80 Basis Points, $8,000 on a Million, and the Fee Gap Nobody Prices In

I spent a Tuesday afternoon last week doing something nobody enjoys. I rebuilt a single trade from scratch, leg by leg, and wrote down every fee on the way.

The trade was simple. Swap one stablecoin for another, move the proceeds, land dollars in the right account. Two clicks on paper. In practice it's four line items, and one of them came in about 80 basis points heavier than I expected.

That gap is the whole story. Not because 80 bps sounds dramatic. Because almost nobody prices it in until after the trade clears.

Where The 80 Basis Points Actually Lives

Here's the part that gets flattened in most comparisons. A trading fee isn't a number. It's a conditional. It depends on the pair, your volume tier, whether you're making or taking, and which stablecoin you're holding on either side of the trade. Kraken's stablecoin conversions sit at a different rate than its deepest-liquidity pairs, and when you stack that against what a competing venue charges on the same nominal size, the spread between them works out to roughly 80 basis points.

Eighty basis points is 0.8%. On a $10,000 rotation that's $80, which feels like nothing. On $100,000 it's $800. On $1 million it's $8,000, and if you're a treasury desk rotating size weekly, you're lighting something like $416,000 a year on fire for the privilege of holding a dollar-pegged asset in a slightly different wrapper.

That's the number I keep coming back to. Stablecoins are supposed to be the cheap part of crypto. The boring part. The part where you park capital between trades and barely think about it.

And yet the fee on moving a dollar is often higher than the fee on moving something that can drop 20% while you're asleep.

If converting USDC to USDT costs more than buying a volatile token, what exactly are you paying for? Because it isn't risk. The venue isn't taking directional exposure on that trade. It's matching two sides of a pegged pair and collecting a spread on the way through.

Now layer on the rest of the route, the stuff that never makes it into the headline rate. Payouts: the wire fee, the ACH window, the withdrawal minimum that forces you to move more than you wanted. Account access: whether the cheap venue will even let you in from your jurisdiction, and whether the pair you need is listed for your region at all. Remaining route costs: the network fee to move the balance, the slippage on a book that's thinner than the ticker suggests, the spread you cross twice if you've to hop through an intermediate asset.

The 80 basis points is real. But it's the opening bid, not the final invoice.

The Fee Schedule Is The Product

Pull the camera back and this stops being about one venue's pricing page. It's about what exchanges are actually selling.

Stablecoin flow is the best business in crypto. Low inventory risk, high turnover, institutional clients who need it every week and can't opt out. A venue that owns that flow owns a recurring revenue stream that doesn't care whether the market is up, down, or sideways. That's the bet. That's where the conviction sits.

So when one exchange prices stablecoin conversions meaningfully above the cheapest available route, that's not an accident or an oversight. It's a read on how much friction its customers will tolerate before they bother to move. And for a lot of desks, the answer is a surprising amount, because moving custody is annoying, compliance is annoying, and the ops team already knows the buttons on the current platform.

Here's my first hot take. Most fee comparisons in this industry are marketing dressed up as journalism. They quote the taker fee on the deepest pair, ignore the withdrawal cost, ignore the off-ramp, ignore the fact that the cheapest venue in the table won't onboard you, and declare a winner. That's not a comparison. That's a screenshot.

Second take, and this one's less comfortable. Stablecoin pairs should be close to free. The risk to the venue is minimal, the inventory turns over in hours, and the volume is enormous. Any platform charging a meaningful premium over single-digit basis points on a pegged pair is charging you for convenience, not for risk. That's fine as a business model. Just don't call it a market rate.

What I'd Do With This

Don't shop on the headline fee. Shop on the full route.

Write down four numbers before you move anything: the conversion fee at your size and tier, the withdrawal or payout cost to get fiat or the target asset where it needs to go, the network fee, and the spread you'll cross. Add them. That total is your real cost, and it's usually 2x to 3x whatever the pricing page advertised.

Then ask the question that actually matters. Is the venue you're already on close enough that switching isn't worth the operational headache? Sometimes the answer is yes. Eighty basis points on a small rotation is noise, and a migration project costs more in people-hours than it saves in fees.

But if you're moving seven figures a month, the math flips fast. At that size, 80 basis points isn't a fee. It's a line item someone should be fired over, or at minimum, a line item someone should be negotiating.

I asked a treasury lead at a mid-sized fund once why his desk stayed on a venue with fees he openly complained about. He paused before answering. The kind of pause that means the real answer is next.

"Because the wire clears in two hours and the auditor already has the statements," he said.

That's the story the pitch deck won't tell you. Fees don't win or lose on price alone. They win on the cost of leaving. So check the route, not the rate. And check what it would actually take to walk away, because that number is the one your exchange is really pricing.