Building an investment bank takes more than registering a company and hiring analysts. Regulatory approval comes first, then capital, then technology and people, in an order that cannot be shuffled without stalling the whole project. This article maps that sequence for founders, fintech executives and institutional teams weighing the build in the European Union: the authorisation routes and what each costs in time, the capital floors that actually bind, the architecture choices that decide whether the firm can grow, and the thresholds that separate a working boutique from a firm with a second act.
The foundation layer: licensing, capital structure and early architecture
Nothing happens before the regulatory and capital foundation is in place, and most teams underestimate both how long it takes and how far it reaches. The authorisation determines which products the firm may offer, which counterparties will deal with it, and which technology arrangements it can legally enter. It also fixes the reporting obligations, which is why the fintech platform work that follows is shaped by a decision taken months earlier by lawyers. Founders used to iterating their way to a product find this uncomfortable, and correctly so: the sequence really does run in one direction.
The sequencing problem is circular. Authorisation as an investment firm under Articles 5 to 7 of MiFID II (Directive 2014/
Route | Time to market | Indicative cost | Principal risk |
|---|---|---|---|
Acquire a licensed entity | Fastest, subject to change of control approval | Roughly €2 million to €8 million | You inherit its history, including any remediation |
Operate as a tied agent (MiFID II Article 29) | Weeks to a few months | Low upfront, revenue share ongoing | The principal firm owns the permission and the client relationship |
Greenfield authorisation | Longest, measured from a complete application | Legal, compliance and build costs, no acquisition premium | Runway burns with no revenue and no certainty of date |
The capital floors are lower than most founders expect and less relevant than they assume. Article 9 of the Investment Firms Directive (EU) 2019/
Venture capital is one of the few realistic sources at this stage, and it prices the regulatory timeline as the main risk, because an eighteen-month wait burns runway without producing revenue. Model two tranches rather than one. A pre-authorisation tranche covers legal, compliance, technology and the salaries of people who must be in post before the permission arrives. A post-authorisation tranche covers the first eighteen months of operating losses, which begin the day the firm is allowed to trade and not the day it does. That also gives the investor a milestone to price against rather than a single act of faith. Pre-revenue budgets for a European boutique typically land between €2 million and €5 million, and the range is wide because compliance headcount, not technology, is what moves it.
One decision from this phase outlives it. Architecture chosen to save three months in year one is paid for through years two and three, and the payment is usually taken in the form of a rebuild at exactly the moment the firm is trying to scale.
Technology stack decisions that determine scalability
The licensing structure tells a regulator what the firm is. The technology stack decides what it can do at volume, and the two interact in ways that only surface under growth. Three layers carry the load:
- execution: order management, trading connectivity, prime brokerage links
- data: market data, client data, and the regulatory reporting pipelines
- client: CRM, deal management and the client portal
Almost nobody builds the execution layer from scratch any more, and a new entrant should not: established order management systems can be operational in a quarter, and the differentiation is not there. The data layer is where new firms consistently underinvest, and it is not optional. Transaction reporting under Article 26 of MiFIR (Regulation (EU) 600/
Cloud is now a compliance question as much as an engineering one. Moving capital expenditure into operating expenditure genuinely helps a firm with finite runway, and the major providers publish financial services compliance frameworks for that reason. DORA changed what sits alongside the choice: Articles 28 to 30 require a register of ICT third-party arrangements and specific contractual terms including exit, and Article 29 asks whether an arrangement reinforces concentration risk. A tested exit plan is now part of the architecture, not part of the paperwork.
Hiring sequence and the compliance-first principle
Capital and technology create the conditions to operate. The hiring sequence decides whether they are used or wasted on avoidable regulatory failure. The functions below have to be identifiable, resourced and independent before client-facing activity starts, and a supervisor will ask who holds each of them by name.
- A permanent compliance function under Article 22 of Delegated Regulation (EU) 2017/
565, with a compliance officer responsible for it and for reporting to the management body. - A risk management function under Article 23 and, where the firm's scale and complexity require it, an independent internal audit function under Article 24.
- An AML compliance officer and a designated member of the management body under Article 8(4)(a) of Directive (EU) 2015/
849, which the AML Regulation (EU) 2024/ 1624 replaces from 10 July 2027. - At least two people of sufficiently good repute directing the business, under Article 9(6) of MiFID II. Four eyes is a condition of authorisation, not a governance nicety.
The consequence is counterintuitive for anyone from a startup background: compliance and legal hires come before revenue-generating ones, and they are not part-time consultants bolted on after launch. A head of compliance with investment banking experience in a major European financial centre commands roughly €120,000 to €180,000 base, and that cost belongs in the pre-revenue capital plan, where it also raises the fixed overheads requirement. Investors have caught up: a firm that cannot name the individuals holding these functions tends to stall at the term sheet.
Choosing who builds the platform?
Our free guide sets out how to evaluate a software development partner for regulated work: what to ask about delivery evidence, testing and reporting pipelines, and what the reference calls will not tell you.
Scaling the operation: revenue architecture, client acquisition and growth thresholds
Getting authorised is the first problem. The second arrives faster than founders expect: a revenue model that carries the cost base while also funding the next phase. Those two pull against each other, and the firms that scale are the ones that resolve the tension deliberately rather than quarter by quarter. Three revenue streams are available in principle, and very few new entrants can pursue more than one at the start.
- advisory fees, from M&
A, capital markets advice and restructuring, with the lowest capital and infrastructure requirement - execution commissions, from brokerage and market making, which need materially more capital and a working execution stack
- principal investment, from proprietary positions and co-investment, which carries the heaviest capital charge of the three
Advisory first is the path that works, and the reason is structural rather than cautious: execution and principal investment both push the firm into the higher initial capital band under IFD Article 9, and into a far larger own funds requirement, before either has produced a track record. Fee income and closed deals are what buy the capital and the people to add execution in year three or four. As a planning anchor, a three-banker advisory boutique can reasonably target €2 million to €4 million of annual fee income in its second full year, assuming mandates convert at roughly a third and average transaction fees run in the low hundreds of thousands. Those are numbers to argue with, not a benchmark to quote.
Phase | Primary revenue | Capital requirement | Milestone that unlocks the next phase |
|---|---|---|---|
Year 1 | Advisory only | Initial capital plus the fixed overheads requirement | First mandates closed, reporting pipeline proven |
Years 2 to 3 | Advisory plus selective execution | Higher band under IFD Article 9, own funds scale with the book | Repeat clients, and a cost base covered by fees |
Year 4 and beyond | Full service, or acquisition | Capital planning tied to the risk-to-firm and risk-to-market measures | A franchise a buyer or an investor can price |
Client acquisition is the part no platform solves. In investment banking it is almost entirely relationship-driven, which makes the founding team's network the firm's main asset in years one and two. A firm launching without senior bankers who carry portable client relationships is not an investment bank in any working sense. It is a technology platform waiting for a business model, and it will spend eighteen months to two years building a pipeline before the first mandate closes.
One structural advantage of authorising in the EU deserves more attention than it gets. Articles 34 and 35 of MiFID II let an authorised investment firm provide services across the whole single market on the strength of its home authorisation, through cross-border services or a branch, with a notification rather than a fresh application. That changes the arithmetic of the licensing cost: it is one process, amortised across the market the firm can then reach.
Venture capital returns as a question in the scaling phase, when moving into execution or principal investment raises the capital requirement again. Investors want either a credible liquidity event, usually acquisition by a larger institution, or a transition to profitability, because investment banking does not fit the return profile a ten-year fund is built around. Growth equity or a strategic investment from an established institution is often the better fit, and saying so early saves a quarter of conversations that were never going to close.
The last discipline is the one most firms skip. Set an explicit revenue and margin threshold at the start of each growth phase, with a named date on which someone has to state whether it was met. A growth initiative without a defined success criterion runs indefinitely, because no single quarter ever looks bad enough on its own to trigger the decision.
Regulated build and scale work we have shipped
Building the platform behind the licence?
WislaCode builds regulated financial software in the EU: reporting pipelines, client portals, integrations with execution and data providers, and the evidence trail a supervisor expects.
How much capital do you need to start an investment bank?
The regulatory floor under Article 9 of the Investment Firms Directive (EU) 2019/2034 is €750,000 for a firm dealing on own account, €150,000 for one holding client money or securities, and €75,000 for one doing neither. The floor is not the budget. A European boutique typically needs somewhere between €2 million and €5 million of pre-revenue capital to cover authorisation, compliance staffing, technology and around eighteen months of operating losses.
How long does it take to get authorised as an investment firm in the EU?
Article 7(3) of MiFID II requires the competent authority to decide within six months of a complete application. Elapsed time is usually longer, because completing the application is the slow part and an incomplete submission restarts the clock rather than pausing it. Teams that engage regulatory counsel before submitting rather than after tend to shorten the calendar meaningfully.
Can a fintech startup build an investment bank from scratch?
Yes, but the path is more constrained than in other fintech verticals. Capital requirements and mandatory governance functions mean a team cannot iterate its way to authorisation: the compliance function, the risk function and two directors of good repute have to exist before the permission does. Plan for eighteen to twenty-four months of pre-revenue expenditure and a compliance-first hiring sequence.
What is the difference between a boutique and a full-service investment bank?
A boutique concentrates on advisory work, M&A, restructuring and capital markets advice, and does not hold client assets or deal on its own account. A full-service firm adds execution, prime brokerage and often principal investment. The difference shows up directly in the capital band under IFD Article 9 and in the operational infrastructure each requires.
How do investment banks attract their first clients?
Through the personal relationships of the founding bankers, almost exclusively, in the first two years. A firm launching without principals who carry portable client relationships from prior roles typically spends eighteen to twenty-four months building a pipeline before the first mandate closes. No amount of technology shortens that.
Can venture capital firms invest in investment banks?
They can and do, particularly where the model is technology-differentiated, but they price the regulatory timeline as the main risk and will ask hard questions about the route to liquidity. Investment banking does not fit the return profile of a ten-year fund. Growth equity or a strategic investment from an established institution is often a better structural fit for the scaling phase.
What technology does a new investment firm need on day one?
At minimum a regulatory reporting pipeline covering transaction reporting under MiFIR Article 26 and, where relevant, trade reporting under EMIR Article 9; a client data management system; and compliant communication and record-keeping. Order management and trading connectivity are needed only if the firm executes. Many advisory boutiques run a genuinely lean stack for their first two years.
What are the biggest mistakes teams make when building an investment bank?
Three recur. Underestimating the authorisation timeline and burning capital before the permission arrives. Hiring revenue-generating staff before the compliance and risk functions exist, which a supervisor treats as a defect rather than a sequencing preference. And building technology that cannot scale without a rebuild. Each is recoverable, and each costs six to twelve months of runway.
How do investment banks make money in their early years?
Advisory fees, almost exclusively, for the first two to three years. A mandate typically combines a monthly retainer during the engagement with a success fee expressed as a percentage of transaction value. A small team closing a handful of transactions a year can reach breakeven on that model without holding client assets or dealing on own account.
Does an EU authorisation let a firm operate across Europe?
Yes, and it is the main structural advantage of authorising in the EU. Articles 34 and 35 of MiFID II allow an authorised investment firm to provide services across the single market on the strength of its home authorisation, either cross-border or through a branch, on notification rather than a fresh application. One authorisation process is amortised across every market the firm can then reach.