Brazil in the mid-1980s tried to grow its own computer industry behind a wall of import bans — and discovered that a wall protects a market, but does not provide technology. The story of Unitron Eletrônica and its Macintosh clone is the perfect lens through which to see how every side breaks against reality: those who copy, those being copied, and those who write the rules of the game.
In 1984, Brazil passed a law establishing the Reserva de Mercado regime — a reservation of the domestic microcomputer market for national manufacturers. Imports of finished machines were effectively banned. The logic was as straightforward as a transatlantic cable: keep IBM, Apple, and Compaq out of the Brazilian market, and local companies would have room to build muscle, master the technology, and one day go international. Política Nacional de Informática — that was the name of this state strategy — drew on the experience of postwar Japan and South Korea, where protectionism really had helped grow electronics giants. The difference lay in the scale of the economy, the depth of the engineering tradition, and — crucially — access to core components. Brazil did not make its own processors. It did not manufacture disk drives. It depended on imported components even for the machines it assembled at home. The reservation regime created a greenhouse, but the seeds still had to be imported.
The regime was overseen by the Secretaria Especial de Informática (SEI) — a special agency that decided which projects received approval and which did not. SEI was regulator, arbiter, and, in a sense, gardener all at once: it determined exactly what would grow behind the wall. Every new computer project by a Brazilian company required its blessing. It created a strange situation: the market was shielded from foreign competition, but domestic competition depended on a bureaucratic decision. A company could develop a brilliant machine — and run into a rejection. Or it could get approval for a mediocre product because it fit the political climate better. Brazilian engineers worked in conditions where the technical problem was woven so tightly into the political one that separating them was impossible.
Unitron Eletrônica was founded on April 23, 1974, in São Paulo — ten years before Reserva de Mercado became law. By the mid-1980s, it was an experienced company, not some garage startup. And when its engineers decided to create a Macintosh clone, they understood the scale of the task. The original Mac 512K, released by Apple in September 1984, was a machine with character: closed architecture, proprietary chips, a unique graphical interface. Copying it was nothing like assembling yet another IBM PC-compatible computer from standard components.
At first, Unitron tried to take the legal route. The plan was to license the technology from Apple — buy manufacturing rights, pay royalties, operate within the law. But Brazil’s market-reservation legislation required that at least 51% of any joint venture belong to a national company. Apple would not agree. For Cupertino, giving a local partner a controlling stake meant surrendering control of its own technology in a market where the company could not sell directly anyway. The deadlock was symmetrical: Unitron could not legally acquire the technology; Apple could not legally sell in Brazil. Both sides were up against the same wall — just from opposite sides.
So Unitron turned to reverse engineering. The company received a $10 million loan from a state bank — a sum that shows the project was viewed not as a gamble, but as a strategic investment. CTI (Centro de Tecnologia da Informação) — Brazil’s information technology center — joined the effort. And to dismantle and reproduce Apple’s proprietary chips, they brought in National Semiconductor of Santa Clara, California. State-level irony: an American corporation, using American money, was helping Brazilians reproduce an American computer while bypassing American pressure.
What fascinates me is the engineering side of this story — the way Brazilian specialists took the Macintosh apart molecule by molecule and put it back together. The task broke down into several layers, each requiring its own approach.
The Motorola 68000 processor at 8 MHz was a standard component, available on the open market. The 512 KB of RAM posed no problem either: DRAM chips were made by dozens of factories around the world. The 3.5-inch, 800 KB disk drive came from Japan’s Chinon. The monochrome 9-inch screen, with a resolution of 512×342 pixels, reproduced the very image seen by users of the original Mac. These were the machine’s “bones,” and obtaining them was a matter of logistics, not engineering.
The real difficulty began with the proprietary elements. Apple used several custom ASIC chips in the Macintosh — the disk-drive controller, the real-time clock, programmable logic arrays (PAL chips). These components could not be bought: they had been designed specifically for the Mac and were not sold separately. This was where National Semiconductor’s help came in handy: its specialists reverse-engineered the chips, reconstructing their logic and creating functional equivalents. The process resembled the work of a restorer reconstructing an entire composition from a fragment of fresco — except that afterward, the fresco had to be made to work.
The most ambitious element was the 128 KB ROM. In the original Macintosh, the ROM contained key parts of the operating system, low-level hardware routines, and the famous Toolbox — the set of functions that made a Mac a Mac. Unitron insisted it had not copied the ROM contents bit for bit: according to Rainer Brockerhoff, a consultant to the team, the firmware was written in C and assembly — functionally equivalent code created from scratch using the Inside Macintosh specifications. Apple would later claim the exact opposite: after obtaining samples and examining them in Cupertino, it declared the ROM an almost exact copy of its own — and this dispute would become the main weapon in the campaign against the clone. Legally, the matter would have been simple: bit-for-bit copying of a ROM is an indisputable copyright violation, while reproducing functionality through original code is a gray area. But politics would intervene, and the courts would not be the ones deciding.
On September 23, 1985, at the 5th International Informatics Exhibition, Unitron unveiled its machine to the public. The Unitron Mac 512 became the world’s first Macintosh clone. Not the first personal-computer clone of any kind — by then, the IBM PC had already acquired dozens of copies, from Compaq to Taiwanese no-name manufacturers. But a clone specifically of the Mac, the machine Apple had built as an impregnable fortress of closed architecture. For the Brazilian public, cut off from the original by the import ban, it was an event. For Apple, a red rag.
Cupertino’s response was predictable and harsh. Apple alleged an infringement of intellectual property. But the company’s legal position in Brazil proved weaker than one might have expected. In 1987, two Brazilian universities — UFRJ (the Federal University of Rio de Janeiro) and UFRGS (the Federal University of Rio Grande do Sul) — conducted an expert review at SEI’s request. The conclusion of both reports — dated May 1987 — was encouraging for Unitron: the machine had indeed been reverse-engineered rather than copied, and Apple, as SEI’s files revealed, had not registered a patent for the Macintosh in Brazil. The company that in the United States fiercely defended every element of its design had simply failed to secure patent protection in the Brazilian market, where it could not sell its products anyway. It is like installing an armored door on your apartment, but forgetting to close the window — because you are sure no one will come in from that side.
But Apple had a weapon that required no Brazilian patents — the government of the United States. By the mid-1980s, trade relations between Brazil and the United States were already tense. Reserva de Mercado irritated Washington: American technology companies were losing access to Latin America’s largest market. The Unitron case became the perfect argument for those in the American administration demanding tougher measures.
The United States moved from diplomatic protests to direct threats. At stake were Brazilian exports of shoes and orange juice — goods with not the slightest connection to computers, but critically important to the Brazilian economy. The logic was crude: if Brazil did not protect the intellectual property of American companies, American tariffs would hit Brazilian farmers and shoemakers. Oranges versus microchips — trade wars are rarely elegant.
Under this pressure, in December 1987, Brazil adopted the Lei de Software — Law No. 7646, establishing legal protection for software. The law was not a direct capitulation: Brazil itself understood the need for regulation in this area. But the chronology speaks for itself — the law appeared amid specific trade threats. In March 1988, SEI rejected the Unitron Mac 512 project. Formally, under the new legal framework. In practice, under pressure coming at once from outside and above.
Unitron did not surrender. On March 29, 1988, the company filed an application for a new project — the Unitron 1024, an improved version of the clone. The application reached the highest authority — the CONIN council (Conselho Nacional de Informática). The vote, held on December 19, 1988, deserves a frame of its own: 8 votes against, 7 in favor. All seven independent representatives on the council voted for Unitron. Every vote against came from government ministers. The Minister of Aeronautics abstained. The deciding vote that buried the project came from the Minister of Science and Technology. The vote looks like an X-ray of Brazilian politics at that moment: independent experts considered the project viable and legal, but the political leadership, caught between its own industrial strategy and international pressure, chose to kill it. Unitron shut down after that.
The arithmetic of that vote gets me. One vote. The company’s entire history, the ten-million-dollar state loan, years of engineering work, a ROM written from scratch — all of it came down to a one-vote margin cast by an official balancing not between “legal” and “illegal,” but between Washington and Brasília.
The Reserva de Mercado regime outlived Unitron by only a short while. In 1990, Fernando Collor de Mello became president of Brazil, and his administration set a course toward economic liberalization. Import barriers on computers began to come down. By the early 1990s, the Brazilian market had opened to foreign manufacturers, and Apple could finally sell its products in the country — the very country where, just a few years earlier, its technology had been reproduced, dismantled, declared legal by a university review, and banned anyway.
The results of the reservation regime are judged differently. Brazil did indeed grow several computer companies — but most of them made IBM PC clones, not original designs. During the years of isolation, the technological gap with global leaders did not narrow. It widened. Brazilian users paid more for local machines than originals cost on the world market, and received products that lagged a generation or two behind. The greenhouse grew plants — just not the ones people had counted on.
For Apple, the Unitron story remained an unpleasant episode in which the company was both victim and beneficiary. A victim — because its technology was copied. A beneficiary — because the scandal became leverage against the Brazilian government and ultimately helped open the market. For the Brazilian engineers who wrote their own ROM in C and assembly, dismantled ASIC chips, and assembled a working machine, it was a display of competence that never got the chance to become a product.
It is easy to tell Unitron’s story as an anecdote about Brazilian pirates who copied the Macintosh and got caught. That version is convenient, compact, and wrong. Unitron did not operate in a legal vacuum: the company worked within a system created by the state, received state funding, passed a university review, and was found lawful by every domestic criterion. It was destroyed not by justice, but by a political decision made under external pressure, by a margin of one vote.
It is just as easy to portray Brazil’s Reserva de Mercado as a regime that “allowed piracy.” That too is a distortion. The regime gave local companies a protected market — but it did not abolish copyright or guarantee that reverse engineering would be approved. SEI could reject a project, and did. The software-protection law was passed before the reservation regime formally ended. The system tried to be both protectionist and lawful — and discovered that those two goals clash when the technologies involved were created beyond the wall.
Jobs’s line about pirates and the navy was a motivational slogan for a team working inside Apple. He meant audacity, a refusal to submit to corporate bureaucracy, a willingness to break internal rules for the sake of results. The Unitron engineers, by a sad irony, became pirates in the literal sense — that is what they were called, though they worked by the rules of their own country. The difference between a pirate and a privateer has always been defined by the flag on the mast, not by what happens on deck.
The company did not die at once: after the ban, Unitron cut its workforce and scraped by on side products — like a stage controller built on an Apple II clone — while still repaying the loan for a machine that was never destined to become merchandise. Several Unitron Mac 512 units survive in private collections and museums. Brazil’s IT industry took another path — through market opening, assembly plants, and, later, software companies.