Bitcoin is developing a massive security crisis

For the last 16 years, Bitcoin has functioned as a decentralized payment system in which transactions are verified and recorded into a canonical blockchain by miners. Miners are paid (intrinsic to the protocol) for finding a block via the block reward, and by transaction fees. Extrinsic to the protocol, mining companies can earn money by engaging in energy or other Bitcoin-related markets or can accept payments (or bribes) out-of-band.

Block rewards are paid out to miners approximately every 10 minutes, when a miner finds a block.

The block reward was initially 50 BTC, in 2012 it halved to 25 BTC, in 2016 to 12.5 BTC, in 2020 to 6.25, and last year to 3.125 BTC. It will halve again in 2028. In USD terms, this amount initially shot up exponentially. In the early days block rewards were pennies to dollars, today the block reward is about $350,000. What's important to note about this increase is it only increases over time if the value of Bitcoin doubles faster than the block reward halves. From 2013 to 2017, the peak price went up by 20x, from 2017 to 2021, the peak price went up by about 3.5x, from 2021 to 2025 hasn't even doubled (although to be fair, the cycle hasn't crashed quite yet). In any case, the trend is clear, Bitcoin will not be doubling in value every four years.

It has always been known that we should not expect the value to double every four years forever and ever. But it has always been believed by the Bitcoin community (even going back to the blocksize wars 10 years ago) that transaction fees would eventually replace the block reward so that miners can continue to be paid. Bitcoin prophets over the last 8 years have predicted fees would one day go into the thousands of dollars, perhaps $20,000 a transaction; Michael Saylor has compared Bitcoin transactions to real estate transactions, and other bulls have suggested that banks will be "settling with each other" many times a day and won't mind the massive fees.

(For those unfamiliar, transaction fees are determined by sort of an auction; Each block has finite space – those wishing to transact can post transactions with some amount of fees attached, and the miners take a fixed amount of these and put these in a block. High demand means high fees, low demand means low fees.)

Here's the problem. There is now an overwhelming amount of empirical evidence showing that transaction fees will not pick up the slack. Today, transaction fees are near zero. Historically, transaction fees have spiked up from time to time, but this was usually due to some sort of speculative frenzy. In the earlier days (say 2017) transaction fees became expensive when much of the gambling on Bitcoin's price was on or adjacent to the blockchain: you could buy Bitcoin on a sketchy exchange and then send it somewhere secure. Today, we are in "paper Bitcoin summer" where nearly all of the gambling is through other vehicles, many of which are available through your tradfi Fidelity account. There's another type of gambling that happens, occasionally on the Bitcoin chain, where certain transactions are swapped, (examples; Runes, Ordinals) but these have tanked in popularity and most degens have left for blockchains with slicker gambling UX such as Solana.

For people just using Bitcoin, to, idk, pay each other? There's no such activity. This was the original use case, but now we have Tether and hundreds of other of crypto options (if you want to send remittances or send hard-to-trace payments), not to mention all the easy tradfi options like Venmo and Zelle.

Miners are now competing for revenue that appears to have begun a trend toward exponential decrease. To make things worse, miners have invested heavily over the last few years, and they are all in competition with one another.

Hashprice is a measure of how much miners are paid per hash: Four years ago, a miner could expect to be paid over $4 for producing an exahash; today this exahash will reward the miner about $0.55.

Miners, in turn, have begun to pivot toward AI/HPC in order to make use of the massive electrical infrastructure many have created or acquired to mine Bitcoin.

Things could get wild

Here's where it gets interesting.

Publicly traded corporations now account for almost 40% of the global hash rate up from about 27% a year ago; this is a massive acceleration. Home and hobbyist miners and small miners are being swept away for good as the pubcos obtain new mining equipment which they plug into cheap electricity.

Another massive development that hasn't fazed the Bitcoin community too much: Activist investors have taken large stakes in these pubcos and tried to steer their activities. For example, DE Shaw, a very tradfi hedge fund, took an activist role in forcing Riot to pivot towards AI/HPC.

Mining companies have been encouraged by the market to pivot. They have become just another FOMO play, just like Bitcoin treasury companies and alt coins. As reality started to set in, the pubcos about mining revenue, these pubcos realized they could get a jump in share value by adding some AI/HPC, and this then became another fun degen game, trying to ride the mining pubcos up the AI bubble.

Bitmain has overproduced s21s, their latest s23s are hitting the market. Meanwhile they have competition: Jack Dorsey's Proto miner claims to be very efficient and will add to the available hashrate for miners.

The result will be corporations stockpiling mining equipment at a discount, and with AI as a profitable option to spend electricity on, they could be sitting on latent hashrate that is not used.

The security phase shift

I've predicted for several years that we will undergo a "security phase shift." I had not expected this until around 2031 or so, but it may be much close. This starts when there is slack in the available hash capacity. Throughout most of Bitcoin's history, much of the hashrate has been profitable to run, as prices surged faster than chips could come off the foundries. But this could be changing soon.

Smaller miners are being priced out. In a few years there will be few to none. The publicly traded miners will be doing much of the mining, and subject to their shareholders desire to maximize profits, these will be simply pawns in a larger game, subject to whims of other markets such as AI. As profits from mining decline, the mining division will get merged, sold off, or eliminated, and what will be left will be a few corporations with a glut of mining resources who then realize that in order to maximize profit, they can do things like selfish mining or harmonic mining between AI/HPC and Bitcoin. This will lead to chaos, which will resolve in a massive consolidation into a small cabal who will negotiate anti-competitive agreements, effectively performing a 51% attack to keep the hashrate low and profits maximal.

(Note that some of the following might run into issues with anti-trust. IANAL, so I have no idea if a US court could enjoin a US based corporation from selling hashrate to a Grenada incorporated LLC in response to a home-miner claiming anti-trust standing – a lot of the rules here are quite unclear and I don't have confidence the regulators are equipped. )

1) Selfish mining. This incentive misalignment has long been known, but it was often thought that either miners would never actually do this (it's a "not cool" thing to do), or that no miner would ever have the adequate resources to do this. In order to pull this off, a single miner needs to possess 34% of the hashrate and has to be OK with using the rules to gain an "unfair" advantage.

Then it's just math, by choosing when to reveal blocks, a selfish miner can create enough reorgs so that the competing miners are proportionally more harmed than the selfish miner. When the difficulty readjusts, this miner then has an advantage. This exploit first becomes available when the miner has one third of the hashrate, and becomes more pronounced when they add hashrate. As other miners drop out or join the selfish miners, this snowballs into a 51% monopoly attack.

2)Direct collusion. If there are a few mining firms whose latent (not necessarily deployed) hashrate is more than 50% of the active hashrate, they could agree to deploy less hashrate, except when attacking blocks for non-cabal miners. You can also do this without colluding, engaging in anti-competitive sniping against pools or smaller player. It's not against the rules, and it's easy if you have access to a deeper bench of hashrate: they can't fight back. It's essentially a game of attacking the weak.

3)Harmonic mining. This is a fun one that becomes available when you have some sort of alternate use of your electricity such as AI/HPC. Turn on your miners for two weeks, run them full capacity and then, when the difficulty adjusts, turn them off. Instead go and do inference or AI/HPC whatever with that same electricity. Now the difficulty will have increased, but you're now removing your own hashrate so the result will be a slower block time. You don't care; you are running inference for the AI/HPC money. The other miners who continue to mine are being rewarded with lower hash price. Eventually they will get to the 2016 blocks - because it took them so long the difficulty will now be a bit lower; now you pivot back. You now have a lower difficulty and can grab a bunch of those blocks at a much lower rate (higher hashprice). Rinse and repeat.

This can be done in collusion with other miners but you don't need to actually have some secret phone call; you can just start to do this and others will catch on. The end result will be the effect of punishing all the miners who do not have AI/HPC as an option.

The story at the end of the day is that there will be a few firms left, after the mergers and acquisitions, at some point they will hop on a call and hash out an arrangement by which they all get to keep a share of the revenue. By eliminating competition the competition, profits margins become healthy again.

Where does it go from here?

I see a few options for Bitcoin going forward.

1) Death Spiral. If the market comes to grips with the fact that this is not the decentralized payment protocol that was long promised, some participants leave, and the death spiral begins - as price drops, security woes worsen, leading to further price drops, and so on.

2) Love the Centralization. The second option, which is think is more likely, is that Wall Street and whoever is holding this decides, "who cares?" Tradfi can just override the original ethos, replace it with some cabal who essentially monopoly mines the Bitcoin, expand the chain size and frequency, and the games go on indefinitely.

True Bitcoiners do not like this outcome, but I think it's the most likely. Most of the money now active in Bitcoin is tradfi. Tradfi not only does not care about decentralization, they probably even prefer that Bitcoin is mined by other tradfi. It actually works much better that way. Just cut out the whole decentralized Rube Goldberg machine (Nakamoto consensus is so 2009) and replace it with a few corporations just stamping transactions into an expanded blockchain.

3) Fill in the blank, some catalyst, something something. For years it's been, "well some catalyst might happen" but at this point, we're all still waiting. With all of the consolidations and financial heavy hitters swinging Bitcoin mining around, if the revenue is collapsing there will be massive centralization.