The Great Fintech Séance

There is an ancient alchemy at play in modern markets. It is the art of the Multiple Expansion.

If you tell Wall Street that you are a bank, a credit card processor, or a consumer lender, they will yawn, hand you a 10x price-to-earnings multiple, and tell you to run along and worry about capital requirements. But if you perform a digital seance—if you spray-paint "proprietary AI platform" onto your plumbing and claim you are programming consciousness rather than just managing ledger balances—the market will suddenly throw a party and hand you a tech valuation.

Currently, four massive financial platforms find themselves in this exact hot seat. They are companies desperately trying to escape their true natures. None of them want to be valued for what they actually do. They want you to believe they've built a digital oracle, when in reality, they've just built a set of highly polished trapdoors designed to optimize their own transaction volume.

1. Upstart: The Black Box That Forgot How to Count

We've officially entered the era of Faith-Based Credit. Once upon a time, if you wanted a loan, a stern person in a suit looked at your FICO score. It was a dry, objective, and incredibly boring system. You paid your bills, the number went up; you didn't, it went down.

Then came Upstart (UPST), the company that decided the FICO score was too "legacy." Why rely on boring old payment history when you can use "proprietary AI" and 1,600 non-conventional variables to analyze a borrower's potential? It is astrology for your wallet, powered by a GPU.

The problem with an AI credit model is that it is only as smart as the era it was trained in. Upstart's "Oracle" spent its formative years in a world of 0% interest rates and government stimulus. It learned to hunt in a zoo where the prey was tied to a post. Now that interest rates are real and inflation is persistent, the algorithm is suddenly discovering that "potential" doesn't actually pay the rent.

Whenever the default rates tick up, the corporate defense is always a beautiful, poetic phrase: "The model is recalibrating." It implies that the AI isn't wrong; it's just "learning" that poor people can't pay back loans during a macro contraction. We used to call that bad underwriting. Now, we call it machine learning.

If a human loan officer makes a million bad calls, they are fired. If an AI does it, it's just "gathering data for the next iteration."

2. SoFi: The Neobank with a Soho House Complex

If Upstart is the "Oracle," SoFi (SOFI) is the "Socialite." They started by refinancing Ivy League student loans—essentially betting on the elite—and have since tried to rebrand as the "AWS of Fintech."

They don't have customers; they have "members." It is a bank that desperately wants to feel like an exclusive lifestyle club, complete with stadium naming rights and member-exclusive events. But beneath the lifestyle paint job, SoFi is a heavily regulated financial institution with actual capital requirements and real-world credit risk.

To justify its premium valuation, SoFi points to its Galileo and Technisys platforms. They want you to believe this is an "AI-driven financial backbone" that will power the entire fintech ecosystem. But in reality, they are trying to convince the market to value them like a high-growth software company instead of a bank that has to worry about interest rate spreads.

They will offer you an "AI financial coach" to guide your "member journey," but at the end of the day, a high-yield savings account is still just a deposit, and a mortgage is still just a debt. No amount of silicon valley vocabulary changes the fact that they are holding your cash and hoping you don't default on your loan.

3. PayPal: The Tech Grandfather's AI Face-Lift

PayPal (PYPL) is the pioneer that stayed at the party too long. They are watching younger, sleeker rivals like Apple Pay and Google Pay effortlessly take their market share. In a desperate bid to show they still have a pulse, they have gone all-in on "AI-driven personalization."

The narrative is that PayPal's AI will predict what you want to buy before you buy it, streamlining the checkout process to save you precious seconds. The reality? Their "groundbreaking" AI features mostly involve the app remembering your shipping address and suggesting a coupon.

These are features most of us would call "the bare minimum of internet shopping since 2005." It is the digital equivalent of your grandfather putting on a supreme hoodie and claiming he's a "disruptor." Adding AI to a checkout button doesn't make you Nvidia; it just means you've automated the process of sending spam emails with the customer's name on them.

4. Robinhood: The Dopamine Arcade and the Nudge Economy

Then we have Robinhood (HOOD). While the other three use AI to justify their valuations, Robinhood uses it to solve a much more practical problem: How do you keep a retail trader clicking "Buy" when they are actively losing money?

This is the Dopamine Algorithm. Robinhood has mastered the art of behavioral nudges. They don't need to build a sentient AI that beats the market; they just need an AI that predicts exactly which push notification, color change, or "market alert" will trigger you to trade a zero-day option at 10:00 PM.

They've introduced a "24-Hour Market" and "prediction contracts," turning the global financial system into an interactive mobile game that never sleeps. If you lose money on a trade, you aren't met with a cold warning; you are met with polished interfaces and "AI-curated insights" that make the loss feel like a temporary setback in a video game. It is behavior modification masquerading as wealth management. They don't need you to win; they just need you to play.

"The unreal has no existence; the real never ceases to be."
― Bhagavad Gita

Corporate narratives constructed entirely out of software buzzwords have no true existence. They are phantoms. The reality of capital requirements, persistent inflation, and rising default rates—the real—is the only thing that actually endures. No matter how many GPUs you buy, the classic rules of arithmetic remain undefeated.

The Four Horsemen of Fintech

Data Sources & References