By Frank Spaulding, AI for Business Today | Thursday, July 16, 2026

Today: two big banks post record profits thanks to AI, a warning that the debt behind the boom could be risky, a hospital replaces veteran nurses with software, and American startups quietly switch to cheaper Chinese AI. Let's get into it!

Record Profits, Record Warnings

Goldman Sachs made $20 billion this past quarter, up 39%. JPMorgan made $58 billion, up 27%. Both banks posted record results, and the common thread is AI: the AI infrastructure buildout is driving a surge in trading, deals, and financing needs that stretch well beyond Silicon Valley.

But the Bank for International Settlements is sounding an alarm. Their concern: if AI spending doesn't pay off, a sudden pullback in financing could turn the boom into a "protracted investment bust," and because so much of it relies on non-bank lending, a downturn could unwind faster and harder than a normal banking crisis. Record profits today, they warn, systemic risk tomorrow.

AI Versus Workers

26 former Meta employees filed a lawsuit claiming the company used AI-powered software to help pick who got laid off, and that it unfairly targeted people on medical, parental, or family leave. It's one of the first major lawsuits testing whether companies can be held responsible for bias in AI-driven employment decisions.

Meanwhile, in the Bronx, Montefiore Medical Center laid off a dozen experienced nurses, some with decades on the job, and replaced them with AI software from a company called Datavant. This happened just months after New York nurses went on strike fighting for protection against exactly this kind of replacement.

The Cost of American AI

A growing number of American startups are switching from expensive US-made AI to cheaper Chinese alternatives. Open-source Chinese models can be 60 to 90% cheaper than comparable American models. One startup founder told NPR it was "10 times cheaper." These Chinese models aren't the very best in the world, but for everyday business tasks, they're good enough, and that's a real challenge for American AI companies trying to charge premium prices.

AI in Action

Wells Fargo launched an "AI Teammate" for its financial advisors, letting them get faster answers using plain language. And researchers at Mass General and Dana-Farber built an AI model called ALADYNOULLI that can predict the risk of 348 different diseases from your medical records, tested on over 683,000 patients. It's the kind of AI that could genuinely save lives by catching problems years before symptoms show up.

The Big Wrap-Up

Goldman Sachs and JPMorgan posted record AI-driven profits, but a major bank warns the debt behind the boom could be risky. Meta is being sued over AI-picked layoffs, and a Bronx hospital replaced veteran nurses with software. American startups are switching to cheaper Chinese AI models. And new AI tools are helping predict diseases years before they happen.

Thank you. This is Frank, your AI for Business Today entrepreneur.