EvergreenKenya loves talking about startups. We celebrate the Silicon Savannah, fill conference halls with panels on digital transformation, and share the news every time a local company raises a few million dollars. Suddenly, everyone is proud of the ecosystem.
There is plenty to be proud of. In 2025, Kenya was the largest destination for tech funding in Africa, attracting $1.04 billion in combined equity and debt, according to Partech's annual report.
But there is another Kenya that rarely makes those headlines. It is the founder with three employees, the developer building something after work, the two friends trying to turn a prototype into a business. It is the company with ten customers, almost no revenue and six months of runway.
For them, Kenya can be a hostile place to build. A large part of the reason is that our systems struggle to tell the difference between a company still trying to find out whether it has a business and one that has already made it.
The cost of being small
Starting a company is supposed to be the cheap part. You have an idea, you build something, and you put it in front of customers. It probably doesn't work, so you change it and try again. If you're lucky, you eventually find something people actually want. That loop is what a startup is.
Our systems, however, tend to expect structure, compliance and administrative maturity long before the business has any of those things.
Registering a private limited company through the Business Registration Service costs KSh10,650. On its own, that isn't outrageous. But incorporation is only the front door. Behind it, depending on what you're building, sit tax registrations, electronic invoicing, county permits, sector licences, payroll obligations, bookkeeping and data-protection requirements.
Many of these are reasonable one at a time. Together, they produce something every founder recognises: friction. And friction costs most when you are smallest.
A large company has accountants, lawyers, compliance officers and an HR team. A startup has Kevin. Kevin is the CEO. He also runs sales, fixes the website, answers support tickets and is trying to persuade an investor to fund another six months. Now Kevin is learning tax law, too.
None of this is an argument for letting startups operate outside the law. It is an argument that regulation has a cost, and that cost does not fall on every company equally.
When compliance becomes a product requirement
Take something as ordinary as invoicing. KRA's eTIMS system requires businesses to issue invoices through its electronic framework. There are accommodations for small businesses, and KRA offers its own software free of charge, but it is still one more system a founder has to learn and build into daily operations.
Or take data. Almost every technology company handles personal information, and Kenya's data protection law rightly sets rules for how it is collected and used. Some small organisations below certain revenue and staff thresholds are exempt from mandatory registration, although businesses in specified sectors must register regardless of size.
Each rule may be perfectly defensible. The trouble starts when they stack. One requirement becomes five, five become fifteen, and somewhere along the way the founder stops building the product customers need and starts building the bureaucracy required for permission to build it.
We regulate the company we hope it will become
The deeper problem is philosophical. We tend to regulate startups for the risks they might one day create, not the risks they create today.
A startup with 200 users is not Safaricom. A fintech moving KSh100,000 a month is not a bank moving billions. A four-person SaaS company is not Microsoft. Yet our regulatory instinct often starts with the mature version of a business, imagines what could go wrong at scale, and builds obligations around that possibility.
But most startups never get big. Most experiments fail, and that is not a flaw in entrepreneurship; it is how entrepreneurship works. Ten people try something. Eight fail, one builds a decent small business, and one, perhaps, builds something enormous.
That arithmetic only works if trying is cheap enough for all ten to have a go. Make experimentation expensive, complicated or legally intimidating, and fewer people try. Then we act surprised when fewer companies come out the other end.
Capital doesn't fix this
The headline numbers hide a contradiction. Kenya attracts large amounts of startup capital, but very little of it reaches the average founder.
In 2025, four companies, all in clean energy, took roughly 70 percent of the venture capital AVCA tracked in Kenya. A country can attract hundreds of millions of dollars in startup investment while many of its founders struggle to raise their first KSh5 million. Both are true at the same time.
And money has become harder to find. In the first half of 2026, Kenyan startups had their weakest fundraising stretch since 2021, according to Africa: The Big Deal, as investors grew more selective and backed fewer companies.
When capital is scarce, the cost of simply existing as an early-stage company matters more, not less.
The startup tax nobody measures
There is one more tax that never appears in a Finance Act: time.
Every hour spent wrestling with a government portal is an hour not spent with customers. Every day spent working out which licence applies is a day not spent on the product. Every shilling paid to a consultant whose job is to interpret bureaucracy is a shilling not spent on an engineer, a salesperson or a designer. Every approval that takes weeks delays the moment a company learns whether its idea works.
Unlike corporation tax, you pay this one whether or not you're profitable. Often you pay it before you've earned your first shilling.
Then there's the market
Regulation is only part of it. Getting customers can be just as hard.
Large companies say they want to work with startups. Government says it supports local innovation. Then procurement begins, and suddenly you need years of audited accounts, previous contracts of a similar size, certifications, proof of financial capacity, and references showing you've already delivered exactly the kind of project you're asking for your first chance to deliver.
It is the old job-seeker's trap: you need experience to get the job, and the job to get experience.
The outcome is predictable. Contracts go to established companies because the requirements were written around established companies. The startup is told to come back once it has grown, when the contract itself might have been what helped it grow.
Failure still costs too much
Healthy ecosystems don't just make it easy to start a company. They make it relatively painless to fail. Failure isn't an unfortunate side effect of innovation; it is part of the machinery. You cannot have experiments without failed experiments.
If closing a business is complicated, if founders stay tangled in compliance obligations long after the company has died, or if failure carries heavy financial and social penalties, people become cautious. They pick ideas that already look safe. They copy what has worked elsewhere. They avoid ambitious bets, which is exactly the opposite of what an innovation economy needs.
We need regulation, just smarter regulation
None of this means startups should do whatever they like. A fintech holding people's money needs oversight. A health startup handling medical records needs safeguards. Any company with employees has obligations to them, and any platform storing personal data owes its users protection. Regulation exists for good reasons.
But good regulation is proportionate. It should ask not only what kind of company you are, but how big you are, how much risk you actually create, how many customers you serve, how much money passes through your hands and how mature the business is.
A company with three employees and KSh2 million in annual revenue should not carry the same practical compliance load as one with 3,000 employees and KSh20 billion. The rules should grow with the company.
Give startups room to become companies
Imagine a different model. For their first few years, genuinely small companies could operate within clearly defined regulatory sandboxes. The basics would stay non-negotiable: don't misuse people's money, protect customer data, pay employees properly, keep records and pay the taxes that apply.
Beyond that, obligations would switch on as the company grows. Cross a revenue threshold, and new requirements apply. Start holding significant customer funds, and financial safeguards tighten. Hire past a certain headcount, and further employment rules kick in. Process sensitive data at scale, and stronger compliance follows.
The principle is simple: regulation should scale with risk and maturity.
We already accept this elsewhere. Nobody builds an estate road to the same specification as a six-lane highway. Both need standards; the standards just reflect different levels of risk. Businesses should be no different.
Government should build APIs, not queues
There is also a big opportunity in how businesses deal with the state. A tech company shouldn't need staff typing the same company details into several government systems.
Company registration, tax, social contributions, permits and regulatory filings should work more like infrastructure: one business identity, standardised data, systems that talk to each other, APIs where they make sense and automated reporting wherever possible. Tell the government something once and, with proper privacy controls, it should already know it.
The best regulation isn't something founders learn to tolerate. It's something they barely notice when they're doing things right.
Kenya already has the hard part
That is what makes this so frustrating. Kenya already has the ingredients other countries spend decades trying to build: technical talent, entrepreneurs, a sophisticated mobile-money culture, a large young population, investors, regional influence, and people willing to take absurd risks to build things. Despite everything, the country keeps producing interesting companies.
Imagine what would happen if we made it dramatically easier to experiment. Not easier to commit fraud, dodge taxes or mistreat employees and customers. Easier to try. That difference matters.
We celebrate the forest and punish the seedling
We want billion-dollar Kenyan companies. We want Kenyan technology exported around the world. We want jobs, innovation, locally built solutions, and young people creating work instead of waiting for it.
Yet our systems often behave as if every new business were already a large, established organisation. That is backwards.
Before a company employs a thousand people, someone has to hire the first. Before it pays millions in tax, it has to make its first sale. Before it raises $50 million, someone has to invest $5,000. Before there is a headquarters, there is usually a laptop. And before there is a successful company, there is an experiment, and experiments need room to fail.
Kenya doesn't need fewer rules. It needs rules that understand what they are regulating. Protect consumers firmly where the risk is real. Regulate mature companies accordingly. Punish fraud and enforce standards. But give young companies room to breathe, experiment, fail, change course and try again.
If every seedling must meet the rules written for a forest before it is allowed to grow, we shouldn't be surprised when so few forests appear.
