Bitcoin’s move into the mainstream has mostly been a story about who owns it. Funds, public companies and exchanges now hold a large share of the supply. Far less attention goes to the indust
Bitcoin’s move into the mainstream has mostly been a story about who owns it.
Funds, public companies and exchanges now hold a large share of the supply. Far less attention goes to the industry that keeps the network running.
Saša Čoh, CEO of hashrate marketplace NiceHash, says that gap is turning into a risk.
In a three-part series titled “Letters to Bitcoin Holders,” he argues that the asset’s largest owners rely on its security without directly paying for it.
He also says artificial intelligence is now competing for the inputs that security depends on.
Related: JPMorgan issues its first long-term Bitcoin return forecast
Wall Street’s appetite is visible in the numbers.
U.S. spot Bitcoin exchange-traded funds, which let investors buy Bitcoin exposure through a regular brokerage account, took in $6.3 billion in net inflows in the third quarter of 2026, according to SoSoValue data.
Most of that came in August ($3.5 billion) and September ($2.7 billion). July added roughly $170 million.
The year as a whole looks thinner. Net flows stood at about $985 million year-to-date in early October, after outflows in the first half. The funds held $109.3 billion in combined net assets in early October.
That was down 14.6% from a peak of $128 billion in mid-January. Since launching in January 2024, they have drawn $57.6 billion in cumulative net inflows.
Čoh calls this concentration “what institutional adoption looks like.” His concern lies elsewhere.
Bitcoin is secured by mining, where computers compete to add new blocks of transactions to the ledger and earn newly issued coins plus fees in return.
He notes that two mining pools coordinate roughly half of all block production. Pools are groups where miners combine their computing power and share the rewards. Citing pool data, four pools account for more than 70%.
“Gold holders can ignore the mine. Bitcoin holders are standing on it,” Čoh writes. If gold mining stopped, he argues, gold would only grow scarcer.
If Bitcoin mining stopped, the network could no longer confirm transactions.
The pressure on miners is built into Bitcoin’s design. The block reward was cut in half to 3.125 BTC in 2024 and will be halved again in 2028. Each cut makes miners lean more on transaction fees.
Now, Čoh says, AI is competing for everything miners depend on: power, capital, sites and chips. Mining companies are converting facilities into multi-year computing contracts for AI firms.
Those contracts make “halving-compressed block rewards look unbankable by comparison,” he writes.
The shift is already showing up in the industry’s numbers. Every major publicly traded Bitcoin miner has announced a pivot toward AI data centers, according to a May 2026 analysis by Charles Schwab.
Many have sold Bitcoin from their balance sheets to fund the move, the brokerage noted.
Watch this video to learn more on Bitcoin mining: Explained: What is Bitcoin mining? (6:23)Some of the deals are large.
Schwab’s own modeling helps explain why.
AI inference, which means running already-trained models to answer user requests, earned the highest net revenue per megawatt-hour in its simulation as of April 30. Bitcoin mining came second.
Costs add to the strain. Less efficient miners spend close to $95,000 to produce one Bitcoin, according to Glassnode data cited by Schwab as of May 22.
Bitcoin traded at $77,558 on November 3, at the time of writing.
Mining does keep one advantage, Čoh concedes. It can shut off within moments when the power grid is under strain, while AI data centers need power around the clock.
But that flexibility “lowers mining’s cost floor. It does not raise mining’s revenue ceiling,” he argues.
His central point is about hashprice, the daily revenue earned per unit of computing power.
Hashprice is calculated from the block reward, fees and Bitcoin’s price. Nothing in that formula reflects how much holders value the network’s security.
Čoh describes security as a public good. Every holder benefits, so each one has a reason to let others pay.
Public goods stay underfunded, he argues, until a mechanism lets individual participants buy a defined share for themselves.
He rejects the idea that institutions should simply buy mining companies, warning that this would deepen concentration rather than ease it.
The answer, he writes, is not institutional ownership of mining but institutional access to an open market for it.
His alternative is open hashrate markets, where buyers acquire computing power at market prices without owning machines, and decide where it works.
NiceHash, founded in 2014 and based in Zug, Switzerland, runs one such marketplace.
Purchased hashrate, Čoh writes, does not have to flow back into the large pools; it can be directed to solo block production outside the dominant coordination points, with rewards delivered straight from the chain.
He is explicit that such allocations would not change the overall picture on their own. A single institution’s allocation, he writes, “does not, by itself, move a Nakamoto coefficient.”
The point is a precedent: “Decentralization was never going to arrive as a single decision. It arrives as an allocation habit.”
He is also open about the fact that his proposed fix runs through the kind of market his company operates.
Even so, his letters raise a question large holders have rarely priced: who pays to keep Bitcoin secure as the block reward shrinks?
If you follow our market coverage, don’t miss the Roundtable 100. Updated every Friday, it's a power ranking of the 100 most investable tech stocks, including digital assets. Check out the latest RTB 100 rankings at roundtable100.com.
Related: Harvard’s largest stock holding is a rocket company
