Where should you incorporate your startup?

1. Do I actually need to incorporate yet?

Incorporation is useful when it gives the work a legal owner. An investor needs a company to fund. A material customer may need a counterparty. Founders may need to assign IP, hire people, open a business account or move liability away from themselves. Each is a real trigger because something important becomes difficult or unsafe without an entity.

A domain name, pitch deck and early customer interviews do not necessarily need a company. Forming too early can lock in founder ownership, tax treatment and a jurisdiction before the business knows what it requires. Waiting preserves choices; waiting after money, valuable IP or consequential contracts arrive can leave the founders personally exposed and the ownership record untidy.

• Incorporate now: an investor is ready, a contract needs signing, people need hiring or material IP needs a clear owner.
• Keep testing: the problem, team or financing route is still changing and no meaningful asset or obligation needs an entity.
• Do not improvise: never accept an investment personally and promise to move it into a company later.

2. Start with where the company will actually operate

Draw the company 24 months forward. Mark where each founder will work, where important decisions will be made, who will create and own the IP, where the first employees will sit, and which entity will sign customers. Add banking, grants, regulated activity and employee-equity plans only where they could change the structure.

Then compare architectures, not formation websites. One local company keeps ownership and operations together. A regional parent can give a distributed team one shared company while local obligations remain where people work. A US parent may give investors a familiar cap table, but it can sit above the same local payroll, tax, employment and accounting work rather than replacing it.

Orrick's guide for German startups makes the trade-off concrete: a US holding structure can improve access to US early-stage capital, but it creates a two-tier structure requiring coordinated corporate, tax and accounting work. The US company is a parent, not an escape from the operating country.

The founders' physical work also matters. The OECD's 2025 update specifically addresses when cross-border remote work can create a taxable presence. Incorporation, corporate residence, payroll and permanent-establishment exposure are related but not interchangeable questions. Map the people and decision-making first, then ask qualified advisers in the affected countries to confirm the result.

• Local: ownership, work and likely early capital cluster in one country.
• Regional: a cross-border team needs one shared parent, but the financing path is not specifically American.
• US: a credible US capital or operating requirement is valuable enough to justify a parent plus local execution.

3. Is this genuinely a US VC-backable business?

“We may raise in the US” is not yet a financing strategy. US venture investors still need a reason to believe this team can build a company large enough to matter to their fund. At the earliest stage, that conviction usually begins with one of two kinds of evidence.

The first is exceptional founder–problem fit: rare technical or domain ability, privileged access to the problem, and an insight that becomes more convincing as the founders pressure-test it. Sequoia describes this as the company's right to exist: a compelling market, a timely wedge and a founder advantage that makes this team unusually suited to capture it. The biography matters only when it changes the odds.

The second is customer evidence. A large enough reachable pool has a burning problem, the current alternative is materially weak and customers will commit money, time or behaviour to a better answer. For most founders, YC's seed guidance expects more than a large TAM: a clear customer and market, a product matching the need, and adoption at an interesting rate. Customer praise without commitment is not the same evidence.

Either route still needs a venture-scale destination, a credible “why now” and a believable way to reach the US market. Incorporating in Delaware can remove legal friction after an investor develops conviction. It cannot manufacture the conviction itself.

• Founder evidence: a non-obvious insight, relevant ability and access that materially improve the chance of solving the problem.
• Customer evidence: a bottom-up pool of reachable buyers with an urgent problem and observable willingness to change.
• Common floor: the wedge can expand into rapid, substantial growth and the team can credibly pursue the market.

The first SAFE chooses the financing company

The company named on a SAFE receives the investment and promises the future shares. Its board approves the instrument; its cap table records the dilution; and its next equity round triggers the conversion. The document is not neutral plumbing around a company choice. It makes the choice real.

YC publishes its standard post-money SAFE forms for US companies. Its current non-US forms cover Canada, Cayman and Singapore, and it tells founders elsewhere to use local counsel. A German, Austrian or UK company should not simply paste its name into the US form.If named investors expect a standard US SAFE and the company passes the US-backability test, a Delaware parent can make the financing mechanics considerably cleaner. If the likely round is European, use an instrument designed and reviewed for that company instead.

If the answer is US, Delaware still leads

Delaware is no longer the only defensible US corporate home. It is still the lowest-friction default for a conventional venture-backed startup because investors, counsel and transaction processes know what they are looking at. Cooley continues to treat it as the established starting point, even as founders reconsider the automatic choice.

Delaware's own formation guidance is unusually direct: the reason to choose the state is the predictability and efficiency of its law, not low cost. A Delaware corporation still needs a registered agent, annual reporting and franchise-tax compliance, and it may need to qualify wherever the business actually operates. Familiarity removes financing friction; it does not remove the operating stack.

Nevada offers a more protective governance posture in some cases, but that benefit must be intentional, counselled and acceptable to likely investors. A founder expecting a standard institutional round should not create a new diligence question merely to appear contrarian. Delaware remains the practical default; Nevada is a specific governance decision.

Apply the framework to the three-founder case

The company needs to exist because investors are ready to receive SAFEs. Next, map the operating plan: where the founders will work, where the IP will be created, which country will employ the first hires and whether any founder should hold shares through a local personal company. Those decisions should be reviewed before stock and IP are issued, not repaired afterwards.

If the next realistic investors are US or international funds that expect standard US SAFEs, and the team has exceptional founder fit or strong evidence of a venture-scale customer problem, a Delaware parent is a coherent answer. Establish the local operating layer deliberately rather than pretending the US certificate removes it.

If the reachable investors, customers and team remain primarily European, a local or regional parent with locally reviewed financing documents may be cleaner. If the investor path is still hypothetical, speak to the ten most plausible investors before paying indefinitely for US optionality. A jurisdiction should remove a constraint the company actually has.