From the 4004 microprocessor in 1971 to the x86 architecture that runs the world — and Intel's existential battle to reclaim manufacturing leadership in the AI era.
All
Person
Event
Conflicts
Disaster / crisis
Science
Culture
Companies
Countries
Dashed left edge = started before this timeline ·
Dashed right edge = still ongoing
Intel 46
1968Gordon Moore's Law — the prediction that built an industryIn 1965 — three years before founding Intel — Gordon Moore published a paper in Electronics Magazine observing that the number of transistors on an integrated circuit had doubled every year since 1959, and predicting this would continue. Later revised to a doubling every two years, 'Moore's Law' becomes the most famous and consequential prediction in technology history. The semiconductor industry organises itself around Moore's Law as a roadmap: engineers, equipment makers and chip designers all target the same cadence. Intel's entire business model is built on delivering Moore's Law.
1968Intel founded — Moore and Noyce leave FairchildGordon Moore and Robert Noyce found Intel Corporation on 18 July 1968 in Mountain View, California, with $2.5 million from venture capitalist Arthur Rock — one of the first significant venture capital investments in Silicon Valley history. Moore and Noyce had both been among the 'Traitorous Eight' who left Shockley Semiconductor in 1957 to found Fairchild Semiconductor — the company that invented the integrated circuit and seeded Silicon Valley. At Intel, Noyce is the visionary; Moore is the technologist; and their third hire, Andy Grove, will prove the greatest manager.
1969Intel 3101 — first product, first SRAM chipIntel ships its first product — the 3101 Schottky bipolar 64-bit static random-access memory (SRAM) chip — in April 1969. The 3101 is faster than competitor offerings and immediately wins customers. Intel's second product, the 1101 MOS SRAM, follows later in 1969. These early memory chips establish Intel's engineering reputation and generate the revenue that funds its subsequent work on DRAM and, ultimately, the microprocessor. Intel at this point is a memory company, not a processor company.
1970Intel 1103 — DRAM replaces core memoryIntel introduces the 1103 dynamic random-access memory (DRAM) chip in October 1970 — the first commercially available DRAM chip and the product that makes Intel's fortune in its early years. The 1103 stores 1,024 bits of data and is cheaper and denser than the magnetic core memory it replaces. By 1972 it is the best-selling semiconductor component in the world. Intel's DRAM business funds the research that will produce the microprocessor — the product that will eventually render DRAM a commodity and force Intel out of the memory business entirely.
1971Intel 4004 — the first microprocessorIntel introduces the 4004 on 15 November 1971 — the world's first commercially available microprocessor, a complete central processing unit on a single chip. The 4004 is designed by Federico Faggin, Ted Hoff and Stan Mazor for a Japanese calculator company (Busicom), which requests a custom chip set. Intel's Ted Hoff proposes a general-purpose programmable processor instead of a fixed-function chip set. Intel buys back the rights from Busicom for $60,000. The 4004 contains 2,300 transistors and runs at 740kHz. The entire computing revolution flows from this chip.
1972Intel 8008 — first 8-bit microprocessorIntel introduces the 8008 in April 1972 — the world's first 8-bit microprocessor, capable of addressing 16 kilobytes of memory. The 8008 is designed for a computer terminal (the Datapoint 2200) but becomes the basis for the first hobbyist personal computers. The 8008 contains 3,500 transistors and runs at 500kHz–800kHz. Its successor, the 8080 (1974), is the chip that the Altair 8800 — the first personal computer kit — is based on. Every personal computer of the 1970s and 1980s traces its lineage to Intel's 8-bit processors.
1974Intel 8080 — the Altair chipIntel introduces the 8080 microprocessor in April 1974 — a major architectural advance over the 8008, with a larger instruction set, more registers and the ability to address 64KB of memory. The 8080 is the processor in the MITS Altair 8800 (1975) — the first personal computer kit, which spawns the hobbyist computing movement. Bill Gates and Paul Allen write the first Altair BASIC interpreter for the 8080, founding Microsoft. The 8080 contains 6,000 transistors and establishes Intel's dominance of the nascent personal computer market.
1975Andy Grove — the manager who made IntelAndrew 'Andy' Grove, Intel's third employee, rises to become President in 1979 and CEO in 1987. Grove — a Hungarian-born Holocaust survivor who came to America with nothing — transforms Intel from an engineering-driven startup into one of the world's most efficiently managed companies. His management philosophy — 'Only the Paranoid Survive,' the title of his 1996 book — emphasises confronting business threats directly rather than hoping they go away. Grove makes two decisions that save Intel: exiting DRAM in 1985 and committing entirely to microprocessors.
1976Intel 8085 and Zilog Z80 — competition emergesIntel introduces the 8085 in 1976 — an improved 8080 requiring fewer support chips. But the more significant event is Zilog's introduction of the Z80 — designed by former Intel engineer Federico Faggin — which is more powerful than the 8085 and becomes the dominant processor in home computers of the late 1970s (Sinclair ZX Spectrum, TRS-80). Intel's response to the Z80 competition is the 8086 — a 16-bit processor that will define PC computing for the next four decades.
1978Intel 8086 — the x86 architecture bornIntel introduces the 8086 on 8 June 1978 — a 16-bit microprocessor that introduces the x86 instruction set architecture. The x86 architecture — with its registers, addressing modes and instruction encoding — becomes the most enduring computer architecture in history. Every Intel Core processor sold today, every AMD Ryzen, every server in the world's data centres runs x86 code. The 8086 contains 29,000 transistors and runs at 5–10MHz. Its cheaper sibling, the 8088 (1979), is chosen by IBM for the IBM PC — cementing x86's dominance.
1979Intel 8088 — chosen for the IBM PCIntel introduces the 8088 — a cost-reduced 8086 with an 8-bit external data bus — in July 1979. IBM chooses the 8088 for the IBM PC in 1981, primarily because Intel can deliver it faster than Motorola's competing 68000. The decision to use the 8088 for the IBM PC, and IBM's open architecture, creates the x86 PC standard that Intel will dominate for decades. Had IBM chosen Motorola's 68000 — used in the Apple Macintosh — computing history would be fundamentally different.
1981IBM PC launches on Intel 8088IBM launches the IBM PC on 12 August 1981, powered by Intel's 8088 processor running at 4.77MHz. The IBM PC's open architecture — using standard components including Intel chips and Microsoft DOS — allows clone manufacturers to build compatible machines, creating an enormous market for Intel processors. Every IBM PC clone that follows must use an Intel-compatible processor (or an exact clone of one). The IBM PC launch is the most important moment in Intel's history — it creates the x86 PC ecosystem that sustains Intel for four decades.
1982Intel 286 — protected mode computingIntel introduces the 80286 (commonly called the 286) in February 1982. The 286 introduces protected mode — a hardware mechanism that prevents programs from interfering with each other's memory, the foundation of modern multitasking operating systems. The 286 is used in the IBM PC/AT (1984) and runs at up to 12MHz with 134,000 transistors. It is the first Intel processor that Microsoft's DOS cannot fully exploit — the tension between Intel's advancing hardware and Microsoft's ageing software architecture will define PC computing for a decade.
1984Intel exits DRAM — the strategic pivotBy 1984, Japanese manufacturers — Fujitsu, NEC, Hitachi — have devastated Intel's DRAM business through aggressive pricing. Intel's DRAM operations are losing money. Andy Grove and Gordon Moore face the defining decision of Intel's history. Grove's famous question: 'If we got kicked out and the board brought in a new CEO, what do you think he would do?' Moore: 'He would get us out of memories.' Grove: 'Why shouldn't you and I walk out the door, come back and do it ourselves?' Intel exits DRAM in 1985 and focuses entirely on microprocessors.
1985Intel 386 — 32-bit computing arrivesIntel introduces the 80386 (386) in October 1985 — Intel's first 32-bit processor, with 275,000 transistors running at 16–33MHz. The 386 introduces the 32-bit protected mode that all modern software is built on. Crucially, Intel manufactures the 386 itself without licensing it to AMD — breaking the cross-licensing agreement that had allowed AMD to make Intel-compatible chips. The 386 establishes Intel's monopoly on the leading edge of x86 processor technology. IBM refuses to use it, briefly preferring Motorola's competing designs — a decision that helps Compaq overtake IBM.
1989Intel 486 — the PC goes mainstreamIntel introduces the 80486 (486) in April 1989 — containing 1.2 million transistors and running at 25–100MHz with an integrated floating-point unit and cache. The 486 is the processor that makes PCs genuinely useful for everyday applications — fast enough for Windows, word processing, spreadsheets and early graphics. Intel's 'Red X' advertising campaign, marking out 386 computers as outdated, is one of the most effective comparative advertising campaigns in technology history. The 486 era sees Intel's revenues grow from $3 billion to $8 billion.
1991"Intel Inside" — ingredient branding inventedIntel launches the 'Intel Inside' cooperative advertising programme in 1991 — offering PC makers financial rebates if they display the Intel Inside logo on their computers and in their advertising. The campaign is one of the most successful marketing innovations in technology history: it makes Intel's brand visible to consumers who had never thought about what processor was inside their PC. The Intel jingle — five notes — becomes one of the most recognised audio logos in the world. 'Intel Inside' transforms Intel from a B2B component supplier into a consumer brand.
1993Pentium — Intel renames its processorsIntel introduces the Pentium processor in March 1993 — the successor to the 486, with 3.1 million transistors running at 60–233MHz. Intel names it 'Pentium' rather than '586' because AMD's legal victories establish that numbers cannot be trademarked. The Pentium brand becomes one of the most recognised in consumer technology. The Pentium contains a superscalar architecture — two parallel execution pipelines — that nearly doubles performance over the 486. It powers the PC multimedia revolution of the mid-1990s.
1994Pentium FDIV bug — the $475 million lessonIn November 1994, mathematics professor Thomas Nicely discovers that the Pentium processor contains a flaw in its floating-point division unit — producing slightly incorrect results for certain calculations. Intel initially downplays the issue, stating it affects only specialised mathematical work. IBM halts Pentium PC shipments. Public outrage forces Intel to offer free replacements to all customers — at a cost of $475 million. The Pentium FDIV bug is the most costly product defect in Intel's history and changes how the company communicates product issues.
1995Pentium Pro — the architecture that powers serversIntel introduces the Pentium Pro in November 1995 — the first of Intel's P6 microarchitecture processors, optimised for 32-bit server workloads with out-of-order execution. The Pentium Pro is the foundation of the Pentium II, Pentium III and Pentium M architectures that will dominate the next decade. Intel also launches the USB standard in 1996 — a collaboration with Microsoft, Compaq and others that replaces the tangle of serial and parallel ports on PCs with a universal connector.
1997Pentium II and the iMac challengeIntel introduces the Pentium II in May 1997 — housed in a distinctive cartridge rather than a socket, running at 233–450MHz with 7.5 million transistors. The Pentium II dominates the PC market. But Apple's introduction of the iMac in 1998, using Motorola/IBM PowerPC processors, demonstrates that Intel's architecture is not the only path to a compelling personal computer. The PowerPC's performance per watt exceeds Intel's offerings — a warning Intel fails to heed about the future importance of power efficiency.
1999Pentium III and the internet eraIntel introduces the Pentium III in February 1999, adding SSE (Streaming SIMD Extensions) for multimedia and internet applications. The Pentium III is the processor of the dot-com boom — powering the web servers that host the internet and the PCs that access it. Intel's revenues surpass $29 billion in 2000, with the company employing 86,000 people. Intel's P/E ratio approaches 50 during the internet bubble, making it briefly one of the ten most valuable companies in the world.
2000Pentium 4 — the GHz race and its limitsIntel introduces the Pentium 4 in November 2000 — designed to achieve extremely high clock speeds through a very deep pipeline (20 stages, later 31). The Pentium 4 reaches 3.8GHz but at enormous power cost: top models consume 115 watts and require active cooling. The 'GHz race' hits a physical wall — faster clocks generate exponentially more heat. AMD's Athlon processor, based on a more efficient architecture, outperforms the Pentium 4 at lower clock speeds. Intel's NetBurst architecture is an engineering dead-end that costs the company years of competitive leadership.
2003Intel Centrino — mobile computing beginsIntel introduces the Centrino mobile platform in March 2003 — combining a Pentium M processor, Intel wireless networking and a mobile chipset into an integrated platform for laptops. The Pentium M is derived from the Pentium III's efficient architecture rather than the hot, power-hungry Pentium 4. Centrino laptops achieve dramatically longer battery life than previous Intel-based laptops. Centrino is Intel's most commercially successful marketing platform and the origin of Intel's eventual Core architecture.
2005Apple switches to Intel — Mac goes x86Steve Jobs announces at WWDC 2005 that Apple will switch the Macintosh from IBM/Motorola PowerPC processors to Intel's x86 processors — completing the transition in just 12 months (announced June 2005, completed August 2006). The Apple-Intel partnership is a landmark: it means every major personal computer — Mac, PC, laptop, desktop — now runs on x86 architecture. Intel gains Apple's premium market and the reputational benefit of powering the most admired personal computers in the world.
2006Intel Core 2 — the architecture resetIntel introduces the Core 2 Duo in July 2006 — abandoning the failed Pentium 4 NetBurst architecture and returning to the efficient P6 lineage of the Pentium III and Pentium M. The Core 2 delivers dramatically better performance per watt than the Pentium 4. Intel's Israeli design team, which had developed the Pentium M, leads the Core 2 development. The Core 2 restores Intel's performance leadership over AMD's Athlon 64 and establishes the Core architecture that Intel uses to this day.
2007iPhone launches — Intel turns down AppleApple approaches Intel in 2005–2006 about producing the processor for its planned smartphone. Intel CEO Paul Otellini declines — the margins are too thin and the volume too uncertain. Apple turns to ARM-based designs manufactured by Samsung. The iPhone launches in June 2007 with a Samsung-manufactured ARM processor. Otellini later calls declining the iPhone chip 'the biggest mistake I ever made.' Intel's failure to participate in mobile computing — the largest computing market in history — costs it irrelevance in an entire product generation.
2008Intel Core i7 — Nehalem architectureIntel introduces the Core i7 processor family in November 2008, based on the new Nehalem microarchitecture. Nehalem integrates the memory controller directly onto the processor die — previously it was in a separate chip — reducing latency and increasing memory bandwidth dramatically. The Core i7 establishes the i3/i5/i7 brand hierarchy that Intel uses for the next fifteen years. Nehalem is the foundation of Intel's server dominance: Xeon processors based on Nehalem power the majority of the world's data centres through the 2010s.
2010Intel acquires McAfee for $7.7 billionIntel acquires security software company McAfee for $7.7 billion in August 2010 — one of the most puzzling acquisitions in technology history. Intel's stated rationale is that security needs to be built into hardware rather than applied in software. The acquisition makes no strategic sense to most analysts; security software has nothing to do with chip manufacturing. McAfee's performance under Intel is poor; Intel eventually spins it out as an independent company (rebranded as Trellix) in 2022 after writing off most of the acquisition value.
2011Sandy Bridge — Intels manufacturing peakIntel introduces Sandy Bridge processors in January 2011 — the first Intel processors built on a 32nm process with an integrated graphics unit on the same die as the CPU. Sandy Bridge is widely considered the best Intel processor generation of the decade: excellent performance, reasonable power consumption and strong integrated graphics. The 22nm Ivy Bridge (2012) and 22nm Haswell (2013) follow with refinements. During this period Intel's manufacturing process is definitively ahead of TSMC and Samsung — a lead it is about to squander.
2012Intel Atom fails in mobile — Windows RT disasterIntel's Atom processor — designed for netbooks and tablets — fails to gain significant traction in the smartphone and tablet market that ARM dominates. Intel powers Microsoft's Surface RT tablets and several Android devices, but the power consumption remains too high and the ecosystem too thin to compete with ARM. Intel spends billions subsidising Atom in mobile devices through its 'contra-revenue' strategy — paying manufacturers to use Intel chips. It never achieves meaningful mobile market share and eventually exits the phone chip market entirely in 2016.
201510nm delays begin — the manufacturing crisisIntel announces delays in its transition to 10nm manufacturing in 2015 — the beginning of a manufacturing crisis that will last seven years. Intel's 10nm process, targeting transistor densities far beyond TSMC's competing 10nm, proves extraordinarily difficult to manufacture at acceptable yields. While Intel struggles, TSMC successfully ramps 10nm (2017) and 7nm (2018). For the first time since the 1980s, Intel's manufacturing process is not the world's most advanced. The company that built its entire identity on manufacturing superiority is falling behind.
2017Meltdown and Spectre — fundamental CPU vulnerabilitiesResearchers reveal Meltdown and Spectre in January 2018 — fundamental security vulnerabilities in the speculative execution mechanisms used by virtually all modern processors, including Intel's. Meltdown is primarily an Intel problem; Spectre affects nearly all processor architectures. Patches for Meltdown reduce performance by up to 30% on some workloads. The vulnerabilities, present in processors manufactured since 1995, cannot be fully fixed in hardware without redesigning the processors. Intel faces class-action lawsuits and the revelation that its processors' security model is fundamentally flawed.
2018Brian Krzanich resigns — #MeToo crisisIntel CEO Brian Krzanich resigns in June 2018 after the company discovers he had a consensual relationship with an Intel employee, violating Intel's non-fraternisation policy. The departure leaves Intel without a CEO during a critical period of competitive pressure from AMD and manufacturing challenges. CFO Bob Swan becomes interim CEO, then permanent CEO in January 2019. The management disruption compounds Intel's technical difficulties and delays strategic decisions about manufacturing and product architecture.
2019AMD Ryzen demolishes Intel — the comebackAMD's Ryzen 3000 series processors, manufactured at TSMC's 7nm process, match or beat Intel's equivalent products at lower prices in 2019. Intel, still stuck on 14nm (due to 10nm delays), cannot compete on process technology. For the first time in over a decade, Intel faces genuine competition in the desktop and laptop CPU market. AMD's server CPUs (EPYC) begin winning major cloud customers — Amazon, Google and Microsoft all announce EPYC deployments. Intel's data centre market share begins to erode.
2020Apple M1 — Apple abandons IntelApple announces at WWDC 2020 that it will abandon Intel processors for its own Apple Silicon chips, manufactured by TSMC. The first M1-powered Macs launch in November 2020 and stun the industry: the M1 delivers dramatically better performance per watt than Intel's equivalent chips. The M1 MacBook Air has longer battery life and better performance than Intel-powered MacBook Pros costing more. Apple completes the Intel transition in June 2023. Losing Apple — whose Macs represented a significant Intel revenue stream and enormous reputational value — is Intel's most public competitive humiliation.
2021Intel 10nm finally ships — Alder LakeIntel finally ships mainstream 10nm processors at volume in 2021 — six years after originally planned. The Alder Lake (12th generation Core) processors, using a hybrid architecture combining performance and efficiency cores, are competitive with AMD's best offerings. Intel's 10nm process — renamed 'Intel 7' to avoid the embarrassing comparison with TSMC's more advanced 7nm — delivers reasonable products but at enormous delay cost. Intel has spent six years fighting fires while AMD, Apple and ARM-based chips redrew the competitive landscape.
2021Pat Gelsinger returns as CEO — IDM 2.0Pat Gelsinger — a former Intel engineer who left to become CEO of VMware — returns to Intel as CEO on 15 February 2021, replacing Bob Swan. Gelsinger is the first engineer to lead Intel since Andy Grove. He announces 'IDM 2.0' — Integrated Device Manufacturer 2.0 — a strategy to restore Intel's manufacturing leadership while simultaneously opening Intel's fabs to manufacture chips for other companies (Intel Foundry Services). The strategy requires $100 billion in capital investment over five years and a complete cultural reinvention.
2021Intel announces $20 billion Arizona fabsIntel announces a $20 billion investment to build two new semiconductor fabs in Chandler, Arizona in March 2021 — the largest private US manufacturing investment in decades. The announcement is partly a response to the CHIPS Act discussions in Congress and partly a genuine strategic commitment to US-based advanced manufacturing. Intel positions itself as the only US-owned company capable of manufacturing leading-edge chips — a geopolitical argument that resonates in Washington as semiconductor supply chain vulnerability becomes a national security concern.
2022Intel Foundry Services — competing with TSMCIntel launches Intel Foundry Services (IFS) in 2021 and signs its first major customer — Qualcomm — in 2022. IFS aims to manufacture chips for other companies on Intel's leading-edge process nodes. The ambition is enormous: Intel wants to become the western world's answer to TSMC, manufacturing chips for US defense, automotive and consumer electronics companies who cannot or will not use a Taiwanese foundry. The challenge is equally enormous: TSMC has 35 years of foundry experience; Intel has none.
2022CHIPS Act — $8.5 billion for IntelThe US CHIPS and Science Act, signed into law in August 2022, allocates approximately $8.5 billion in direct funding to Intel — the largest single CHIPS Act award — plus an estimated $11 billion in loans for Intel's US manufacturing expansion. The funding supports Intel's fabs in Arizona, Ohio and Oregon. Intel's argument — that the US needs a domestically owned, leading-edge chip manufacturer for national security — resonates with the Biden administration. Intel's survival as a manufacturer has become a matter of US industrial policy.
2023Intel 4nm — process recoveryIntel ships its 'Intel 4' process node (equivalent to approximately 4nm in TSMC terms) in 2023, powering Meteor Lake — the first Intel processor to use a chiplet architecture with components manufactured on different process nodes. Intel 4 represents a genuine recovery of manufacturing competitiveness after years of delays. The chiplet approach — assembling processors from multiple smaller dies rather than one large die — allows Intel to use the best available process for each component, including TSMC manufacturing for some tiles.
2024Intel Gaudi — AI chip challenger to NvidiaIntel's Gaudi 3 AI accelerator chip launches in 2024, positioned as a lower-cost alternative to Nvidia's H100 and H200 GPUs for AI training and inference. Intel prices Gaudi 3 aggressively — approximately 40% less than equivalent Nvidia hardware — and targets customers frustrated by Nvidia's supply constraints and premium pricing. Early adopters include IBM and several cloud providers. Gaudi 3 represents Intel's most credible attempt to participate in the AI chip boom that has made Nvidia the world's most valuable company.
2024Pat Gelsinger ousted — Intel in crisisPat Gelsinger resigns as Intel CEO on 1 December 2024 under pressure from the board — effectively fired after three years of declining revenues, manufacturing setbacks and a stock price that has fallen over 60% during his tenure. Intel's 18A process node — intended to restore process leadership by 2025 — has suffered yield problems. Intel's market capitalisation has fallen below $100 billion from a peak above $250 billion. CFO David Zinsner and product chief Michelle Johnston Holthaus become interim co-CEOs as the board searches for a permanent successor.
2025Intel 18A — the make-or-break process nodeIntel's 18A process node — featuring RibbonFET gate-all-around transistors and PowerVia backside power delivery — is Intel's most ambitious manufacturing technology in decades and its clearest attempt to reclaim process leadership. Initial 18A test chips show promising results. Microsoft commits to using 18A for a custom chip. If 18A achieves acceptable yields at volume, it validates Intel's IDM 2.0 strategy. If it fails, Intel faces existential questions about whether to remain a manufacturer at all. The fate of 18A is the most consequential question in the semiconductor industry.
2025Lip-Bu Tan becomes CEO — new chapterLip-Bu Tan — former CEO of Cadence Design Systems and a semiconductor industry veteran with deep foundry experience — becomes Intel's CEO in March 2025. Tan's appointment is widely welcomed by the industry: he understands chip design, manufacturing and the foundry business. He faces an extraordinary challenge: restore Intel's manufacturing competitiveness, build a viable foundry business, compete with AMD in PCs and servers, and develop an AI chip strategy — all simultaneously, with limited financial resources and a demoralised workforce.