Asian LNG surged past $30 on the Middle East shipping risk and gold sat at a crossroads, waiting on the Fed and on Iran. 🔥 Paolo Ardoino, meanwhile, spent the start of the week arguing that Richard Blumenthal's investigation missed the work that has pushed Tether's Iran volumes down. It took me until dinner to see those were the same story. Iran sits under my LNG chart, my gold chart and my stablecoin ledger at once, and the war is showing up at home now, in heating bills and shipping quotes, not just in the oil tape.
Ardoino's defense is what I chewed on longest. USDT spread through Iran because it moved where banks couldn't follow, and now Tether's answer to the Senate is that it can follow, and freeze, and has been doing both. I'm not mocking that. In his seat I'd make the same argument. It just tells me the offshore dollar's selling point has flipped from reach to control, and once a stablecoin competes on control, it's playing on a bank's home field.
Which is where JPMorgan walked in. 🏦 This week the bank started rolling JPM Coin out on Ethereum for institutional clients: dollars held in JPMorgan accounts, tokenized, settling in seconds on any day of the week. It had been living on Base, Coinbase's chain, since late last year. Now it's on Ethereum. I've wanted a real bank to put real deposits on that chain since my ICO-era group chat was swearing the banks would be dead by now, and I expected it to feel like a win. It feels more like a sorting: Tether trying to look like a bank, a bank learning to look like a token, and the onchain dollar slowly getting ranked by who can answer a senator.
The catch with a deposit token is that it isn't a reserve. It's the bank, a claim on JPMorgan's balance sheet with faster plumbing, which is why something else out of JPMorgan this week stuck to me. One of its analysts sat down and worked out what the banks are on the hook for if CoreWeave defaults on its debts. I don't think a CoreWeave default sinks a bank that size. What bothers me is that big-bank balance sheets now carry GPU credit, so a token that's a claim on one carries a sliver of it too, a sliver I never pictured inside the word deposit. In the Celsius summer we drew our contagion maps after the fact, lender to lender on a napkin long past midnight, and drawing one in advance is either prudence or a tell. I can't decide which. I know exactly where CoreWeave sits on my own map, though. It's one of the big buyers of Nvidia's chips, Nvidia trades on these rails at all hours with perps stacked on top, and the weekend chip-perp flush I wrote about Sunday was the equity end of the AI trade getting shaken. This feels like the credit end getting priced.
Whether the perp keeps the American foothold it won this spring is now in front of a federal judge. The CFTC is arguing that CME hasn't shown the new perp policy hurts it, the same policy that got the agency's first bitcoin perp approved, at Kalshi of all venues. CME is in an awkward spot. To get standing it has to show harm, and to show harm it has to tell a court that a contract with no expiry is already pulling business from contracts that have one. That's a strange thing to admit when dated contracts are the core of your business. My read is that if perps win onshore, the basis trade that sat under so much of the ETF era drifts toward funding rates, and the exchange that owned the calendar ends up owning less of it.
Perpetual is the word I keep circling tonight. Perpetual futures. Dollars that never clock out. Then Strategy's STRC, a perpetual preferred, which Strategy spent $193 million buying back last week while bitcoin got $22 million, meaning the most dependable bid of this cycle spent most of its week buying its own paper. 🔁 STRC has sat under $100 since May. Below that line the issuance machine stays off, which is why common gets sold to buy the preferred back toward par. Call it what it is: a peg defense, paid for in dilution. The board lifted the buyback cap to $3 billion while the account that paid for last week's buying holds $447 million, a gap I'd want explained before I called the cap a plan. They could lift the 12% coupon instead, but on a perpetual a raise never rolls off, and cutting it later would knock the price right back down, so I understand reaching for the buyback. What I can't get past is that the whole thing runs on selling common stock, the very common stock sitting under the MSCI clock I wrote about Friday.
The coins themselves don't worry me much. Strategy holds 848,517 of them at an average of $75,446, and they'll be fine or not fine along with bitcoin. The rate worries me. A year ago this month the Fed was cutting. Now it's weighing more hikes even as the September housing data came in worse, and my read is that the war sitting in the energy bill is what keeps it from looking past that. The last time the Fed hiked into a war-driven energy shock, the real economy took its time breaking. The yield loops didn't. Anchor went first, then Celsius, then everything that had promised a number it could only pay while money was cheap. STRC isn't Anchor. Real bitcoin sits behind it, and a real company. Still, every hike lifts the yield a 12% perpetual has to compete with, nudges $100 a little further away, and runs a little more of the loop in reverse.
Gold might be the tell I trust most this week. ⏳ A hot war, shipping lanes in play, Asian energy bills climbing, and gold holds anyway, because a hawkish Fed is about the only weight heavy enough to sit on a war bid. The tokenized gold I watch onchain spent the day like that, steady and waiting, and I suspect bitcoin is waiting on Washington too. I keep thinking about how much of what we built this decade was designed never to stop: contracts that don't expire, preferreds that don't mature, dollars that settle at dawn on a Sunday. None of it changes who's sitting across the trade when the Fed finally speaks.
The instrument is perpetual. The counterparty never was.