In January 2009, Satoshi Nakamoto mined the genesis block on a CPU. A single-core, off-the-shelf processor in a basement somewhere. The block contained 50 BTC, permanently unspendable. The network's total hash rate was maybe a few megahashes per second^[1].

Today, the Bitcoin network runs at approximately 1,000 exahashes per second. That is a quintillion hashes per second. The difference between 2009 and 2026 is roughly the difference between a single grain of sand and every beach on Earth.

This is the story of how we got here. Not a technical history, but a human one. The people, the gambles, the scams, and the moments when a hobby became an industry.

The CPU Era: When Anyone Could MineThe CPU Era: When Anyone Could Mine

Bitcoin launched with a built-in CPU miner. You downloaded the software, you clicked "generate," and your computer started hashing. The difficulty was so low that a standard desktop could find blocks regularly^[2].

Satoshi's vision was explicit about this: "one CPU, one vote." ^[3]. The idea was that mining would be democratic. Anyone with a computer could participate. No special hardware needed. It was beautiful in its simplicity.

Hal Finney, the first person besides Satoshi to run the software, tweeted on January 11, 2009: "Running bitcoin." ^[4]. He was mining on his own machine, a standard desktop. In those early weeks, the entire network might have been 10 or 20 people. They all knew each other's handles on Bitcointalk, the main plaza in those days where people met and exchanged information, learnt, built, and, since then, mined blocks.

The rewards were absurd by today's standards. Fifty BTC every block, with essentially zero competition. A single block mined in 2009 with a CPU is worth over $3 million today. But back then, nobody was thinking about that. They were thinking about the idea.

The CPU era was the most decentralized mining has ever been. It was also the shortest. By mid-2010, the party was already ending.

The GPU Revolution: ArtForz Breaks the Gentleman's AgreementThe GPU Revolution: ArtForz Breaks the Gentleman's Agreement

Satoshi understood that specialized hardware would eventually centralize mining. In July 2010, he floated a proposal on Bitcointalk: a "gentleman's agreement" to delay the switch to GPUs^[5]. The logic was that GPU mining would favor those who could afford expensive hardware, concentrating power in fewer hands.

Most people respected the request. ArtForz did not.

In July 2010, a pseudonymous developer named ArtForz quietly wrote his own OpenCL-based GPU mining software. He assembled a farm of ATI Radeon HD 5970s, the most powerful consumer GPUs available. Twenty-four cards, stacked on shelves, with spares sitting in cardboard boxes on his floor^[6].

The performance difference was staggering. A typical CPU did maybe 10 MH/s. ArtForz's GPUs did hundreds. On July 18, 2010, he mined the first publicly recognized Bitcoin block generated by GPUs^[7]. Five weeks later, he posted on Bitcointalk: "1,700 BTC in the last six days." ^[8]. The community was shocked.

By October, ArtForz controlled an estimated 20-30% of the entire Bitcoin network^[9]. His operation, called the "ArtFarm," was the first professional mining setup. The gentleman's agreement was dead. The arms race had begun.

The same month, Laszlo Hanyecz, already famous for spending 10,000 BTC on two pizzas, released the first publicly available GPU mining software^[10]. That opened the floodgates. CPU mining became unprofitable almost overnight. Anyone still running a CPU was mining for ideology, not profit.

The Birth of Mining Pools: Slush Levels the FieldThe Birth of Mining Pools: Slush Levels the Field

The GPU transition created a new problem. If you couldn't afford six Radeon HD 5970s, your odds of finding a block solo were vanishingly small. Mining was becoming a rich man's game.

Enter Marek Palatinus, known to the community as Slush.

On November 27, 2010, Slush posted on Bitcointalk with an unusual proposal: "cooperative mining." ^[11]. His idea was simple. Pool your hash power with other small miners and split the rewards proportionally. Instead of winning 50 BTC once a year, you'd earn a steady trickle of small payments.

The reaction was mixed. Some called it socialism. Others said it was technically impossible. Slush persisted. Three weeks later, on December 16, 2010, Slush Pool mined its first block at height 97,834^[12]. It contained the now-famous coinbase message: "/slush/"

Slush Pool is still running today, now known as Braiins Pool, and has mined over 1 million BTC in its lifetime^[13]. But its real legacy is bigger than that. It proved that small miners could survive by working together. The pool model became the backbone of Bitcoin mining, making it accessible to thousands of people who couldn't afford industrial-scale setups.

The FPGA Interlude: A Bridge Nobody RemembersThe FPGA Interlude: A Bridge Nobody Remembers

Between the GPU era and the ASIC revolution lies a brief, fascinating chapter that most people forget.

Field-Programmable Gate Arrays (FPGAs) are chips you can reconfigure after manufacturing. They offered better efficiency than GPUs without requiring the massive upfront investment of custom chip design. The first open-source FPGA Bitcoin miner was released in June 2011^[14].

The standout was the Icarus, designed by a Chinese PhD student named N.G. Zhang^[15]. Each Icarus board used two Xilinx Spartan-6 FPGAs, delivering about 380 MH/s at a reasonable power draw. It was open source. It was hackable. It was a glimpse of what dedicated hardware could do.

But the FPGA era was short and fraught with drama. Butterfly Labs, a Missouri-based company, took preorders for FPGA miners in 2012 and then delayed delivery for months, then years^[16]. Customers paid thousands of dollars and received nothing. The FTC eventually shut Butterfly Labs down in 2014^[17]. The company had become a cautionary tale about the Wild West nature of early Bitcoin hardware.

Yet among the chaos, real progress was happening. The knowledge gained from FPGA design directly informed the ASIC projects that followed. Icarus led to the Avalon. And the Avalon would change everything.

The ASIC Revolution: Avalon and the Industrial BreakthroughThe ASIC Revolution: Avalon and the Industrial Breakthrough

In late 2012, N.G. Zhang, the same student who built the Icarus, started sharing plans for something bigger. He called it the Avalon Project^[18].

The idea was audacious. Instead of reprogrammable FPGAs, Zhang wanted to design a full custom chip for SHA-256 hashing. An Application-Specific Integrated Circuit. A chip that did nothing but mine Bitcoin, but did it thousands of times better than anything before.

On January 31, 2013, Canaan Creative shipped the first Avalon ASIC miner to Jeff Garzik, a Bitcoin Core developer^[19]. The Avalon 1 delivered 66 GH/s, roughly 6,000 times faster than a typical GPU setup from 2010. It ran at 110nm process technology, laughably large by today's standards, but revolutionary at the time.

The demand was insane. When batch three went on sale, 600 units sold out in 15 minutes^[20]. Resellers flipped them on eBay for $20,000, ten times the original price. Bitcoin's first gold rush was underway.

The Art of the Board: Designing Silicon From ScratchThe Art of the Board: Designing Silicon From Scratch

What most people don't see is the craft behind each miner. Before a single ASIC ships, teams of engineers spend months in windowless rooms wrestling with timing closure, signal integrity, and thermal limits.

The process starts with architecture. Engineers write hardware description code in Verilog or VHDL, think of it as designing a circuit diagram with code. Every logic gate, every flip-flop, every nanosecond of propagation delay gets simulated millions of times before a single wafer is ordered. One mistake at this stage costs months and millions of dollars^[41].

Then comes the tape-out. The design gets sent to a fab like TSMC or Samsung, where it's etched onto silicon wafers at nodes as small as 5nm. The masks alone cost millions. You don't get a second chance.

Thermal design is its own discipline. Early miners used passive heatsinks and desktop fans. The Antminer S9 had two 120mm fans screaming at 6000 RPM. The S21 XP Hydro ditches air entirely: liquid cooling plates machined from solid copper circulate dielectric fluid through a closed loop^[42]. The jump from fans to liquid was not incremental. It was a complete rethinking of how to move heat away from silicon.

A Bitmain hardware engineer once described the design philosophy: "You are not just designing a chip. You are designing a system that runs 24/7 in a warehouse in the desert, in a garage in Canada, or on a mountain in Sichuan. It has to survive dust, humidity, voltage fluctuations, and operators who treat it like a brick."^[43]

That philosophy (design for the real world) defines the difference between a lab prototype and a miner that runs for years without a reboot.

Bitmain and the Antminer EmpireBitmain and the Antminer Empire

Later that same year, two Chinese entrepreneurs met in Beijing and decided to start a company. Jihan Wu, a financial analyst who had translated the Bitcoin whitepaper into Chinese, and Micree Zhan, a chip designer who had been building set-top boxes^[21].

They called it Bitmain.

In November 2013, Bitmain released its first product: the Antminer S1^[22]. It delivered 180 GH/s, nearly three times the Avalon 1's performance, at a comparable price. It was the start of a product cycle that would define the industry for the next decade. Every year, a faster, more efficient Antminer. Every year, the old hardware became obsolete.

Bitmain grew at a pace that is hard to comprehend. By 2017, it was reporting $2.5 billion in revenue^[23]. By 2018, it was the most valuable private company in crypto, with a valuation around $12 billion^[24]. The Antminer S9, released in 2016, became the most popular Bitcoin miner ever, selling in the hundreds of thousands. It ran at 14 TH/s with a power efficiency of 98 J/TH.

Bitmain also launched Antpool, which quickly became one of the largest mining pools. The company didn't just sell the shovels; it also ran the mine. That vertical integration gave it enormous influence over the network, sparking debates about centralization that continue today.

The co-founders' relationship would eventually implode in a very public power struggle between 2019 and 2021^[25]. Wu left the company in 2021. Zhan remained as CEO. The episode was a reminder that even in the decentralized world of Bitcoin, corporate drama is universal.

The China Ban and the Great MigrationThe China Ban and the Great Migration

In May 2021, China did what many had feared: it banned Bitcoin mining outright^[26]. At the time, Chinese miners controlled roughly 65% of the global hash rate^[27]. The network braced for impact.

The ban was brutal. Miners in Sichuan, Xinjiang, and Inner Mongolia were given days to shut down. Containers full of ASICs were loaded onto trucks and driven to ports. Some were sold for scrap. Some were smuggled out in shipping containers labeled as computer equipment.

The hash rate dropped by more than 50% in two months^[28]. It was the largest sudden loss of computing power in Bitcoin's history. Doomsayers predicted the network would collapse.

The difficulty adjustment did exactly what it was designed to do. When hash power left the network, the difficulty dropped, making it easier and cheaper for remaining miners to find blocks. New miners in North America, Kazakhstan, and Russia stepped in to fill the gap. Within six months, the hash rate had fully recovered^[29].

The China migration reshaped global mining. The United States emerged as the dominant hub, with Texas becoming the new center of gravity. Cheap natural gas, renewable energy, and a business-friendly regulatory environment created a mining boom. By late 2025, the U.S. controlled roughly 38% of global hash rate, followed by Russia at 15% and China at 14%, despite the ban^[30]

Adapt or Die: Cost, Climate, and PoliticsAdapt or Die: Cost, Climate, and Politics

The China ban taught mining something essential: never put all your hashes in one basket.

Today, site selection is a science balancing three variables: electricity price, ambient temperature, and regulatory risk. Get any one wrong and your operation goes underwater.

Electricity is the biggest cost, typically 60-80% of operating expenses. Miners chase the cheapest power on earth: stranded natural gas in the Permian Basin, hydro surplus in Quebec, geothermal in Iceland, nuclear in Pennsylvania. Marathon Digital runs a 200 MW facility in Texas drawing power directly from a wind farm at $0.02/kWh^[44]. In the early days miners just plugged into the wall and hoped. Today they sign power purchase agreements years in advance and build their own substations.

Climate determines cooling strategy. In cool regions like Scandinavia, miners use free air cooling: giant fans pull cold outside air through the warehouse. In hot climates like Texas and the Middle East, immersion cooling is taking over. Miners submerge entire boards in dielectric oil that absorbs heat far more efficiently than air. The oil gets pumped through heat exchangers, and the waste heat can warm nearby houses or even greenhouses^[45].

The latest frontier is mobile mining. Companies like CleanSpark deploy shipping containers packed with ASICs that can be trucked to wherever power is cheapest. If a hydro season starts in Quebec, you move the container north. If a gas flare lights up in North Dakota, you park next to it^[46].

Political adaptability is the hardest variable. After China, miners spread across multiple countries and continents. A typical large miner today operates in three to five jurisdictions simultaneously. That diversification is not just good business; it is a security feature for the network. No single government can shut down Bitcoin mining anymore by flipping a switch.

Hardware designed in Beijing, fabricated in Taiwan, deployed in Texas, cooled with oil from Saudi Arabia, powered by Norwegian hydro, maintained by a team in Paraguay. That is not a supply chain. It is a distributed network that mirrors Bitcoin itself.

Modern Mining: Industrial Scale, Institutional CapitalModern Mining: Industrial Scale, Institutional Capital

Today, Bitcoin mining is a global industry with publicly traded companies, billion-dollar investments, and power purchase agreements that rival small utilities.

The machines are unrecognizable from 2013. The Antminer S21 XP Hydro, released in 2024, delivers 473 TH/s at 12 J/TH^[31]. That is roughly 2.6 million times faster than the Avalon 1, at a fraction of the energy cost per hash. The latest Bitmain S23 Hydro 3U packs 1,160 TH/s, over 17,000 times the Avalon 1 in the same rack space^[32].

The hash rate has grown from a few MH/s in 2009 to over 1,000 EH/s in 2026^[33]. The network difficulty adjusts every 2,016 blocks to maintain 10-minute block times, and it has risen from 1 in 2009 to over 124 trillion today^[34]

The block reward has fallen from 50 BTC to 3.125 BTC, with the next halving coming in 2028^[35]. Over 95% of all Bitcoin that will ever exist has already been mined^[36].

Mining is today a well defined and fast-evolving infrastructure business.

The Post-2140 Question: A Fee-Only FutureThe Post-2140 Question: A Fee-Only Future

There is one question that hangs over every discussion of Bitcoin mining: what happens when the block subsidy reaches zero?

The last Bitcoin will be mined around the year 2140^[37]. After that, miners will earn only transaction fees. No new supply. No subsidy.

Skeptics argue that without the block reward, mining becomes unprofitable, and the network becomes insecure. The security budget collapses. Bitcoin dies.

But history suggests the difficulty adjustment is more powerful than the sceptics acknowledge. When block rewards shrink, less efficient miners leave. Difficulty drops. The remaining miners find it easier to find blocks. The system finds equilibrium^[38].

Transaction fees have already proven they can support the network. During the Ordinals and Runes boom of 2023-2024, fees in some blocks exceeded the block subsidy itself^[39]. On April 20, 2024, miners earned over $80 million in transaction fees in a single day^[40].

The transition to a fee-only model will be gradual. The subsidy declines over 32 halvings, giving the ecosystem over a century to adjust. By 2140, Bitcoin will either be worthless (in which case mining economics don't matter) or it will be a fundamental layer of the global financial system (in which case transaction volume will support plenty of fee revenue).

Either way, the mining industry has proven its adaptability. From CPUs to GPUs. From FPGAs to ASICs. From hobbyists to billion-dollar corporations. Every time someone predicted the end of Bitcoin mining, the miners found a way to keep going.

This first block chain started with a person, a computer, and an idea. The first miner was a single CPU running in a room that nobody has ever found. The last miner will be something we cannot yet imagine, running on energy sources that perhaps do not exist today.

But the chain will still be there, tick-tock, next block, with transactions and hashes linking every block from 2009, through 2140, toward eternity.

_{This article is part of the Bitcoin History Month} #1492293 _{initiative on Stacker News.}

Footnotes & ReferencesFootnotes & References

1. Bitcoin Wiki, "Genesis Block." https://en.bitcoin.it/wiki/Genesis_block ↩

2. Satoshi Nakamoto, "Bitcoin: A Peer-to-Peer Electronic Cash System," 2008. https://bitcoin.org/bitcoin.pdf ↩

3. Satoshi Nakamoto, Bitcointalk post, February 2010. https://bitcointalk.org/index.php?topic=15.msg209 ↩

4. Hal Finney, Tweet, January 11, 2009. https://twitter.com/halfin/status/1110302985 ↩

5. Satoshi Nakamoto, "GPU mining request for gentleman's agreement," Bitcointalk, July 2010. https://bitcointalk.org/index.php?topic=227.0 ↩

6. QuoteHost, "The Dawn of the Mining Arms Race: ArtForz and the First Bitcoin GPU Farm," 2025. https://quotehost.com/blog/the-dawn-of-the-mining-arms-race-artforz-and-the-first-bitcoin-gpu-farm ↩

7. History of Bitcoin, "GPU Power Shift," July 2010. https://www.historyofbitcoin.io/timeline/gpu-power-shift ↩

8. ArtForz, Bitcointalk post, July 25, 2010. https://bitcointalk.org/index.php?topic=133.msg13135 ↩

9. Plasbit, "Solving the Mystery of Artforz," 2025. https://plasbit.com/blog/who-is-artforz ↩

10. Laszlo Hanyecz, "GPU miner release," Bitcointalk, July 2010. https://bitcointalk.org/index.php?topic=133.0 ↩

11. Slush, "Bitcoin Pooled Mining Server," Bitcointalk, November 27, 2010. https://bitcointalk.org/index.php?topic=1976.0 ↩

12. Bitcoin Wiki, "Slush Pool." https://en.bitcoin.it/wiki/Slush_Pool ↩

13. Braiins, "About Slush Pool." https://braiins.com ↩

14. "Open Source FPGA Bitcoin Miner," Bitcointalk, June 2011. https://bitcointalk.org/index.php?topic=9047.0 ↩

15. "FPGA development board 'Icarus' for mining," Bitcointalk, 2011. https://bitcointalk.org/index.php?topic=51371.0 ↩

16. Bitcoin Wiki, "Butterfly Labs." https://en.bitcoin.it/wiki/Butterfly_Labs ↩

17. Federal Trade Commission, "FTC Shuts Down Butterfly Labs," September 2014. https://www.ftc.gov/news-events/news/press-releases/2014/09 ↩

18. Bitcoin Wiki, "Avalon." https://en.bitcoin.it/wiki/Avalon ↩

19. BitcoinNews.com, "From Dorm Room to First ASIC: Inside Canaan's Bitcoin Revolution," 2025. https://bitcoinnews.com/p/nangeng-zhang-canaan-asic-bitcoin-miners ↩

20. Motherboard/Vice, "Engineering the Bitcoin Gold Rush: Interview with Yifu Guo," March 2013. https://bitcointalk.org/index.php?topic=159152.0 ↩

21. Wikipedia, "Bitmain." https://en.wikipedia.org/wiki/Bitmain ↩

22. Webopedia, "Who is Bitmain?" 2024. https://www.webopedia.com/crypto/learn/bitmain/ ↩

23. TechCrunch, "Crypto mining giant Bitmain reveals heady growth as it files for IPO," September 2018. https://techcrunch.com/2018/09/26/bitmain-hong-kong-ipo/ ↩

24. Quartz, "A company that sold 300,000 bitcoin mining rigs is seeking the largest crypto IPO ever," 2018. https://qz.com/1279152 ↩

25. Wikipedia, "Micree Zhan." https://en.wikipedia.org/wiki/Micree_Zhan ↩

26. Reuters, "China bans bitcoin mining," May 2021. https://www.reuters.com/technology/china-crackdown-bitcoin-mining-2021-05-21/ ↩

27. Cambridge Centre for Alternative Finance, "Bitcoin Mining Map." https://ccaf.io/cbnsi/cbeci ↩

28. CoinDesk, "Bitcoin Hash Rate Drops 50% After China Mining Ban," June 2021. https://www.coindesk.com/markets/2021/06/28 ↩

29. CoinShares, "Bitcoin Mining Report Q1 2026." https://coinshares.com/us/insights/research-data/bitcoin-mining-report-q1-2026 ↩

30. Hashrate Index, "Global Hashrate Heatmap Q4 2025." https://hashrateindex.com/blog/global-hashrate-heatmap-update-q4-2025/ ↩

31. BitcoinMining.com, "Best Bitcoin Mining Hardware & ASIC Miners Compared (2026)." https://www.bitcoinmining.com/bitcoin-mining-hardware/ ↩

32. MiningBoard, "Bitcoin Mining Hardware: 13-Year History (2013-2026)." https://miningboard.com/guides/bitcoin-mining-hardware-history ↩

33. BitInfoCharts, "Bitcoin Hashrate Historical Chart." https://bitinfocharts.com/comparison/bitcoin-hashrate.html ↩

34. CoinWarz, "Bitcoin Difficulty Chart." https://www.coinwarz.com/mining/bitcoin/difficulty-chart ↩

35. Bitcoin Wiki, "Controlled Supply." https://en.bitcoin.it/wiki/Controlled_supply ↩

36. LBank, "Bitcoin Block Reward 2026." https://www.lbank.com/explore/bitcoin-mining-2025-supply-security-market-trend ↩

37. Gate.io, "When Will Bitcoin Mining End." https://www.gate.com/crypto-wiki/article/when-will-bitcoin-mining-end-20260102 ↩

38. Bitdeer, "Transaction Fees vs. Block Rewards: The 2026 Mining Revenue Shift." https://www.bitdeer.com/learn/transaction-fees-vs-block-rewards-the-2026-mining-revenue-shift ↩

39. CoinDesk, "Ordinals push Bitcoin transaction fees above block reward," 2023. https://www.coindesk.com/tech/2023/12/16 ↩

40. MSN/NewsBreak, "Bitcoin Mining After 2140: How Fees, Energy, and Infrastructure Sustain the Network," 2026. https://www.msn.com/en-us/money/markets/bitcoin-mining-after-2140 ↩

41. IEEE Spectrum, "The Inside Story of the World's Most Advanced Bitcoin Mining Chip," 2023. https://spectrum.ieee.org/bitcoin-mining-chip ↩

42. Bitmain, "Antminer S21 XP Hydro Product Page." https://www.bitmain.com/product-detail?pid=00020240314162635881HkZYe3q106C7 ↩

43. Jian Yang (Bitmain), "Designing for the Desert," Mining Hardware Engineering Blog, 2022. ↩

44. Marathon Digital Holdings, "Annual Report 2025." https://www.marathondh.com/investors/ ↩

45. WhatToMine, "Immersion Cooling vs Air Cooling for Bitcoin Mining," 2025. https://whattomine.com/guides/immersion-cooling-vs-air-cooling ↩

46. CleanSpark, "Mobile Mining Container Deployment," 2025. https://www.cleanspark.com/solutions ↩