The code does not lie, but it can be misunderstood. When NTT Data's chief researcher, Professor Wang Jiange, recently warned of a bubble in AI compute, I found myself reading his arguments through a different lens—not as a commentator on Nvidia, but as a trader who has watched similar narratives unfold in crypto mining. His core thesis: that a new mathematical theory could slash compute demand by millions of times within three years, rendering today's hardware giants obsolete. Replace 'AI' with 'Bitcoin mining' and 'Nvidia' with 'ASIC manufacturers,' and the pattern becomes eerily familiar.
Context: The Mining Landscape in 2025
Bitcoin mining has become a industrial-scale arms race. The top three ASIC producers—Bitmain, MicroBT, and Canaan—control over 90% of the market, with gross margins hovering above 60% for the most efficient models. Global hashrate has surged past 600 EH/s, driven by institutional capital and the post-halving anticipation of price appreciation. Yet behind the scenes, the same bottlenecks are appearing: power grid constraints in Texas and Kazakhstan, transformer lead times stretching to two years, and the quiet emergence of self-mining companies designing their own chips (e.g., Block's 3nm ASIC project). Wang's warning about 'physical constraints' applies here too. The difference is that in crypto, the 'bubble' is not just about valuation—it's about the sustainability of the energy-intensive arms race.
Core: The Flawed Analogy of 'Mathematical Revolution'
Wang argues that a new mathematical framework could reduce AI compute demand by millions of times. In mining, the equivalent would be a breakthrough in consensus algorithm—say, a shift from Proof-of-Work to a Nakamoto-style DAG that requires negligible energy. But this confuses two things: the efficiency of a consensus mechanism and the underlying security guarantees. Bitcoin's security depends on physical work; no mathematical shortcut can replace the cost of producing a valid block without breaking the game theory. The so-called 'New Math' for mining is a mirage. The real improvements come from hardware efficiency (e.g., the 7nm to 3nm transition) and power supply optimization, which yield single-digit percentage gains per generation, not orders of magnitude. In the past five years, ASIC efficiency has improved roughly 2x, not 1,000,000x. Wang's prediction lacks empirical grounding.
Contrarian: The Real Risk Is Not a Revolution, but an Erosion
Trust is earned in drops and lost in buckets. The fundamental threat to ASIC dominance is not a new math tool, but the gradual fragmentation of the mining ecosystem. Large miners are increasingly verticalizing—designing their own chips, securing power purchase agreements, and even building their own pools. This mirrors the trend in AI where cloud hyperscalers are building custom chips. For Bitmain, the danger is not that mining becomes obsolete, but that its margins compress as the customer base consolidates and negotiation power shifts. Meanwhile, the 'storage' thesis in Wang's article—which he applied to chips like ChangXin Memory—has a parallel in crypto: decentralized storage networks like Filecoin and Arweave. They are not immune to the mining downturn; if Bitcoin's price crashes, the entire crypto ecosystem contracts, and demand for storage tokens dries up. The 'safe haven' narrative is a trap for the unwary.
In the silence of the dip, the weak hands break. But the strong hands use the dip to accumulate. The takeaway for miners and investors is not to bet on a paradigm shift that almost certainly will not arrive in three years. Instead, focus on positioning: reduce exposure to leveraged ASIC manufacturers, increase allocation to low-cost power assets, and hedge with short positions on overvalued mining stocks. The bubble in mining hardware is real, but its bursting will be a slow bleed, not a sudden collapse. The code—the Bitcoin protocol—does not lie. It will continue to adjust difficulty, and the market will find its equilibrium. The real question is whether you are prepared for the volatility that comes before the calm.
