Open Banking Grows Teeth: PSD3 Inches Closer, Thought Machine Clears $100m, and BaaS Finally Grows Up
As of this week, open banking in Europe is less about theoretical rules and more about commercial proof. The rulebook is nearly ready. The vendors are, for once, nearly ready too.
If you have been watching the open banking file in the European Union, you will know the running joke: PSD3 is always almost there. For most of 2024 and 2025, almost there meant another political compromise and another delayed vote. The latest signals from Brussels, picked over this week by Vixio, Norton Rose Fulbright and the European Banking Authority, suggest the file is finally closing in on the Official Journal.
PSD3: Where We Actually Stand
Norton Rose Fulbright reiterated that the PSD3 and Payment Services Regulation (PSR) texts were finalised politically in April 2026. The remaining steps are the formal European Parliament plenary vote, signature, and publication in the Official Journal, originally expected in June or July 2026 and now, per law-firm trackers, more likely in the autumn. Clifford Chance and Sopra Banking briefings reach broadly the same conclusion. Once in force, PSD3 is a directive, so member states get 18 months to transpose. That puts applicability on a Q2 to Q3 2028 track.
What Actually Changes
The headline changes land in four areas: more granular API uptime and latency metrics; mandatory contingency fall-backs when APIs go down; a clearer allocation of fraud liability for impersonation and authorised push payment fraud; and stronger enforcement powers for national competent authorities and the EBA, including quicker licence withdrawal for persistent API failures.
FIDA Alongside
The Financial Data Access Regulation (FIDA) continues to track behind PSD3. Norton Rose Fulbright's June note said formal adoption remains likely in mid-2026, with implementation in late 2027. The operationally awkward question is whether PSD2's API estate can be re-used for FIDA's broader open-finance scope. The answer, uncomfortably, is: it depends on which data category you are exposing.
Core Banking Finds Commercial Proof
The core banking category has long been accused of being an ARR mirage: lots of signed logos, slow actual migrations. The last few months have forced a reassessment.
Thought Machine's $100m Milestone
Thought Machine disclosed in September that its annual recurring revenue surpassed USD 100 million as of Q2 2026, according to Finovate and FinTech Futures. Its 2025 revenue reached £74.8 million (up from £47.6 million in 2024), per filings. In May 2026, the company closed a Series E and followed up with a further £30 million transaction, split into £9 million of primary funding via convertible loan notes and £21 million of secondary share sales. The company has committed to hiring over 100 new engineers in 2026 and a London IPO is reportedly back on the table.
Mambu Goes Agentic
In September 2026, Mambu launched what it describes as an agentic AI core, a layer of AI agents on top of its cloud-native platform designed to automate configuration, product set-up and operational monitoring. If 2015 to 2020 was about moving the core to the cloud, 2026 to 2028 looks set to be about moving the implementation effort to AI.
10x and the Composability Question
10x Banking continues to anchor the composable-banking narrative alongside Backbase. The CIO question has moved from which vendor survives to which operating model fits our bank's actual change capacity.
BaaS Finally Looks Like a Business Model
Banking-as-a-Service has had a difficult 24 months. US enforcement actions in 2024 and 2025, European DORA alignment demands and a few high-profile fintech partner failures have made the category sound more like a cautionary tale than a growth story. That is finally shifting. Analyst projections cited by Sopra Banking suggest the global BaaS revenue pool will grow from USD 1.7 billion in 2021 to over USD 17.3 billion in 2026, a tenfold increase.
The Treezor Precedent
The acquisition of Treezor by Shares earlier this month, as reported by Finextra, is the clearest signal that scale-ups now treat BaaS infrastructure as a buyable asset. Expect at least two more mid-size BaaS consolidations by Q1 2027.
What To Watch Over The Next Fortnight
Three concrete items: PSD3 Official Journal publication (watchword: autumn 2026); Q3 earnings colour from core banking vendors including Thought Machine, Temenos and Mambu; and BaaS M&A activity among Nordic, French and German providers where the economics without scale have never worked.
The Takeaway
The headline is that open banking in Europe is finally boring in the right way. The rules are nearly done. The vendors are nearly ready. The operating models are nearly agreed. For CIOs and CTOs, the practical question has shifted. When PSD3 lands next spring and FIDA follows a year later, does your platform make the next decade cheaper to serve, or just more expensive to comply with?
Published by FinanceX Magazine, 21 October 2026.
