The surge in blockchain and digital cryptocurrencies like Bitcoin is undeniable. In a recent earnings call, NVIDIA’s CFO Colette Kress noted that “strong demand for cryptocurrency exceeded our expectations.” Soaring chip prices have attracted more manufacturers, including Samsung, to enter the market—a trend largely driven by the booming cryptocurrency mining sector.
For chipmakers, this represents both an opportunity and a challenge. While profits are rising, the industry is also wary: if the cryptocurrency market falters, can artificial intelligence (AI) step in to sustain growth?
The Real Culprit Behind the GPU Shortage
Cryptocurrency “mining” is at the heart of the GPU supply shortage. In simple terms, mining involves computers across a network racing to solve complex mathematical problems. The first to solve and verify a transaction is rewarded with new coins and transaction fees—much like extracting minerals from the earth, hence the term “mining.”
To control inflation, Bitcoin’s total supply is capped at 21 million coins. Introduced in 2009, its production rate halves every four years until all coins are mined by around 2140.
Mining has evolved through five key phases:
- CPU Mining
- GPU Mining
- FPGA Mining
- ASIC Mining
- Large-Scale Cluster Mining
Each step brought exponential gains in processing speed:
- CPU: 20 MHash/s
- GPU: 400 MHash/s
- FPGA: 25 GHash/s
- ASIC: 3.5 THash/s
- Cluster Mining: 3.5 THash/s × number of units
As mining difficulty increased, so did hardware requirements. While CPUs were initially used, their general-purpose design made them inefficient for heavy computation. By early 2013, miners had largely shifted to GPUs and FPGAs. The 2017 crypto boom accelerated this shift, causing GPU shortages and price spikes.
Contrary to popular belief, AMD GPUs aren’t always the preferred choice. Different cryptocurrencies use different algorithms:
- Ethereum (ETH) uses the I/O-intensive DaggerHashimoto algorithm, favoring AMD’s Polaris architecture.
- Zcash (ZEC) relies on the computation-heavy Equihash algorithm, where NVIDIA cards excel.
Miners choose hardware based on efficiency, availability, and cost. Mid-range cards like the AMD RX 570 often offer the best balance of performance and power consumption. High-end models like the GTX 1080 Ti or AMD Vega 64 are less cost-effective due to high prices and energy use.
ASIC miners, designed for specific algorithms, offer even greater efficiency. Bitmain, a leading Chinese manufacturer, produces the Antminer S9—a dedicated Bitcoin mining machine using custom 16nm ASIC chips. Another major player, Bitfury, sells full data center setups to enterprise clients.
Though FPGAs offer flexibility, their lower efficiency has limited their role in mining. Still, they remain relevant for certain algorithms and might see renewed interest in the future.
Rising Demand for ASICs and Market Impact
With large-scale ASIC farms dominating Bitcoin mining, individual miners using smaller setups face diminishing returns. This trend is boosting the ASIC market significantly.
MediaTek’s Q4 2017 earnings report revealed that Bitmain had fully booked TSMC’s 12nm production capacity. That year, Bitmain’s chip sales reached ¥14.3 billion, making it the second-largest IC design company in China after Huawei’s HiSilicon.
With reported net profits exceeding ¥1 billion in the first half of 2017, Bitmain exemplifies how cryptocurrency mining is driving growth in semiconductor design and manufacturing.
Yet this rapid growth brings uncertainty. Chip manufacturers must consider how long the crypto boom will last and how to pivot if demand falls.
From Mining to AI: A Strategic Shift
Anticipating market changes, companies like Bitmain are expanding into AI. In 2017, Bitmain launched its SOPHON TPU, an ASIC chip designed for AI applications.
The company is focusing on three strategic areas:
- AI Inference Chips: Developing deep learning processors with rapid iteration cycles—typically every nine months—along with industry-specific hardware solutions.
- Big Data and Private Cloud: Partnering with industries to offer customized data and cloud services based on their AI chips.
- Robotics: Enhancing capabilities through acquisitions like the purchase of萝卜机器人 (LuoBo Robot).
Bitmain is targeting key verticals including security, internet services, and urban big data.
Although the AI chip market already has strong competitors like Cambricon and Horizon Robotics, Bitmain’s years of experience in high-performance chips make it a serious contender.
The AI chip sector is still growing fast, but increased competition and potential cooling in both crypto and AI markets may lead to consolidation. Only a few major players are likely to survive long-term.
Synergy Between Cryptocurrency and AI Development
The growth of digital cryptocurrencies is accelerating the adoption of advanced chip manufacturing processes. Companies like Bitmain are not only supporting the crypto ecosystem but also supplying much-needed technology to the AI industry.
Even if the cryptocurrency market declines, the computational demands of AI—such as machine learning, data analysis, and real-time inference—will continue to drive innovation and demand in the semiconductor sector.
Both blockchain and AI rely heavily on computational power. Advances in one field often benefit the other, creating a cycle of mutual technological progress. 👉 Explore more strategies in tech development
Frequently Asked Questions
What is cryptocurrency mining?
Cryptocurrency mining is the process of validating transactions and adding them to a public ledger (blockchain). Miners use powerful computers to solve complex mathematical problems, and the first to solve a problem earns new coins and transaction fees.
Why are GPUs used in mining?
GPUs are efficient at performing the parallel computations required for mining many cryptocurrencies. They offer a better balance of performance, availability, and cost compared to CPUs for certain algorithms.
What are ASIC miners?
ASIC (Application-Specific Integrated Circuit) miners are chips designed exclusively for mining a specific cryptocurrency algorithm. They are much more efficient than general-purpose hardware but lack flexibility.
How does cryptocurrency mining benefit AI?
Mining drives demand for high-performance processors, leading to innovations in chip design and manufacturing. These advances in computational power and efficiency are also crucial for developing and deploying AI applications.
Can the AI industry sustain chip demand if crypto falls?
Yes. AI applications—such as deep learning, autonomous systems, and big data analytics—require immense and growing computational resources. This demand is expected to support the semiconductor industry independently of cryptocurrency trends.
What’s the future of mining hardware?
As mining difficulty increases, efficiency becomes more critical. We can expect more specialized hardware, better energy management, and possibly hybrid systems that can switch between mining and other tasks like AI processing.