Gottfried Wilhelm Leibniz
Described arithmetic using only 0 and 1, noting how simple the rules of addition become.
Why it matters: Every digital computer counts in binary.
📄 Leibniz GW. Explication de l'arithmétique binaire. Mémoires de l'Académie Royale des Sciences, 1703.
Charles Babbage (with Ada Lovelace, 1843)
Designed a mechanical general-purpose computer with a ‘mill’ (processor) and ‘store’ (memory). Lovelace wrote the first published program for it.
Why it matters: The architecture of every computer since, a century early.
📄 Menabrea LF, notes by A.A.L. (Lovelace). Taylor's Scientific Memoirs 3:666 (1843).
George Boole
Wrote logic as algebra with true/false values and the operations AND, OR and NOT.
Why it matters: The mathematics of every logic gate.
📄 Boole G. An Investigation of the Laws of Thought. London, 1854.
William Eccles & Frank Jordan
Two vacuum-tube amplifiers cross-coupled so the circuit stays in one of two states — 1 bit of memory.
Why it matters: Registers and counters in every processor.
📄 Eccles WH, Jordan FW. The Electrician 83:298 (1918); British patent 148,582.
Claude Shannon
His master's thesis showed that relay circuits can carry out Boole's algebra — and that circuits can be designed with algebra.
Why it matters: Called the most important master's thesis of the century: the birth of digital circuit design.
📄 Shannon CE. A Symbolic Analysis of Relay and Switching Circuits. Trans. AIEE 57:713 (1938).
J. Presper Eckert & John Mauchly; John von Neumann
ENIAC (17,468 vacuum tubes) ran at 5,000 additions a second. Von Neumann's report described keeping the program in memory with the data.
Why it matters: The ‘von Neumann architecture’ is still how computers are organised.
📄 von Neumann J. First Draft of a Report on the EDVAC. Moore School, 30 June 1945.
John Bardeen, Walter Brattain & William Shockley
At Bell Labs a germanium crystal with gold contacts amplified a signal — the first transistor.
Why it matters: Small, cool, reliable switches replaced vacuum tubes (Nobel 1956).
📄 Bardeen J, Brattain WH. Phys. Rev. 74:230 (1948).
Maurice Karnaugh (1953); Jay Forrester (1949–51)
Magnetic-core memory stored bits in tiny iron rings; Karnaugh maps made simplifying gate circuits easy.
Why it matters: Reliable main memory for 20 years; K-maps are still taught today.
📄 Forrester JW. J. Appl. Phys. 22:44 (1951); Karnaugh M. Trans. AIEE 72:593 (1953).
Jack Kilby; Robert Noyce
Kilby (Texas Instruments) put several components on one piece of germanium; Noyce (Fairchild) made a practical flat silicon chip with metal wiring on top.
Why it matters: Every chip descends from it (Kilby: Nobel 2000).
📄 Kilby JS. US patent 3,138,743 (filed 1959); Noyce RN. US patent 2,981,877 (filed 1959).
Mohamed Atalla & Dawon Kahng
Bell Labs built the metal–oxide–semiconductor field-effect transistor, which can be made tiny and flat.
Why it matters: The most-manufactured device in history — over 10²² made.
📄 Kahng D, Atalla MM. IRE-AIEE Solid-State Device Research Conference (1960).
Frank Wanlass & Chih-Tang Sah
Pairs of n-type and p-type transistors: a gate uses almost no power except when it switches.
Why it matters: Nearly every chip today is CMOS.
📄 Wanlass FM, Sah CT. ISSCC Digest, p.32 (1963); US patent 3,356,858.
Texas Instruments
A family of TTL logic chips — 7400 NAND, 7404 NOT, 7408 AND, 7474 flip-flop… — with standard pinouts.
Why it matters: The building blocks of minicomputers, arcade games and lab circuits for decades.
📄 Texas Instruments, SN5400 series (1964); SN7400 commercial range (1966).
Gordon Moore
Predicted the number of components on a chip would keep doubling regularly (every 2 years, from 1975).
Why it matters: Held for about 50 years: from thousands to tens of billions of transistors.
📄 Moore GE. Cramming more components onto integrated circuits. Electronics 38(8):114 (1965).
Robert Dennard
Stored a bit as charge on a capacitor beside one transistor; Intel's 1103 (1970) was the first commercial DRAM.
Why it matters: The RAM in every computer and phone.
📄 Dennard RH. US patent 3,387,286 (filed 1967, granted 1968).
Federico Faggin, Ted Hoff, Stanley Mazor & Masatoshi Shima (Intel)
A whole CPU on one chip: 2,300 transistors, 4-bit, 740 kHz.
Why it matters: Began the microprocessor age — PCs, phones and smart devices.
📄 Faggin F, Hoff ME, Mazor S, Shima M. IEEE Micro 16(6):10 (1996), history of the 4004.
Fujio Masuoka (Toshiba)
Invented flash memory (NOR 1984, NAND 1987) that keeps data with the power off and erases in a ‘flash’.
Why it matters: USB sticks, SSDs and all phone storage.
📄 Masuoka F et al. IEDM Technical Digest, p.464 (1984); p.552 (1987).
Ross Freeman (Xilinx); Sophie Wilson & Steve Furber (Acorn)
Xilinx sold the first field-programmable gate array (logic you can rewire by software); Acorn's ARM1 was a simple, low-power RISC CPU.
Why it matters: ARM designs now power almost every phone; FPGAs prototype new chips.
📄 Xilinx XC2064 (1985); Furber SB. VLSI RISC Architecture and Organization (1989).
Nvidia
GeForce 256 was marketed as the first ‘graphics processing unit’; in 2006–07 CUDA opened GPUs to general computing.
Why it matters: GPUs now train almost all modern AI.
📄 Nvidia GeForce 256 (1999); Nickolls J et al. ACM Queue 6(2):40 (2008).
AMD & Intel
Clock speeds stopped rising (heat), so chips gained several cores working in parallel.
Why it matters: Today's CPUs have 8–100+ cores.
📄 Sutter H. The Free Lunch Is Over. Dr. Dobb's Journal 30(3) (2005).
Intel (after Hisamoto, Hu et al., 1989–2000)
The transistor channel became a thin vertical fin wrapped by the gate on three sides, cutting leakage at 22 nm.
Why it matters: Kept Moore's law going for another decade.
📄 Hisamoto D et al. IEEE Trans. Electron Devices 47:2320 (2000); Intel 22 nm announcement, May 2011.
ASML, TSMC & Samsung
Machines printing with 13.5 nm light from tin plasma entered mass production at the 7 nm node.
Why it matters: The only way to print today's smallest features.
📄 TSMC N7+ volume production (2019); ASML NXE:3400 systems.
Apple; AMD
Apple M1 put CPU, GPU, AI engine and memory in one package (16 billion transistors); AMD built CPUs from several smaller ‘chiplets’.
Why it matters: New ways to keep chips improving as shrinking slows.
📄 Apple M1 announcement, 10 Nov 2020; Naffziger S et al. ISSCC (2020).
Samsung, TSMC & Intel
Stacked nanosheets surrounded by the gate on all four sides: Samsung 3 nm (2022), TSMC N2 and Intel 18A entering volume production in 2025.
Why it matters: The successor to the FinFET.
📄 Samsung 3 nm GAA production announcement, 30 June 2022; TSMC N2 (2025).