Brazil

High-growth firms sit on top of a huge population of small companies in a tax regime built for them. Most of those small companies are not startups. The 2021 law tries to mark the difference. Daily life still marks it by whether anyone outside the family has put in risk capital.

The sociedade limitada is the ordinary private company. A sociedade anônima is the corporations-law vehicle the startup statute touched. Choice of vehicle is a lawyer’s work because liability, investment instruments, and tax do not line up in one form. GRIP will not pick the form.

Local venture funds, corporate investors, and development institutions (BNDES and FINEP among the public ones) all exist. They are not substitutes. Foreign funds add an exchange registration and a tax question. The angel-investor provisions in the 2021 law are a statutory category, not evidence that angels are the main source of scale-up capital.

The strongest technical base in the startup system is software for a domestic service economy: finance, retail, logistics, health administration. Deep hardware is narrower and more tied to public research institutions than to the Paulista avenue office market.

Portuguese is the operating language. English is an overlay for fundraising and for some engineering teams, not for customers. Senior product and security talent is scarce relative to graduate volume. Services and bank technology departments are the training ground and the competing employer.

The University of São Paulo is the main research university people mean when they point at the city. Unicamp is in Campinas, in the same state and not in the same city. Federal universities elsewhere, and public research institutes, matter for science-based firms. A campus tour is not a map of customers.

The startup law sits beside the Central Bank for anything that touches payments, the securities regulator for anything that looks like a public offer, and the tax authorities at three levels. A sandbox, where a regulator offers one, is a supervised exception. It is not a general permission to launch.

The cost that does not show in rent is tax compliance and the price of employment. Labour charges and litigation risk are part of why companies stay small on the formal books. Doing Business ranks are not used. They were discontinued, and Brazil was one of the countries the method was bad at describing even before that.

Domestic aviation connects the major cities. Urban transport inside São Paulo is a constraint on where staff will work. Logistics across a continental country, with uneven roads and a serious port system on the coast, is the physical problem for any company that is not pure software.

The domestic market is the point: a large, unequal, Portuguese-speaking consumer and enterprise economy. Mercosur is a trade arrangement, not a single market a startup can treat as home. Selling in Spanish-speaking neighbours is a second company problem, not a feature of São Paulo.

Innovation that clears the market is often a way of distributing a service Brazil’s incumbents did badly: payments, retail logistics, credit to a customer the bank would not underwrite. Scientific innovation exists and has a longer path through public funding and regulation. Both are real. They belong in separate notes.

Scale means surviving formal employment, tax, and a customer who pays late, long enough to matter outside one city. The public market is B3, in São Paulo. Listing is rare relative to the number of companies that call themselves startups. A cross-border exit has to pass Brazilian exchange and tax rules on the way out.