Bank of America's GLP-1 Bet Is Not About Health

A few days ago, Brian Moynihan told CNBC that Bank of America spends more than $250 million a year on weight-loss drugs for its employees. Five years ago the bill was zero. Today it accounts for 13 percent of the bank's $2 billion annual healthcare budget, covering roughly 211,000 people.

Moynihan called it a "good investment." The usual reporting treated it as an act of progressive employer benevolence. Neither frame is wrong, but both miss what's actually going on.

The more interesting question is why a bank is making a healthcare bet that doesn't make financial sense — at least not in the way Moynihan describes it.

The economics of employer-funded GLP-1 coverage are not flattering. The average annual drug cost per person is around $6,540. A case study by AssuredPartners found that GLP-1 users saved roughly $560 per year in other medical costs. That's negative ROI by a factor of more than ten. An Aon analysis published in 2026 found a more encouraging signal — six percent lower medical cost growth among GLP-1 users with diabetes after 30 months — but that's a slower growth rate, not a cost saving, and it takes two and a half years to appear.

There's also the turnover problem. The average employee stays at a company about four years. Moynihan acknowledged this himself: some employees will leave before the bank realizes the long-term savings from improved health. The drug costs money immediately. The health benefit is speculative and arrives on a timeline most workers won't stay long enough to experience.

Then there's the adherence issue. About 45 percent of GLP-1 users stop treatment within a year. A significant fraction of that $250 million pays for prescriptions that people abandon before completing a full course.

Against all this, most employers are retreating, not advancing. PwC cut GLP-1 weight-loss coverage from its US employee plans in May, restricting it to diabetic patients only. A survey of nearly 300 employer health plans found that 36 percent of companies cover these drugs for both diabetes and weight loss — flat compared to 2025. Only 9 percent are considering expanding coverage. Ten percent of companies that currently cover these drugs plan to stop.

GLP-1 medications now account for 11.4 percent of annual pharmacy claims, up from 6.9 percent in 2023. Nearly eight in ten employers report that these drugs are driving up their healthcare costs. Only nine percent expect prices to decrease.

So Bank of America is spending $250 million on a treatment that costs more than it saves, that most users abandon within a year, whose long-term payoff may arrive after employees have left, and that most other large employers are actively pulling back from. Why?

The answer is in a detail most coverage of Moynihan's comments skipped. Nearly 30 percent of employees say they would switch jobs for GLP-1 coverage. That figure comes from the 2026 NFP U.S. Benefits Trend Report, and it's the number that makes the $250 million start looking less like a health program and more like a retention wall.

When 29 percent of your workforce says they'll leave if you don't cover these drugs, the math changes. You're no longer evaluating the treatment on its medical return. You're evaluating the cost of keeping your people from walking out the door. And $250 million — about $1,185 per employee — may be cheaper than the alternative.

Bank of America's revenue for the first half of 2026 was roughly $62 billion. The GLP-1 spend works out to less than half a percent of revenue. In absolute dollars the number is impressive enough to headline a CNBC interview. As a fraction of the bank's scale, it's a rounding error.

But the real question isn't whether the bank can afford it. It's whether the stated rationale matches the actual incentive structure. Moynihan cited cardiovascular benefits and called the program a worthwhile investment in his workforce. He paired drug access with health coaching to monitor lifestyle changes. None of that is false. But the health rationale and the retention rationale are not the same thing, and confusing them leads to bad decisions.

If GLP-1 coverage is a retention play, then the bank should track it like one: measure turnover rates among eligible employees, compare the cost of replacement hiring and training, and evaluate whether the drug is actually keeping people who would otherwise leave. If it's a health program, the bank should track clinical outcomes, adherence rates, and downstream medical costs. The two goals require different metrics, and treating one as evidence for the other is a category error.

I suspect Moynihan knows both are true. The drugs probably do improve health. They probably also keep employees from jumping ship. But "good investment" is a phrase that collapses the two into one, and that collapse obscures the real decision the bank is making.

The test is simple. Watch what happens next. If Bank of America is truly banking on the health argument, it will start publishing data — aggregated or anonymized — showing improved outcomes among its GLP-1 users. If it's banking on retention, the program will stay quiet and the turnover metrics will do the talking. The fact that other employers are pulling back while Bank of America doubles down suggests the bank sees something different about what this money actually buys. That observation is worth paying attention to, even if it isn't what the CEO said on camera.