BitMine Said It Was ‘Largely Finished’ Buying ETH. Then It Deleted the Line.

BitMine's October 11 filing added that management believes the company is largely finished buying ETH. The October 18 filing removed it — with no explanation, while the company bought another 12,535 ETH.

In the week to October 11, BitMine’s filing added a line: management says the company is largely finished buying ETH. In the week to October 18, that line was gone — with no explanation. The deletion is the most informative thing either filing said.

BitMine weekly releases, October 11 and October 18, 2026
BitMine Said It Was 'Largely Finished' Buying ETH. Then It Deleted the Line.
Udo Keppler, Public domain, via Wikimedia Commons

The Case, As Filed

BitMine Immersion Technologies (NYSE: BMNR) publishes a weekly update. I have been reading them the way you read testimony — not for the headline number, but for what changed between two consecutive statements under no apparent obligation to explain the change.

Here is the pair that matters.

Week to October 11. Crypto, cash, and equity holdings: $16.2 billion. 6,028,946 ETH, marked at $2,544. Staked: 5,067,309 ETH. And in the forward-looking statements — the legal boilerplate at the bottom that nobody reads — a new sentence appeared, citing a management statement that BitMine is largely finished buying ETH.

Week to October 18. Holdings: $15.1 billion. 6,041,481 ETH, marked at $2,365 — its lowest weekly mark since the August 16 update. Staked: 5,067,309 ETH, unchanged. And the sentence about being largely finished buying is gone. The wording on acquisitions reverts to what the filings said through the week to October 4: continued weekly ETH purchases. No explanation is given for the removal.

The company bought 12,535 ETH that week. It bought 12,532 the week before. The buying never stopped. What stopped was the sentence saying the buying would stop.

The market, reading the October 11 filing
BitMine’s biggest buyer is done. ETH loses its bid.
BitMine, October 11
Management believes the company is largely finished buying ETH.
BitMine, October 18
…12,535 ETH purchased this week. (No mention of stopping.)
Me, reading both
The only thing that got withdrawn was the sentence, not the bid.

Cross-Examining the Sentence

I want to interrogate this properly, because two readings are available and they produce opposite trades.

Reading one: the company signalled an exit and then lost its nerve. Tom Lee said on October 7 that BitMine needed roughly 100,000 more ETH. He has said publicly — at Token2049 in Singapore — that 5% is a hard ceiling: BitMine will not keep accumulating past 5% and will not hold more than 5% of total ETH supply. At 4.9%, the company is one or two good weeks from a self-imposed wall. The October 11 line was honest. The October 18 deletion was the market’s reaction arriving, and the company blinking.

Reading two: the line was never a policy, it was a disclosure artifact. Forward-looking-statement language exists to flag risk, and a sentence about finishing a purchase programme belongs there precisely because it is uncertain. Removing it seven days later, with purchases continuing at the same rate, is what it looks like when a company decides a sentence created a misunderstanding it did not intend.

I don’t know which is true, and the filing does not say. But here is what I can establish without guessing at motive: the sentence existed, then it did not, and the buying continued at roughly the same rate through both. Whatever happened, the market’s “BitMine is done” narrative was built on one clause in one boilerplate paragraph that lasted seven days.

6,041,481 ETH ÷ 122.1 M supply = 4.9% · target 5% · ceiling stated as hard, by the company

The Numbers That Didn’t Move

Two figures in these filings deserve more attention than the deletion, and neither is about conviction.

Staked ETH has been 5,067,309 since August 9. Not one weekly update has changed it. Meanwhile the treasury has grown by 12,532 and 12,535 in consecutive weeks. Do that arithmetic and you get a ratio drifting downward: the share of the treasury that is staked falls every week the company buys and does not stake. At $2,365, Lee’s own projection puts annual staking revenue at $309 million on the staked balance, or $369 million if the whole treasury were staked, at a 2.58% seven-day yield. The gap between those two numbers is about $60 million a year of yield the company is carrying rather than collecting.

Cash and marketable securities fell to $582 million — the fourth consecutive weekly decline, from $714 million on September 20. The release gives no reason. Four weeks is not a quarter, and I am not going to build a solvency story on four data points. But a treasury company that is still buying, still not staking its new coins, and drawing down cash every week for a month is a company whose balance sheet is doing three things at once, and only one of them is the thing it is famous for.

Week to Oct 11Week to Oct 18
Total holdings$16.2 B$15.1 B
ETH held6,028,9466,041,481
ETH marked at$2,544$2,365 (lowest since Aug 16)
ETH bought that week12,53212,535
ETH staked5,067,3095,067,309 (unchanged since Aug 9)
Cash & marketable securities$612 M$582 M (4th straight weekly drop)
“Largely finished buying”PresentRemoved
BitMine Said It Was 'Largely Finished' Buying ETH. Then It Deleted the Line.
Julie Ceccaldi, CC0, via Wikimedia Commons

Why the Flywheel Needs the Buying

Here is the mechanism, and it is the part I think most coverage of digital-asset treasuries leaves out.

A treasury company’s equity can trade above the value of the assets it holds. When it does, the company can issue shares, buy more of the asset, and increase the asset-per-share figure — which is the thing that justifies the premium in the first place. The loop sustains itself as long as both halves hold.

treasury_flywheel:
    while mNAV > 1:
        issue equity  ->  buy ETH  ->  ETH per share rises
                                   ->  premium justified
    requires: continued issuance AND continued buying
    if buying stops:
        premium has no mechanism left to defend it
        company becomes: a leveraged holder of 4.9% of supply

Note what the second branch says. A treasury company that stops buying does not become a passive index fund. It becomes a leveraged holder with a stated ceiling, an unrealized loss, and a yield it is not fully collecting. That is a materially different animal, and it is the animal BitMine becomes the week it stops — whether or not it ever says so in writing.

I’ve written about what staking actually pays and what it costs you to be in it before, and the relevant point here is narrow: at a 2.58% yield on $12 billion of staked ETH, the yield is a real business. It is also roughly a rounding error next to the price exposure, which is the thing that moved the headline number from $16.2 billion to $15.1 billion in seven days.

Figures above are from BitMine Immersion Technologies’ weekly releases dated October 11 and October 18, 2026, as reported by CryptoPotato and PR Newswire. Staking yield is BitMine’s own seven-day figure. The 122.1 million ETH supply figure and the 4.9% calculation are BitMine’s.

BitMine Said It Was 'Largely Finished' Buying ETH. Then It Deleted the Line.
Julie Ceccaldi, CC0, via Wikimedia Commons

What Lee Actually Said About the Market

It is worth separating two things that got merged in the coverage.

Lee’s own explanation for recent weakness is macro, not treasury-specific. He warned on October 13 that a slump in South Korean equities could hit crypto, and in Monday’s release he tied reduced risk appetite to concerns over AI funding and model safety, the escalating US–Iran war, anxiety ahead of the US midterm elections, and rising bond yields worldwide. His chain of causation is: AI stocks sell off, that spreads to Korea and Taiwan where those companies dominate the indices, crypto is widely held in all three places, therefore crypto catches the spillover.

That is a coherent account. It is also an account in which BitMine is a victim of the same selloff as everyone else, rather than a cause of it. Which is exactly the ambiguity the deletion created: if the company is finished buying, its absence is a cause. If it is still buying 12,535 a week, its presence is a stabiliser and the weakness is macro. The filings support the second reading and, for seven days, contained a sentence supporting the first.

The Line I’d Leave You With

Herbert Stein put it as economically as it can be put: if something cannot go on forever, it will stop.

A programme that buys every week for fifteen months, targeting a fixed share of a fixed supply, has an ending written into it — and BitMine wrote that ending into its own ceiling. It said 5% is the ceiling. It said it will not go past 5%. It is at 4.9%, and it is buying about 12,500 a week.

So the end of the buying is not a question of conviction or nerve. It is arithmetic, and it lands within a couple of months at the current rate. The genuinely open question is what happens to the premium when the mechanism defending it switches off — and whether anyone holding BMNR for the flywheel has priced the version of the company that exists after the last purchase.

  • Watch the sentence, not the share count. If “largely finished” returns, that is a signal. If it stays absent while purchases continue, the October 11 version was noise.
  • Watch the staked number. It has not moved since August 9. If it finally rises, the company has shifted from accumulating to optimising. If it keeps not moving, $60 million a year of yield stays on the table.
  • Watch cash. Four straight weekly declines from $714 million to $582 million. A fifth, or a stated reason, changes the read on how much room the buying actually has.
  • Discount the “ETH loses its biggest buyer” framing. BitMine bought 12,535 ETH last week. Whatever ended, it was not the buying.

(The End)

Markets

Bitcoin Ended the Day Flat. $812 Million of Leverage Did Not.

2026-10-15 2:55:12

Markets

Is Strategy a Company or a Fund? MSCI's Five Ratios, Read From the Rulebook

2026-11-2 3:13:58

❯
Profile
Cart
Coupons
Search