The Hardware and Software Landscape
In September 2019, the Bitcoin mining industry operated at a critical juncture between two hardware generations. The Bitmain Antminer S17 series, released earlier in 2019, delivered approximately 56 terahashes per second (TH/s) at a power efficiency of around 45 joules per terahash, representing a meaningful upgrade over the aging S9 generation that still dominated many mining farms. The MicroBT Whatsminer M20S offered comparable performance at roughly 48 TH/s, creating genuine competition in the high-efficiency ASIC market. Meanwhile, Canaan Creative was preparing its next-generation AvalonMiner units and laying the groundwork for what would eventually become the first Bitcoin mining company to list on NASDAQ in November 2019.
The global hashrate in September 2019 stood at approximately 80-85 exahashes per second (EH/s), a figure that would later be recognized as a local peak before a gradual decline into early 2020. Mining software had matured significantly, with most large-scale operations running custom firmware on their ASICs to optimize power consumption and hashboard performance. Major mining pools — including F2Pool, Poolin, BTC.com, and AntPool — collectively controlled over 70% of the network’s hashpower, with F2Pool and Poolin increasingly gaining share from Bitmain-affiliated operations as the industry decentralized slightly from its Chinese manufacturing roots.
The geographic distribution of mining activity remained heavily concentrated in China, particularly in Sichuan and Xinjiang provinces, where abundant hydroelectric and coal-powered electricity provided some of the lowest energy costs globally. Estimates suggested that approximately 65% of global Bitcoin mining occurred in China during this period, with significant secondary operations in the United States, particularly in Washington State and Texas, where renewable energy and deregulated electricity markets attracted new entrants.
Hashrate and Difficulty
Bitcoin’s mining difficulty in September 2019 reflected the sustained high hashrate environment, with the network adjusting every 2,016 blocks (approximately every two weeks) to maintain the target block time of 10 minutes. The difficulty level had been climbing steadily throughout 2019, reaching all-time highs as new-generation ASICs came online and mining farms expanded capacity built during the bear market of 2018 in anticipation of the next price cycle. The network difficulty in late September 2019 stood at roughly 13.7 trillion, a level that required significant capital investment in hardware to mine profitably.
When Bitcoin’s price collapsed from approximately $10,100 to below $8,500 on September 24 — a drop exceeding 15% in a single session — the immediate impact on mining economics was severe. Revenue per terahash fell precipitously, squeezing operations running older S9 hardware that had been operating on thin margins even at higher price levels. The network’s hashrate did not drop immediately, as miners face sunk costs in hardware and long-term electricity contracts, but the economic pressure began to build. Historical data from subsequent months would show that the hashrate peaked in September 2019 and then gradually declined as the least efficient miners were forced offline, eventually bottoming around 70 EH/s in November before recovering into early 2020.
The difficulty adjustment mechanism acted as the network’s economic stabilizer. As less efficient miners capitulated and hashrate declined, the network reduced difficulty, making it easier (and more profitable) for remaining miners to find blocks. This self-correcting mechanism ensured that Bitcoin’s block production remained roughly on schedule at one block every 10 minutes, regardless of price volatility. The adjustment following the September crash was modest — a small decrease — but it signaled the beginning of a trend that would see multiple downward adjustments in late 2019.
Profitability Metrics
Before the September 24 crash, Bitcoin mining with modern S17-class hardware was comfortably profitable at most global electricity rates. At a BTC price of approximately $10,100 and an electricity cost of $0.05 per kilowatt-hour — a rate achievable in parts of China and the Pacific Northwest — an S17 Pro generated roughly $3-4 per day in net profit after electricity costs. Older S9 units were generating $0.50-1.00 per day, making them marginally profitable at best. Block rewards stood at 12.5 BTC (the halving to 6.25 BTC would not occur until May 2020), with transaction fees adding roughly 0.5-1.0 BTC per block on average.
The price crash to $8,500 compressed these margins dramatically. S17 profitability dropped to roughly $1-2 per day at the same electricity rates, while S9 operations at $0.05/kWh fell to break-even or slightly negative territory. For miners paying $0.07/kWh or higher — common in many European and some American operations — even the S17 became unprofitable or barely breakeven. The total daily revenue for the entire network fell from approximately $13 million to $11 million overnight, a $2 million daily revenue reduction that directly impacted mining operations worldwide.
Adding to the pressure, Bitcoin’s price had already been trending downward throughout September, falling from around $10,600 at the beginning of the month. This extended decline meant that mining profitability had been eroding for weeks before the dramatic September 24 crash delivered the coup de grâce to marginal operations. Some miners began hedging their expected Bitcoin output through over-the-counter (OTC) deals or futures contracts to lock in revenues, a practice that was becoming increasingly sophisticated among large-scale operations.
Environmental Impact
The September 2019 mining landscape consumed an estimated 70-75 terawatt-hours of electricity annually, comparable to the entire power consumption of countries like Colombia or Bangladesh. The concentration of mining in China’s Sichuan province during the wet season — when hydroelectric power was abundant and cheap — created a natural seasonal pattern where mining’s environmental footprint was partially offset by renewable energy. During the dry season, miners migrated to Xinjiang and Inner Mongolia, where coal-powered electricity dominated, significantly increasing the carbon intensity of Bitcoin mining operations.
The profitability squeeze following the September crash had an unintended environmental benefit: as less efficient miners shut down older hardware, the network’s overall energy consumption decreased slightly while the share of mining powered by newer, more efficient ASICs increased. This dynamic — where market downturns accelerate the retirement of inefficient hardware — represents one of the few natural mechanisms constraining Bitcoin mining’s environmental impact. The industry was also beginning to see early experiments with flare gas mitigation and stranded hydroelectric utilization, particularly in North America, where environmental regulations and cheap natural gas were creating new mining locations.
Strategic Outlook
The September 2019 price crash served as a stress test for the Bitcoin mining industry, revealing the fragility of operations built on thin margins and aging hardware. The miners who survived — and eventually thrived — were those who had invested in next-generation ASICs, secured long-term electricity contracts at competitive rates, and maintained sufficient treasury reserves to weather extended bear periods. The episode also accelerated the professionalization of the mining industry, with publicly traded companies like Hut 8 and Argo Blockchain gaining conditional listings and attracting institutional capital to fund expansion.
Looking ahead, the approaching May 2020 halving — which would reduce block rewards from 12.5 to 6.25 BTC — loomed as a far more significant challenge to mining economics than any single price crash. Miners who optimized their operations during the September 2019 downturn positioned themselves to survive the halving, while those who failed to adapt faced consolidation or exit. The hashrate decline that began in September 2019 ultimately represented not a failure of Bitcoin’s security model, but a healthy recalibration — proof that the network’s economic design functions as intended, cycling out inefficient participants and rewarding those who operate with discipline and foresight.
Disclaimer: This article is for informational purposes only and does not constitute financial advice. Cryptocurrency investments and mining operations carry significant risk. Always conduct your own research before making investment decisions.
S17 at 56 TH/s was a game changer in 2019. Anyone still running S9s after that price crash was bleeding money on electricity alone
Canaan preparing for a NASDAQ listing while the hashrate was peaking at 85 EH/s. Timing was aggressive given the price environment
the M20S at 48 TH/s was genuinely competitive with the S17. MicroBT quietly ate a huge chunk of Bitmains market share around this period
Custom firmware on ASICs was the real competitive advantage. Stock Bitmain firmware left so much efficiency on the table