Pontes stops being a forecast: the ECB's tokenized settlement bridge goes live with 13 banks, and full capability is still two years out.

Tracee flagged the 21 September target three weeks early, from a warning, not a roster. The roster is what is actually new here.

Separate what tracee already reported on 6 September from what the go-live actually added.

One claim confirms a forecast. One claim is a roster nobody had. The rest oversell a launch day that is narrower than the coverage suggests.

Here is what plugs into what, and what runs alongside it outside the Eurosystem entirely.

• The rail is now real, not a slide. Thirteen institutions moved from a named intention to a named deployment.
• The fork predates Pontes. Fnality has run a central-bank-money-referenced private network since 2023; Deutsche Bank and Santander back both it and Pontes.

Tracee's 6 September briefing decoded ECB Executive Board member Isabel Schnabel's Jackson Hole warning: central banks put their own money on-chain, or dollar stablecoins become the default settlement layer for tokenized markets. That briefing had a date and a warning. This one has a roster: 13 named institutions, four named DLT operators, and a central bank that put its own balance sheet, however small, on the line.

One of the 13, KfW, already has a real-money deadline riding on this rail. Tracee's June briefing on KfW's eWpG bond reported that its EUR 100M issuance committed coupon and redemption cash flows to Pontes ahead of a December 2026 payment. That payment is now the first scheduled test of whether the bridge that just launched actually clears real money on time.

The quote, from ECB Executive Board member Piero Cipollone, names the bet plainly: trust is the product, not throughput. Nothing in the launch materials claims Pontes moves faster or cheaper than Fnality's private rail. The claim is narrower and, for a central bank, more defensible: settlement in central bank money carries no counterparty risk that a privately issued token, however well collateralised, can fully remove.

• Eight hours a day, not always-on. Trading runs 8:00 to 16:00 CET on business days. The 22.5-hour and round-the-clock milestones are dated 2028, not launch day.
• The ECB's own commitment is undisclosed. "A small portion" of its funds is not a number. Until the ECB publishes one, its own participation reads as a symbolic anchor more than a liquidity commitment.
• Programmability waits too. Expanded programmability is scheduled from mid-2028 alongside round-the-clock operation, not available now.
• A private incumbent already exists. Fnality has settled wholesale transactions in central-bank-referenced tokens since 2023. Pontes adds a central-bank-operated alternative; it does not create the category.
• Its own long-term status is contested. Germany's banking association has reportedly framed Pontes as a near-term interoperability layer, with the ECB's separate Appia project positioned as the durable architecture. The ECB has not itself resolved that framing in public.

The same month produced three distinct answers to who moves money in a tokenized market. Canada's OSFI, in tracee's 10 September briefing, ruled tokenized deposits are legally unremarkable and left banks to build their own products under existing supervision. Swift's DBS-Citi pilot, covered 7 September, routed a tokenized-deposit payment over a messaging network banks already trust, no new central-bank infrastructure required. The ECB took the more expensive route: it built and now operates the settlement bridge itself, with its own balance sheet attached, however narrowly.

That is also the difference between Pontes and the wholesale CBDC model tracee covered in India's SEBI and RBI bond settlement, where the central bank's digital currency sits on both sides of the trade. Pontes does not issue a new instrument at all; it lets existing central bank money settle a tokenized asset leg. That is a narrower, more conservative design than a wholesale CBDC, and it is also why the ECB can coexist with Fnality rather than compete to replace it.

What changed on 21 September is operational reality replacing an announced date: 13 named institutions, four DLT operators, and a central bank willing to put an undisclosed slice of its own funds through the bridge. What has not changed is the substance regulators and treasurers actually care about, round-the-clock availability, programmability, and disclosed settlement volume, all of which sit on a 2027 to 2028 roadmap rather than in this launch. Treat 21 September as the ECB proving it can ship on a date it named itself, not as proof the architecture has scaled.

Watch three things:

• KfW's December 2026 coupon. The first real-money cash flow scheduled to clear through Pontes, and the first test of whether the bridge settles on time.
• Whether the 2027 settlement-finality and multi-currency milestone holds. Tracee has flagged the same slippage risk on GENIUS Act rulemaking; a Eurosystem roadmap is not immune to it.
• Whether the ECB or Germany's banking lobby settles the Pontes-versus-Appia framing. A public answer would tell participants whether to treat this launch as durable infrastructure or a transitional step.