On 9 October 2026, two companies made up 94.5% of the market value in the CoinBucha Listed Bitcoin Index: SpaceX and Tesla. Each of them holds bitcoin worth about 0.1% of its own market value.
So an index meant to follow listed companies that hold bitcoin was, in practice, following a rocket company and a car company. We changed the rules two days later. This is what went wrong, the rule we wrote, and what it did.
How it happened
The index weights companies by market capitalisation. Under the old rules (methodology 1.x), any listed company with a sourced bitcoin holding could be admitted, and the 20% cap on a single company was only a display view: the published level itself was not capped.
Admit two of the largest companies in the world because they hold a small bitcoin treasury, and they swallow the index. Nothing in the data was wrong. The rule was.
The rule we added (R6, dated 11 October 2026)
A company is now admitted only if it is one of two things:
- A bitcoin business (Rule A). At least 50% of its latest fiscal-year revenue, as reported in its own annual filing, comes from bitcoin mining, bitcoin mining hardware, or exchange, brokerage or custody where bitcoin is a material asset. A business built on another chain's token does not qualify, whatever its size.
- A material holder (Rule B). It holds at least 1,000 BTC, and that holding is worth at least 30% of its market capitalisation at the review. Once in, it leaves only if the same measure falls below 20%. A reading above 300% admits nothing by itself and sends the record back for a check.
And the 20% single-company cap now binds in the level itself, through capping factors set at each review.
What it did
Of the 32 constituents, 28 pass:
- 6 on Rule A alone: Coinbase, Hut 8, Riot Platforms, Ionic Digital, Bitdeer, Keel Infrastructure
- 15 on Rule B alone
- 7 on both measures
Four leave: SpaceX (bitcoin = 0.1% of its market value), Tesla (0.1%), Block (1.6%), Core Scientific (1.4%). Their company pages stay in the register, with the reason.
The bitcoin barely moved: 1,163,627 of the constituents' 1,203,813 BTC (96.7%) is still in the index.
Why we re-based instead of chaining
The old series was three weeks old and nobody cited it yet, so we chose a clean start over chaining a series whose composition we now reject. The index is re-based to 100 on 9 October 2026, the last close before the decision. The 1.x series is closed, not revised: no old figure was changed. The whole change is one dated entry in an append-only corrections log.
From now on, a methodology change takes effect at a pre-announced close. This first one took effect on publication, because nobody outside depended on the series yet.
Same companies, two answers
There are two series on the same 28 companies and the same base date:
- LBTC weights by market capitalisation, capped at 20% per company.
- LBTC-H weights by bitcoin held. Strategy holds about 73% of the constituents' bitcoin, so it carries about 73% of LBTC-H.
Neither is "right". They answer different questions: how the listed bitcoin companies are valued, and where the bitcoin actually sits.
Check it yourself
Rule B is computed on every row from three numbers you can see: the company's holding, the BTC-USD reference close for the review date (82,546.32 on 9 October 2026), and its recorded market capitalisation. Rule A cites each company's own annual filing on its row.
- The rules, every threshold and every source: methodology
- The index and its 28 companies: coinbucha.com/listed-bitcoin
- Downloads and the API: coinbucha.com/data
If you want the monthly read (which companies moved the index, who bought or sold bitcoin according to their filings), the free Bitcoin Stocks Monthly Brief is one email a month.
Information, not financial advice.
Top comments (1)
Fixing the cap in the level and re-basing instead of chaining is the right call, and the corrections-log entry makes it checkable. Two things on Rule B that the post leaves open.
With the 9 October reference close of 82,546.32, the 1,000 BTC floor is about $82.5M. The 30% test only lets that floor bind for companies under roughly $275M of market cap (82.55M / 0.30); for anything bigger the percentage is the binding rule. A $1B company needs about 3,634 BTC, so in practice Rule B is the 30% test, and the 1,000 BTC floor is a small-cap filter.
The second is that the 30% entry and 20% exit move with the BTC price and with the company's own premium or discount to its holdings. A drop in BTC lowers the numerator, but a holder trading below the value of its coins can still pass, so the test measures how the market prices the company as much as how material the coins are. Do you review at fixed dates only, or also when BTC moves a set amount between reviews?
Also, if Strategy carries about 73% of LBTC-H, does the 20% cap apply to that series, or is it held-bitcoin weighting uncapped by design?