Because the adjustment window is not real-time-based, but block-based: every 2016 blocks, an adjustment is done - with the obvious caveat being that the duration it takes to mine 2016 blocks is based on the global hash rate. Also, the downward adjustment is limited to (IIRC) halving the difficulty at max.

So, given the number of miners drops too far too fast, Bitcoin may end up in a situation in which the global mining power is too low to even reach the end of the adjustment window in a realistic timeframe, which in turn triggers even more miners to shut off their equipment and sell their coins, which lowers the price even more and lengthens the timeframe, which in turn...and so on. This situation is known in Bitcoin circles as the "chain death spiral", and it has been a purely theoretical thing - until now, at least.

> Because the adjustment window is not real-time-based, but block-based: every 2016 blocks, an adjustment is done ...

This is definitely true. Luckily for miners there are two other blockchains their mining equipment can make use of which don't suffer from the slow difficulty adjustment. They are both version of Bitcoin Cash (BCH and BSV) and they both have 10-block adjustment periods. This makes them much more resilient to sustained periods of depreciation.

Miners probably have a bit of a sunk cost fallacy, since there's a fixed cost associated with acquiring the hardware. Plus, if they're still bullish on bitcoin, the short term loss is probably worth it in their eyes.

I think you nailed it, although I don't think miners even need to fall for a fallacy for your argument to work: some major miners have made big capital expenditures (efficient hardware, cooling, etc.) to keep variable costs low over the long term. Even if the price drop means they'll never mine themselves out of debt, it could still be the rational thing to do for them to keep the machines on, as long as their reduced variable cost is still below than the market price.

You're implicitly assuming that bitcoin will retain value. If the value of bitcoin halves every 6 months and the cost of mining halves every 7 months, bitcoin will never* be worth mining again.

* Of course it wouldn't be a smooth decline so there may be short periods where it occasionally becomes viable.

Where is the author pulling their mining costs from? I'd love to see their source. Mining costs are mostly concentrated in purchasing the equipment and setting up a factory. Once that's done, your day to day expenses are electricity and labor.

Many of the biggest Bitcoin mining operations are in places like China, near hydro-electric dams, where the electricity cost is extremely low. Labor is also cheap. Since they've already paid the setup costs, they are not going to be shutting down even at $1000 BTC.

There have already been hash rate wars. BCH has tried sending BTC into a death spiral before, by luring over miners to their blockchain. It worked, for a bit. Eventually, transaction fees went up on BTC, and miners were incentivized to return to BTC.

The average block time is 10 minutes for Bitcoin. Hash rate adjusts every 2016 blocks. That's 14 days on average. If the hash rate dropped 50%, it's 28 days.

The way the bitcoin protocol is designed, the mining difficulty and rewards only adjust after N blocks have been mined. It doesn't respond in real time to the hash rate on the network.