{"id":718,"date":"2026-09-06T15:53:12","date_gmt":"2026-09-06T18:53:12","guid":{"rendered":"https:\/\/www.linuxpro.com.br\/2026\/09\/a-historia-da-intel\/"},"modified":"2026-09-06T16:00:38","modified_gmt":"2026-09-06T19:00:38","slug":"a-historia-da-intel","status":"publish","type":"post","link":"https:\/\/www.linuxpro.com.br\/en\/2026\/09\/a-historia-da-intel\/","title":{"rendered":"The history of Intel"},"content":{"rendered":"

\"A<\/p>\n

Intel invented the memory market, then abandoned it; invented the microprocessor almost by accident; put \u201cIntel Inside\u201d in every office on the planet; and then lost \u2014 in sequence \u2014 mobile phones, Apple, and the manufacturing leadership that defined it. Few companies explain as much of computing history. This is Intel's trajectory, from the \u201ctraitorous eight\u201d to the existential crisis of the decade beginning in 2020, with dates on the table.<\/p>\n

1968: the founders who were already legends<\/h2>\n

On 18 July 1968<\/strong>, Robert Noyce<\/strong> (co-inventor of the integrated circuit) and Gordon Moore<\/strong> (author of Moore's Law, formulated in 1965) left Fairchild \u2014 the company they themselves had created after leaving William Shockley in 1957, in the episode of the \u201ctraitorous eight\u201d<\/em> \u2014 to found the company that weeks later would adopt the name Int<\/strong>egrated El<\/strong>ectronics \u2014 Intel (initially incorporated as NM Electronics). Employee number 3 was a Hungarian immigrant named Andy Grove<\/strong>, who would become Silicon Valley's most influential manager.<\/p>\n

1971: the microprocessor is born from a calculator<\/h2>\n

Intel started by selling memory (the SRAM 3101 in 1969, the DRAM 1103 in 1970). The product that would change the world came from an order: Japan's Busicom requested 12 chips for a calculator, and Ted Hoff, Federico Faggin and Stan Mazor<\/strong> condensed them into one. In November 1971<\/strong> came the Intel 4004<\/strong> \u2014 the first commercial microprocessor: 2.300 transistors, 4 bits, 740 kHz. Intel had the good sense to buy back the rights from Busicom.<\/p>\n

1978\u20131981: x86 and the big break of the IBM PC<\/h2>\n

The 8086 (1978)<\/strong> launched the x86 architecture still running in your server today; the budget variant 8088<\/strong> was chosen by IBM for the IBM PC of 1981<\/strong> \u2014 the winning lottery ticket of computing history. With IBM requiring a second source, Intel licensed x86 to AMD in 1982, planting the half-century rivalry we tell in the history of AMD<\/a>.<\/p>\n

1985: Grove's decision<\/h2>\n

Crushed by Japanese memory manufacturers, Intel faced its first existential crisis. The scene is a classic: Grove asked Moore \u201cif we were fired and the board brought in a new CEO, what would he do?\u201d<\/em>. Answer: get out of memory. That's what they did \u2014 Intel abandoned the DRAM it had invented<\/strong> and bet everything on processors. The 386 (1985)<\/strong>, with 32 bits, and the 486 (1989)<\/strong> sealed the decision as the best in the company's history.<\/p>\n

1991\u20132005: Pentium, the US$ 475 million bug and the stumbles<\/h2>\n

The Pentium (1993)<\/strong> and \u201cIntel Inside\u201d marketing (1991) turned an invisible component into a consumer brand. It wasn't all glory: the FDIV bug (1994)<\/strong> \u2014 incorrect floating-point divisions \u2014 cost a recall of US$ 475 million<\/strong> and a lesson in public relations. The decade of 2000 brought architectural stumbles: the Pentium 4\/NetBurst<\/strong> chasing GHz at any cost (and heating up like a stove), and the Itanium<\/strong> \u2014 the 64-bit bet with HP that ignored x86 compatibility and was overtaken by AMD64. Recovery came from the team in Israel with Centrino (2003)<\/strong> and the Core (2006)<\/strong> architecture, which regained the performance crown for a decade.<\/p>\n

2007\u20132020: the three missed opportunities<\/h2>\n

The decline came not from one blow, but from three bets declined: Intel refused to make the iPhone chip<\/strong> (2007, because of low margins \u2014 Otellini would publicly admit the mistake); lost the decade of 2010 struggling with the transition to 10nm<\/strong> while TSMC raced ahead; and in 2020<\/strong> saw Apple replace its processors with its own M1<\/strong> . Along the way, Meltdown and Spectre (disclosed in 2018)<\/strong> exposed speculative execution vulnerabilities that hit its chips hard \u2014 and led to years of mitigations in the Linux kernel.<\/p>\n

2021\u20132026: the rebuild<\/h2>\n

In 2021<\/strong>, veteran Pat Gelsinger<\/strong> returned as CEO with the IDM 2.0<\/strong>plan: become a foundry for third parties and regain process leadership with High-NA EUV lithography. Execution was costly \u2014 historic losses, mass layoffs \u2014 and in December 2024<\/strong> Gelsinger left. Under Lip-Bu Tan<\/strong>, with a multibillion-dollar investment from the US government (the CHIPS Act becoming an equity stake in 2025), Intel fights to prove it is still strategic: Xeons remain in half of datacenters, the 18A\/14A processes are its bet on returning to the top, and x86 \u2014 now in an unprecedented collaboration alliance with AMD \u2014 defends territory against ARM.<\/p>\n

Intel and Linux: the ever-reliable ally<\/h2>\n

In our backyard, Intel has always been among the best citizens: historically one of the largest corporate contributors to the kernel<\/strong>, with open network, graphics and storage drivers decades before that became fashionable \u2014 from e1000<\/code> to i915<\/code>, Linux administrators know its modules by heart. Meltdown\/Spectre also showed the other side of the coin: when silicon fails, the kernel pays the cost of mitigations.<\/p>\n

Timeline<\/h2>\n