What AT1 Bonds Are, and Why They Decide UBS's Future

In March 2023, a batch of bonds that called themselves "Tier 1 capital" — higher quality than ordinary bonds, lower quality than equity — were wiped to zero. The equity holders of the same bank still had something left. That inversion of the usual capital stack was supposed to be impossible under the rules that were supposed to be in place.

The bonds were called Additional Tier 1, or AT1. The bank was Credit Suisse. And that event is the ghost inside every paragraph of a regulatory fight that has been consuming UBS for the past two years.

The Swiss government wants UBS to hold roughly $20 billion in extra capital to back its foreign operations. Not because the bank is in trouble. Because the government decided that after Credit Suisse, the capital backing those operations should only be the hardest kind — Common Equity Tier 1, or CET1. Real equity. No shortcuts.

Lawmakers want to weaken that rule. They're considering letting UBS use AT1 bonds — the very same instrument that caused the Credit Suisse mess — to cover part of the requirement. The government calls AT1 "less secure." Lawmakers call it "less expensive." Both descriptions are true, and that tension is the whole story.

What you're actually looking at here is a classification boundary. The kind of boundary that determines who gets paid, who gets written off, and what a bond is legally allowed to be called.

Banks hold capital in layers. At the top is CET1 — common equity, retained earnings, things shareholders actually own. Below that sits AT1 — technically perpetual bonds, meaning the bank never has to return the principal. They pay a coupon like interest, and they look like fixed income. But they carry a contract that says: if the bank's capital falls below a certain trigger, these bonds can be written down or converted into equity. Sometimes to zero. Always before equity is touched.

The basic point is that AT1 is debt wearing an equity costume. It counts as capital because regulators decided banks need more than just equity to absorb losses. But it's still debt, and debt was supposed to sit below equity in the payoff hierarchy. Credit Suisse broke that rule.

When FINMA, the Swiss banking regulator, ordered the $17 billion wipeout, it did so because Credit Suisse was failing and needed to be recapitalized so UBS could buy it. The AT1 was written down first. Common equity — which had been bought at $1 per share as part of the rescue — still existed. Bondholders who bought Credit Suisse AT1 expecting something safer than equity got nothing. Equity holders got everything relative to them.

That was supposed to be the wrong order.

The court system is still untangling it. In October 2025, Switzerland's Federal Administrative Court overturned FINMA's decision to write down the AT1 bonds. FINMA appealed. The legal case isn't over. But the market already learned its lesson: Swiss AT1 bonds carry a unique risk that European and American AT1 bonds don't. The European Central Bank and Bank of England explicitly said they would not do what FINMA did. The problem wasn't the instrument. The problem was the regulator's discretionary power over the instrument.

Now the same instrument is being debated as the solution to a different problem.

After absorbing Credit Suisse, UBS became Switzerland's sole global bank. Its balance sheet is bigger than the Swiss economy. The government's response was to draft rules requiring UBS to fully back its foreign subsidiaries with CET1. Under current rules, UBS only needs 60% CET1 backing for those operations. The proposal pushes it to 100%. The cost: roughly $20 billion in new CET1, phased in over seven years.

On top of that, the Federal Council added roughly $2 billion more in requirements around how UBS treats capitalized software and valuation adjustments, and about $15 billion in legacy requirements from the Credit Suisse merger itself. Total additional capital across all measures: roughly $37 billion. Annual cost of that extra capital: UBS estimates $3 billion.

UBS called the proposal "extreme" and "out of step internationally." Their competitors in Europe and the US use risk-weighting or partial deductions instead of full CET1 deductions. And while Switzerland tightens, the US is expected to release $138 billion in CET1 capital and the UK plans to reduce minimum Tier 1 requirements from 14% to 13%.

So here's where the AT1 idea comes back in. Lawmakers — pushed by UBS and skeptical of the economic damage — are considering letting UBS use AT1 bonds to cover up to 50% of the foreign unit capitalization. The government originally rejected that option. But the compromise would let UBS issue roughly $10 billion in AT1 instead of raising that same amount in equity. AT1 pays investors around 7% (UBS recently issued tranches at 6.625% and 7%). Equity doesn't pay a coupon — it costs the bank through diluted earnings and opportunity cost. Either way it's expensive, but one kind of expensive is a promise to bondholders and the other is a claim on future profits. Different incentives, different risk, different politics.

The regulatory trigger mechanism matters here too. The committee is discussing introducing a specific capital-ratio floor that would force UBS to stop paying dividends and buybacks if it falls below the threshold. The idea is to make AT1 more "secure" — if the bank can't distribute cash to shareholders before the AT1 gets stressed, the AT1 holders have an extra layer of protection. Whether that actually works depends on what happens between the trigger point and the point at which the trigger fires. There's always a gap.

What this means for the investment case is less dramatic than it sounds, but the direction matters.

UBS currently has about $20 billion in AT1 on its books. The bank has been issuing more: $3 billion in January at 6.625%–7% yields, another $1.5 billion in June at 6.875%. Investor demand was strong in every case — more than $21 billion of bids for the January issue alone. The market appetite exists. The question is whether the regulatory permission exists.

If lawmakers approve the AT1 compromise, UBS can meet a chunk of the new requirement by issuing debt rather than accumulating equity or doing a costly equity raise. That's better for existing shareholders — less dilution, more flexibility, continued capacity for buybacks. The bank already said its 2026 buyback plan depends on getting regulatory clarity.

If lawmakers stick with 100% CET1, UBS has to hold that capital in retained earnings or raise equity. Either path means lower returns on equity, smaller buybacks, or both. The bank's current return on CET1 is around 16–17%. Holding extra capital without earning extra revenue on it mechanically pulls that ratio down. At a $170 billion market cap and roughly 1.9x book value, the market is already pricing in some uncertainty here.

The timeline adds to the option value. The parliamentary committee failed to reach agreement in August and reconvened at month's end. An upper house vote is targeted for September, with final adoption not expected before December at the earliest. The actual rules might not take effect until the end of 2026 or beyond. In the meantime, UBS's capital position is strong enough to absorb the wait — the bank isn't insolvent, and the requirements don't kick in tomorrow.

The real risk isn't that UBS can't raise the capital. It's that the uncertainty itself becomes a tax. The bank reports it has underperformed European and US bank indices by 27% over roughly the past two years of regulatory debate, which works out to roughly CHF 30 billion in lost market value. That underperformance isn't from operational failure. It's from investors sitting on their hands while a classification boundary gets argued out in committee rooms.

The stock has recovered most of that loss in the past year, up roughly 20% year-to-date and trading near its 52-week high of about $56. Part of that move likely reflects the market's growing sense that lawmakers will soften the rules. The question going forward is whether the compromise is enough, or whether the gap between what UBS needs and what the rules allow keeps the overhang alive.

The structural takeaway is simpler than the politics. AT1 is a hybrid instrument that sits at a classification boundary — debt that can be equity, capital that isn't quite capital. That ambiguity is the whole point. It lets banks hold cheaper capital. It lets regulators create a loss-absorbing layer below shareholders. And it let the Swiss government do something in 2023 that broke a hundred years of capital hierarchy. The current debate is just the same classification question asked again: what is good enough to call capital, and who decides.