Definition
Treezor, Swan and Solaris are three European heavyweights of Banking-as-a-Service: authorised (as EMI, PI or credit institution), they provide account, IBAN, card, transfer, direct debit and KYC as an API.
It is the invisible infrastructure that lets hundreds of fintechs and neobanks exist without becoming banks themselves:
- Treezor (FR) — EMI, a subsidiary of Société Générale since 2019. Behind Lydia, the early days of Qonto, Younited, Sumeria and more than 200 fintechs.
- Swan (FR) — an independent EMI, founded in 2019 by Nicolas Benady. A modern API pure player, oriented towards tech teams.
- Solaris (DE, formerly Solarisbank) — a German credit institution (Vollbank, a full bank), founded in 2016. Broader coverage (account + card + credit), the historic infrastructure of Vivid, Tomorrow and, in part, Trade Republic.
| Player | Country | Status | Founded | Subsidiary | Specialty |
|---|---|---|---|---|---|
| Treezor | FR | EMI + PSP agent | 2016 | Société Générale (2019) | FR, multi-service |
| Swan | FR | EMI | 2019 | Independent | Modern API, tech-savvy |
| Solaris | DE | Credit institution (Vollbank) | 2016 | Independent | Credit + account + card |
Treezor
Founded in Paris in 2014 by Éric Lassus, Xavier Labouret and Grégoire Bourdin, authorised as an EMI in 2016, then acquired by Société Générale in 2019. Strong growth since (more than 100 corporate clients, several million Mastercard cards distributed).
- Offering: dedicated IBAN per user, physical and virtual cards (Mastercard), SCT / SCT Inst / SDD transfers, integrated KYC, AML/CFT compliance.
- Clients (~200): Lydia, the early days of Qonto, Younited, Sumeria, Lemonway, Mansa, Defacto, Karmen.
- Strengths / limits: the SG backing (solidity), maximum FR coverage, experience — but agility that is sometimes seen as lower because of belonging to a large group.
Swan
Founded in Paris in 2019 by Nicolas Benady (ex-Smile&Pay), authorised as an EMI in 2020. A €16M Series A (2021), a €37M Series B (2023, Lakestar, Creandum), then expansion into Spain, Germany, the Netherlands and Italy.
- Offering: a modern REST API very oriented towards developer experience, dedicated IBAN, virtual and physical cards, SCT / SCT Inst / SDD transfers, embedded integration in a few days.
- Differentiation: an API pure player not backed by a legacy bank, Stripe-like docs, public pricing, setup in under a month for standard cases.
- Clients: Pennylane, Indy, Spendesk, Klaxoon and more than 100 fintechs/SaaS.
- Strengths / limits: agility, docs, independence — but narrower coverage than Treezor.
Solaris
Founded in Berlin in 2016 by Marko Wenthin and Andreas Bittner (the company builder finleap), with a full banking licence (BaFin, Vollbank). Massive growth and nine-figure fundraising (2018-2021), rebrand from Solarisbank to Solaris (2022), then financial difficulties and restructuring (2023-2024) before a refocusing in 2025.
- Offering: passported German account and IBAN, debit and credit cards, consumer credit (the differentiator tied to the banking licence), trading/investment, integrated KYC.
- Clients: Vivid Money, Tomorrow, Trade Republic (historically, in part), Kontist, Yokoy.
- Strengths / limits: a full banking licence (so credit is possible), strong German anchoring — but financial difficulties in 2023-2024 and a freeze on some new clients.
BaaS vs alternatives
| Approach | Advantages | Disadvantages |
|---|---|---|
| BaaS (Treezor, Swan, Solaris) | Fast launch, product focus | Dependency on the player, reduced margins |
| PSP agent (Lemonway, MangoPay) | Simpler, cheaper | Less control over the service |
| Own licence (PI / EMI / CI) | Total control, maximum margins | Cost (~€1M to €3M), lead times (1 to 2 years), complex operations |
2024-2025 trends
- Consolidation: more than 50 BaaS at the peak (2021), many have disappeared or pivoted.
- Reinforced regulation: BaFin, ACPR and CSSF impose stricter controls (cf. Solaris).
- Margins under pressure: strong competition from licensed players (Qonto, Revolut, Sumeria).
- Internalisation: fintechs that grow take their own licence (Qonto in 2022, Trade Republic, Lydia → Sumeria in 2024), which disintermediates the BaaS.
What they are not
- Not consumer brands: the user sees Lydia, Qonto or Pennylane, never the BaaS.
- Not B2C fintechs: 100% B2B infrastructure.
- Not equivalent to one another: Treezor and Swan are EMIs (no credit), Solaris is a full bank.
- Not all sustainable: a heavily consolidating market.
In the PSD2 / Open Finance ecosystem
The BaaS players are the hidden infrastructure of the ecosystem:
- PSD2: all can take on the AISP/PISP roles, some the full roles.
- PSD3 / FIDA: an extension planned to provide enriched data to their clients.
- VoP, IPR: mandatory rollout since October 2025.
- Embedded finance: they form its base in Europe (Stripe, Adyen for Platforms…).
Concrete examples
- Lydia → Sumeria: started at Treezor (2017), then its own EMI authorisation via Sumeria (2024) — disintermediation.
- Qonto: started at Treezor (2017-2022), then its own PI licence (2022).
- Pennylane: Swan for the account integrated into the accounting SaaS.
- Vivid Money: Solaris for the EU account, with its own licence since 2023.
- Trade Republic: a move from Solaris to its own brokerage + PI licence.
- Indy (FR): Swan for the business account for the self-employed.
- Solaris difficulties: in 2023, BaFin imposed an audit and limited onboarding, affecting Trade Republic and Vivid, with a heavy restructuring in 2024.
- Outside the top 3: Mangopay (FR, marketplaces), Lemonway (FR, marketplaces), Modulr (UK, B2B payments), Banking Circle (LU, cross-border).
- Cost for a fintech: €0.5 to €2 per account per month plus 0.1% to 0.3% per transaction, versus €1M to €3M of setup and as much per year for an own licence.