are model agnostic.
And that's not going to change.
Roberto Vasquez
Founder & CEO
FactorFox | FactorEvo
Experience FactorFox AI with your own operating data, agreements and reports. See what an AI Native financial operating system can do for your business.