AI companies’ revenues are reaching heights not seen since the dotcom bubble. Some companies exceed the 100 million euros mark in recurring annual turnover after barely two years of existence. This speed is impressive, but problematic for finance teams, because their systems are not designed for such rapid growth.
At the house of Airwallexwe spend most of our time talking to founders and CFOs of AI scale-ups and we hear the same story over and over again. These leaders tell me about creating a truly innovative company, raising considerable funds and recruiting at a sustained pace in several markets. Then, they admit, in an embarrassed tone, that their financial infrastructure is “a bit chaotic”.
The truth is often worse than they let it appear.
Are you going through the typical financial journey of scale-ups?
All successful AI companies start the same way, and many follow the same trajectory: a great idea, initial funding, then a frantic race to develop the product. In this whirlwind of activity, financial infrastructure is slipping down the priority list, which is understandable to some extent.
Here’s what happens in most cases. To start, you opt for standard products that do the trick at first. A professional account with a traditional bank. A tool for spending. Another for payment of suppliers. A third for international transfers and finally, an invoicing tool.
Then the company grew at high speed. You need to pay a development team in Estonia, open a sales office in London and expand your presence in Asia-Pacific via Singapore or Hong Kong. Each market involves a new banking relationship, a new payment service provider and new compliance requirements.
And without even realizing it, you find yourself managing six different platforms, each of which only solves one piece of the puzzle. Initially practical solutions turn into a complex tangle of disparate systems. If this resonates with you, read on.
Why your AI business faces unique pressures
These challenges are particularly felt among hypergrowth AI companies. Global investments in generative AI jumped to $56 billion in 2024according to Tech Crunch, an increase of 92% compared to the 29 billion dollars recorded in 2023.
In France, 32 start-ups have exceeded the threshold of 100 million euros in cumulative fundraising in 2026compared to 24 a year earlier, reveals Tech Insider. According to theThe 2026 mapping of France Digitale67% of French start-ups have raised funds for a total of 16 billion euros.
Unlike traditional SaaS companies that can take years to establish a presence in a single market, AI companies often go global from the start. Customer demand is emerging in several regions at once. Investors are pushing for both rapid development and cash flow control. The result is explosive.
You find yourself caught in a vice. On the one hand, your investors expect you to know your burn rate, your cash flow forecasts and your financial autonomy at all times. On the other hand, there is the reality on the ground. You are recruiting quickly in Singapore, London and San Francisco. These three sites require payments in different currencies. Engineers purchase credits from OpenAI. Your cost structure has nothing to do with that of traditional accounting software.
What once took three to five years is now done in twelve months. A financial infrastructure designed for 10 people and a modest turnover is simply not suited to managing 200 people spread across different continents and financial flows numbering in the millions.
When did spreadsheets become your worst enemies?
The vast majority of early-stage business founders do not have a background in finance. Your mission is to inspire a vision and develop the product. You tinker with systems that just about work.
You end up creating complex search models in Google Sheets to answer simple questions. What is our cash flow level? What is our burn rate? How long will we be financially independent?
Jack Zhang, CEO ofAirwallexsummed up the situation perfectly in a post on LinkedIn. “CFOs continue to assemble dashboards, spreadsheets and other tools to answer questions as simple as ‘Where is my cash flow?’ » or “What is my burn rate?” »
In 2025, it shouldn’t be so complicated to answer. Yet for AI scale-ups juggling entities and currencies, these are truly difficult questions.
4 hidden costs that penalize your growth
The impact of disparate financial systems goes beyond operational issues. When we inherit this amalgam, we find ourselves faced with challenges with a direct impact on the growth of the company.
- You are moving blindly
Without a single source of reference for business spending, it becomes virtually impossible to make informed decisions about resource allocation. AI companies are growing in different different sectors. If we look across the Channel, in the first quarter of 2025, almost half of venture capital awarded to healthcare companies in the UK has gone to AI start-upsaccording to a study by HSBC. Each new segment brings its share of complexity.
- Hidden fees undermine your financial autonomy
You lose thousands of euros because each regional office and each customer base has its own payment solution. I’ve spoken to business founders who found they were paying for duplicate software subscriptions in different offices. No system tracked all expenses. And remember, effective currency management reflects financial maturity, so don’t be that team that bills everyone in headquarters currency.
- Your finance teams are underwater
Your finance teams spend 60% of their time reconciling accounts and avoiding disasters, when they should be focusing on forecasting and optimizing budgets. Some discovered supplier invoices processed twice or important payments delayed.
- You can’t answer important questions
The French market is showing signs of maturity. Nearly half of players are profitable or plan to be profitable within 3 yearswhile the cumulative turnover of the sector exceeded one billion euros in 2025. In this context, investors are choosing companies that demonstrate strong growth while demonstrating control over their finances.
If you’re preparing for your next funding round and you’re not able to quickly answer investors’ questions, you have a serious problem.
What does a future-ready financial infrastructure look like?
According to a study by France Digitale and Sopra Steria Ventures, France recorded 1,114 AI start-ups in 2025, 32 of which have completed funding rounds exceeding 100 million euros.
Success belongs to companies that scale quickly while maintaining control. This requires financial systems that provide real-time visibility, automation of repetitive tasks, and easy scaling as you grow.
Jack’s observation about the evolution of the SaaS model towards that of AI seems relevant here. As he noted:
“SaaS gave us visibility. AI has brought us speed. »
Classic SaaS tools have digitized finance, but they don’t actually manage it. The new generation of tools will be different. Autonomous systems that can handle account reconciliations, forecasting and payments, without constant human intervention.
However, we cannot build autonomous finance from disparate systems.
For AI scale-ups, the question is not whether or not to modernize their financial infrastructure, but when is the best time to do so. Most companies wait until the situation is unsustainable: preparation of a financing round, a total lack of tracking of expenses, finance teams threatening to resign.
At this point, fixing the problem while maintaining activity is a lot like replacing the engine in your car while driving down the highway.
The other solution is to treat financial infrastructure with the same urgency as product development. Ensure robust systems are in place early, before they hamper your growth.
Your competitors are growing rapidly. Don’t make the finance function an obstacle to your growth.
At the house of Airwallexthese conversations with AI scale-ups have shaped the way we think about modern financial infrastructure. We know that solving these challenges requires more than one-off solutions.
Want to continue the discussion? Contact Airwallex by clicking here.
