AMD vs Intel

AMD was founded on May 1, 1969 by Jerry Sanders and seven colleagues who had all left Fairchild Semiconductor — the company that had spawned Intel a year earlier. Sanders, a former actor and salesman, had been passed over for the top job at Fairchild. He started AMD with $100,000 in seed capital and a philosophy that "people first, products and profits will follow."

IBM's decision to use Intel's 8086 processor in the original PC came with a condition: Intel had to license the design to a second manufacturer for supply security. Intel chose AMD. This "second source" agreement gave AMD the right to manufacture Intel-compatible chips and launched it as a serious competitor. The deal expired in 1986, beginning decades of legal disputes between the two companies over who owned what intellectual property.

By 2014, AMD was burning through cash, had lost its performance advantage over Intel, and was widely rumoured to be on the verge of bankruptcy. The board appointed Lisa Su as CEO in October 2014. Su, an MIT-trained semiconductor engineer, refocused AMD on data centres and high-performance computing, cancelled underperforming projects, and bet the company on a new chip architecture called Zen. She described her first years as "fixing the foundation before building the house." AMD's stock rose over 3,000% from its 2015 low to its 2021 peak.

AMD launched the MI300X AI accelerator in late 2023 — a chip with 192GB of HBM3 memory and 5.3TB/s bandwidth, offering more memory capacity than NVIDIA's competing H100. Lisa Su described it as "the fastest-ramping product in AMD history." Microsoft, Meta, and Oracle began qualifying it for production workloads. AMD's data centre GPU revenue went from near zero in 2023 to over $5 billion in 2024. For the first time in the AI era, there was a credible alternative to NVIDIA.

In a landmark strategic partnership announced in 2025, OpenAI designated AMD as a "core preferred partner" for AI infrastructure, committing to deploy 6 gigawatts of AMD GPUs — starting with 1 gigawatt of MI450 chips in the second half of 2026. AMD's data centre revenue reached $5.8 billion in Q1 2026 alone — growing sequentially even in AMD's historically weakest quarter, a sign analysts compared to NVIDIA's early AI dominance cycle. AMD guided Q2 2026 total revenue of $11.2 billion. Lisa Su said demand was "sky high." The company that had been weeks from bankruptcy a decade earlier was now a $200 billion chipmaker threatening NVIDIA's AI monopoly.

Gordon Moore published a paper in 1965 observing that the number of transistors on a microchip doubled approximately every two years, while the cost halved. He predicted this trend would continue. "Moore's Law" proved to be one of the most reliable predictions in the history of technology, holding for over fifty years. Intel's entire business strategy was built around the predictable improvement curve that Moore had identified: make chips faster and cheaper at a rate that customers could anticipate and plan around.

Robert Noyce and Gordon Moore founded Intel in July 1968, having left Fairchild Semiconductor. Noyce had co-invented the integrated circuit. Their initial investor, Arthur Rock, raised $2.5 million in two days. Andrew Grove, who joined as employee number three, would eventually become Intel's most transformational CEO.

Intel introduced the 4004 in 1971 — the world's first commercially available microprocessor: a complete CPU on a single chip. It had been commissioned by a Japanese calculator company. Intel engineer Ted Hoff proposed putting the entire computing logic on one chip rather than multiple chips. The 4004 contained 2,300 transistors. A modern chip contains tens of billions. The 4004 made the personal computer possible in principle; a decade of further development made it possible in practice.

IBM chose Intel's 8088 processor for the IBM PC in 1981. IBM's open architecture allowed other manufacturers to clone the PC using Intel chips. The "Wintel" ecosystem — Windows operating system on Intel processors — became the dominant computing platform for three decades. Intel's revenues grew from $800 million in 1980 to $33 billion in 2000 as personal computer adoption spread globally.

Intel spent years falling behind in semiconductor manufacturing, failing to advance to smaller transistor sizes at the pace that Taiwan's TSMC and South Korea's Samsung had achieved. Competitors including Apple, AMD, and Qualcomm were having their chips made by TSMC using processes more advanced than Intel's own factories. CEO Pat Gelsinger, who returned to Intel in 2021, announced an ambitious plan to rebuild Intel's manufacturing capability and become a contract chip manufacturer — directly competing with TSMC. The programme required $100 billion in investment. Gelsinger was replaced in December 2024 after the stock fell 60% and the turnaround stalled. Intel remained the processor inside approximately 70% of the world's PCs — but the AI era was running on Nvidia GPUs and TSMC silicon, and Intel was in danger of becoming a passenger in the industry it had created.