Glassnode data shows the first quarterly decrease in hash rate in six years, which is abnormal in itself. Even during past bear markets, miners kept adding machines out of stubbornness because they could still gamble on the coin price. But this time is different; the problem lies in the rising costs that can no longer be sustained.

Currently, mining one BTC costs nearly $90,000, while the market price is just over $60,000. This gap can't be covered by faith alone. Many mining farms are genuinely losing money—either shutting down or seeking alternative ways out. So, the decline in hash rate isn't a sign of bearish sentiment; it's simply unsustainable.
What's even more interesting is the subsequent change: miners are starting to "not rely on coins to survive."
In the past, they exchanged electricity for BTC; now, they exchange electricity for cash flow. Whoever offers more stable payments gets the hash rate.
AI data centers have just met this demand, and the logic is straightforward—
Mining is a gamble on price, while AI is about collecting rent.
One involves high volatility and uncertain returns; the other offers relatively stable USD cash flow. Plus, AI is currently supported by policies, capital, and narratives, making this shift for miners quite rational.
However, many people tend to think in extremes, asking whether AI/Tokens will become the new BTC, which is a bit of a forced analogy.
Back in the day, BTC's rise was about consensus going from 0 to 1, driven by the narrative of currency; now, AI is more about industrial demand—it's a computing power business, not the same thing.
To be more precise, the current change isn't about who replaces whom, but rather:
Miners, who once "depended on coins," are gradually shifting to "selling computing power" for a living.
BTC was their biggest client in the past; now, AI has become a more stable client.
So instead of saying "rushing into AI," it's more accurate to say—
Whoever can provide continuous cash flow, their hash rate will flow there. $BTC