846,000 BTC on One Balance Sheet: The Concentration Risk Behind Strategy's 2,305 BTC Week

Hook

September 27. Two entities. One number: 2,305 BTC.

Strategy added to its treasury. Strive moved inside the top five of the public-company Bitcoin leaderboard. At a $100,000 print, the combined buy is roughly $230 million — noise against a global spot market that turns over tens of billions a day. Traders will read it as scheduled confirmation, not alpha. That first read is correct.

Stopping there is the error.

The same disclosure set carries a second number that almost nobody is pricing. Strategy holds 846,000 BTC. That is 4.03% of the entire 21,000,000 hard cap, sitting on one corporate balance sheet. Roll in every listed treasury company and the aggregate reaches 1,273,000 BTC — 6.06% of supply. Strategy alone is 66.4% of that pile.

A "corporate adoption" narrative is being sold on data where two-thirds of the adoption is one issuer.

Speed is the only metric that survives the crash. Concentration is the metric that survives the boom and detonates in the bust. This is a piece about the second number, because the first one — 2,305 — is already in the price.

Context: What Was Actually Filed, and Which Layer It Lives On

Strip the branding off and the filing is simple. Two balance sheets moved BTC from the float into custody. One is the largest holder in the asset class. The other just crossed into the leaderboard's top five.

Strategy — the software company formerly known as MicroStrategy — runs a capital machine that most equity analysts still describe in the wrong vocabulary. It is not a software firm with a Bitcoin position. It is a Bitcoin accumulator with a software subsidiary attached for reporting purposes. The mechanics are consistent and repeatable:

• Issue equity through at-the-market programs.
• Issue convertible notes at low coupons when volatility pricing cooperates.
• Issue preferred stock with a dividend obligation.
• Convert proceeds into BTC.
• Hold.
• Repeat.

The engine only turns one way. It works when equity is issued above net asset value per share. That premium is the product. BTC-per-share accretion is the output. Everything else — the software, the brand, the conference circuit — is packaging around that spread.

Strive runs the same idea through a different wrapper. Asset-management structure, BTC exposure assembled through vehicles rather than a pure operating-company balance sheet. It now reports 26,355 BTC, roughly 0.13% of supply, and it has displaced a former occupant to take a top-five seat.

Here is the part that matters for anyone who writes or reads protocol analysis for a living: none of this touches consensus. No soft fork. No client release. No mempool policy change. No fee-market change. No change to difficulty adjustment, script, or the UTXO model. Bitcoin Core does not know these entities exist. From the chain's perspective, 846,000 coins sat in some outputs yesterday and sit in some outputs today. Same opcodes. Same validation rules.

What changed is the distribution of control over the keys — and the psychological structure of the market that watches those keys.

I spent four months in 2017 auditing the Hard Hat Protocol's staking contracts as a junior backend developer in Rome. The lesson from that audit was narrow and permanent. The vulnerability was never in the exotic edge case everyone argued about in the design doc. It was in the boring arithmetic on the accounting path — an integer overflow in the staking logic that nobody had tested at the boundary, because the boundary looked fine on a whiteboard. I reported it through GitHub. The patch shipped before mainnet. A potential $2 million loss died in a code review.

Protocol risk lives in arithmetic. Institutional risk lives in the same place. The difference is that the arithmetic runs on a balance sheet instead of in Solidity, and the boundary conditions are financing windows instead of uint256 limits.

846,000 ÷ 21,000,000 = 0.040286. That is 4.03%. It is not an opinion. It is a division problem, and its denominator is the only one in finance that cannot be expanded by a board vote.

The supply side has a second arithmetic layer. Post-halving, Bitcoin issuance runs at roughly 450 BTC per day — approximately 3,150 BTC per week. Against that denominator, this week's 2,305 BTC is 73% of one week's new supply, absorbed by two entities in a single reporting period.

That ratio is directionally real and rhetorically powerful. It is also, on its own, close to useless as a trading signal. I will show why in the next section.

Core: The Four Numbers That Actually Move, and the One That Is Missing

1. The 73% absorption figure is real — and it is the wrong denominator

New supply is not float. The market for Bitcoin is not cleared by miners; it is cleared by every holder who is willing to sell into every bid. Miners contribute roughly 3,150 fresh coins a week to that process, and they are a small, predictable, price-insensitive seller at the margin.

When the treasury companies absorb 2,305 of those coins, they remove the most price-insensitive marginal seller from the book and replace it with a holder whose entire thesis depends on not selling. That is a genuine structural tightening. It is not a bull signal in the sense the timeline treats it.

The ratio is also a sample of n=1. One week. Two thousand three hundred and five coins. The prior weeks did not print the same number, and the next weeks will not either — because the inflow is gated by equity issuance windows, not by a schedule. Any trader building a model on a single week's absorption percentage is curve-fitting to a press release.

What I want is the trailing 13-week absorption ratio, published as a rolling series. Nobody is publishing it. That gap is the actual information asymmetry.

2. 846,000 BTC is a custody concentration, not a market position

This is where the reporting stops being about price and starts being about plumbing.

A quarter of a percent of the float moving between custodians is a footnote. Four percent of the hard cap moving into a single legal entity's control is a system change.

The distinction most coverage misses is hot versus cold. Exchange-held coins are operationally liquid — they can hit the tape in minutes. Treasury-held coins sit in cold storage under institutional custody arrangements. That means they are, for months or years at a time, entirely absent from the order book. This is bullish while the position builds. It becomes catastrophic in reverse, because the exit from cold storage is not a gradual unwind — it is a governance decision, taken under pressure, executed in size.

Four failure vectors, all of them off-chain and none of them visible in the UTXO set:

• Key management. Multi-signature quorum, signer identity, geographic distribution, succession planning. An institutional custodian's controls are better than an individual's, and they are still a single administrative domain.
• Legal segregation. Whether the coins are held in bankruptcy-remote structures, or on the operating entity's own balance sheet where creditors can reach them.
• Insurance. What is actually insured, at what limit, against which perils, and by whom. Custody insurance rarely covers the scenario people imagine it covers.
• Governance continuity. What happens when the executive who structured the position is replaced by a board with different incentives.

I wrote a 2,000-word post-mortem on Terra Luna two days before the collapse. I spent two weeks inside Anchor's yield-generation model before that, and the conclusion was mechanical, not sentimental: the sustainability model broke at the arithmetic, long before it broke at the price. Yield was being manufactured from reserves that were being drawn down by the same cohort of depositors the yield was attracting. The tokenomics had a terminal condition built into them from the first block.

The treasury-company model is not the same animal. It has no algorithmic peg, no reserve curve, no reflexive mint. But it shares one structural property: its sustainability depends on a financing condition remaining true. For Anchor, that condition was deposit inflow. For a BTC treasury company, it is a persistent equity premium to net asset value.

That is the number nobody quotes. And unlike 846,000, it does not appear in the press release.

3. The capital stack is the real technical surface area

If you want to analyze this the way you would analyze a protocol, stop looking at the coins and start looking at the liabilities.

The stack has three load-bearing elements:

| Instrument | Obligation | Failure Mode | |---|---|---| | ATM equity | None — dilutive | Premium to NAV compresses to zero, issuance stops | | Convertible notes | Coupon + maturity, refundable in cash or equity | Settlement in cash forces BTC sales | | Preferred stock | Fixed dividend, senior to common | Dividend coverage requires ongoing capital markets access |

Read that table as a code audit. Each row is a function with a precondition.

The ATM function's precondition is NAV premium. The convert function's precondition is a low volatility-adjusted borrowing cost and an equity curve that keeps the conversion option in the money. The preferred function's precondition is a cash-flow source that can service a fixed dividend — and the underlying asset, BTC, does not produce cash flow.

This produces the reflexivity loop in its honest form:

1. Equity trades above NAV.
2. Issuance at a premium is accretive per share.
3. Proceeds buy BTC.
4. Buying supports price.
5. Price support sustains the premium.
6. Return to step 1.

That loop is not a conspiracy and it is not fraud. It is a mechanical structure with two stable states and an unstable transition between them. Run it backward and it reads identically:

1. Equity trades below NAV.
2. Issuance becomes dilutive per share.
3. Buying slows or stops.
4. The marginal bid thins.
5. Price weakens further.
6. Return to step 1.

The entry into the second loop is the event the market is not pricing. Note that the transition does not require a Bitcoin crash. It requires only that the equity premium close. A treasury company can hold 846,000 BTC and still be unable to buy a single additional coin, because the coin is not the constraint — the spread is.

This is the point where I would normally open a terminal and check whether the structure is already degrading. It is checkable. The premium is observable in the market. The dividend coverage is inferable from filings. The maturity wall is disclosed. None of it requires privileged data.

4. What I would actually monitor, and the code for it

In 2024 I built a real-time flow monitor tracking institutional activity into the spot ETFs — pulling wallet movement and correlating it against price at the daily close. The dashboard published within minutes of a flow print. That speed edge was the product.

The same architecture applies here, with one substitution: instead of ETF creation baskets, you track disclosed treasury balances and derive the rolling absorption ratio and the concentration index. Here is the skeleton I would run against a filings feed:

# treasury_flow_monitor.py
# Input: weekly disclosed BTC balances per public company
# Output: absorption ratio, concentration index, premium flag

ISSUANCE_PER_WEEK = 450 * 7 # post-halving, approx HARD_CAP = 21_000_000

def absorption_ratio(weekly_buys): # ratio of disclosed corporate buying to new issuance return sum(weekly_buys) / ISSUANCE_PER_WEEK

def concentration(holdings): # share of the largest single holder vs all tracked holders return max(holdings.values()) / sum(holdings.values())

def supply_share(btc): return btc / HARD_CAP

holdings = { "strategY": 846_000, "strive": 26_355, # ... remaining tracked entities }

weekly_buys = [2_305] # current reporting period

print(f"absorption: {absorption_ratio(weekly_buys):.2%}") print(f"concentration: {concentration(holdings):.2%}") print(f"supply share (largest): {supply_share(holdings['strategY']):.2%}") ```

Run it against the disclosed numbers and it returns three values:

| Metric | Value | Interpretation | |---|---|---| | Weekly absorption ratio | ~73% | Marginal tightening, single-period sample | | Concentration index | ~66.4% | Two-thirds of the trend is one entity | | Largest holder supply share | ~4.03% | Structural overhang, not a position |

The third row is the one that changes how you size risk. Floors are illusions until the bot sees the spread. The spread here is the NAV premium, and no spreadsheet on the timeline is plotting it.

5. The leaderboard is spreading — slowly, and top-heavily

The structural change in this filing is not the 2,305 coins. It is the rank order.

| Holder | BTC | Rank | Structural Edge | |---|---|---|---| | Strategy | 846,000 | 1 | Scale, mature financing stack, brand reflexivity | | Strive | 26,355 | Top 5 | Asset-management wrapper, structured entry | | Mining operators | Various | Scattered | Natural accumulation from produced inventory | | All public companies | 1,273,000 | Aggregate | ~6.06% of supply |

Strive's entry means the playbook is being copied. That is the first evidence of diffusion, and diffusion is healthy in principle — a trend with one participant is a company, a trend with twenty is a market.

But look at the distribution. Strategy owns 66.4% of the entire corporate stack. The top five, in aggregate, own materially more than the rest of the universe combined. This is not a diversified institutional base. It is one dominant holder, one fast-follower, and a long tail of companies whose positions are small enough that their selling would be invisible and their buying is not decisive.

When someone tells you "public companies hold X% of Bitcoin," ask which company. The answer is usually the same one.

Contrarian: The Market Is Pricing Adoption. The Data Is Pricing Reliance.

The consensus framing of every treasury-company headline is decentralization: institutional capital is arriving, the holder base is broadening, the asset is maturing.

The disclosed numbers say something narrower. Corporate BTC demand is not distributed. It is dominated — 66.4% of it sits under one management team, one custodian relationship, one financing strategy, and one board.

That is not a diversified bid. That is a single point of failure with a press release.

Here is the angle that almost nobody writes, because it is directionally bearish in a market that prefers the other direction:

The 4.03% supply share is a demand signal today and a supply overhang tomorrow, and only the first half is modeled.

Every coin that moves from the float into a treasury leaves the order book. While it stays there, the float shrinks and the marginal clearing price rises. The moment the accumulation reverses — for financing reasons, governance reasons, or a margin call on a convertible — those coins re-enter the float. Buying is scheduled and slow. Selling, when it happens, is neither.

The asymmetry is structural, and it runs the same way every time: slow absorption, fast distribution. The market has spent two years pricing the first leg without stress-testing the second.

A second blind spot concerns the exit mechanism. Spot ETF holders can redeem. The redemption process is disclosed, automated, and daily. A treasury company cannot redeem. It can only sell — a discretionary corporate decision, taken at a time of the board's choosing, disclosed after the fact. The ETF wrapper has a release valve. The treasury wrapper has a decision. For risk purposes those are not equivalent instruments, even though the media treats them as the same trade.

A third blind spot is bear-market mechanics. In a risk-off regime, a treasury company's equity falls harder than BTC does — it is a levered expression of the same exposure, trading at a premium that depends on sentiment. When the equity breaks, the ATM engine stalls. When the engine stalls, the buying stops. When the buying stops, one of the marginal bids that was supporting the price disappears.

This is not a prediction of collapse. It is a description of the structure. The company does not need BTC to crash. It needs the premium to vanish, and premiums vanish faster than prices do. That is exactly how the mechanism worked in 2022, when the equity traded at a discount and the buy program went quiet for a stretch — a detail that gets memory-holed every time the accumulation resumes.

One more thing the coverage skips. Nobody has disclosed the cost basis, the average entry, the coupon schedule, the dividend coverage ratio, or the collateral terms behind any of this. Every one of those is material to solvency analysis. Every one of those is missing from a filing set that markets are treating as a strength signal. You cannot audit a balance sheet with a press release.

Takeaway: The Number to Watch Is Not 2,305

Three things to track over the next quarter, in order of signal quality.

The NAV premium. If the equity trades persistently below net asset value, the accumulation engine has stalled regardless of what any headline says. Watch the spread, not the balance.

The trailing 13-week absorption ratio. One week at 73% of issuance is a data point. Thirteen consecutive weeks at an elevated ratio is a regime. Nobody is publishing the rolling series, which means the traders who build it hold the edge.

The dividend coverage math. Preferred obligations are fixed. BTC generates no cash flow. Coverage depends on capital markets access, and capital markets access depends on the premium. All three are the same variable.

The disclosure that would actually move this market is not another weekly purchase. It is the first filing where the accumulation stops, or where a custodian change, a collateral term, or a dividend shortfall surfaces. That filing will not be framed as a Bitcoin story. It will be framed as a corporate finance story, in a section of the terminal that crypto traders do not read.

846,000 BTC does not care what anyone believes about it. It only cares who holds the keys, who holds the debt, and which covenant matures first. The coin is not the risk. The wrapper is.

Ask yourself a simple question before the next weekly print arrives: if the equity premium closes tomorrow, how many of the remaining buyers in this trade are buying Bitcoin — and how many are just buying the spread?