On September 25, Sequans Communications sold its remaining 314 BTC. The French semiconductor company now holds zero cryptocurrency. Its press language calls this a return to "core business" โ IoT modules and software-defined radio. The market shrugged. At prevailing prices the notional is somewhere in the low tens of millions of dollars, roughly 0.0000149% of Bitcoin's 21 million supply. Against daily spot volume that is a rounding error. Anyone reading this as a supply event is reading the wrong number. The filing language is clean. The mechanics are not.
The number that matters is the company's mNAV โ market value of equity divided by the market value of the Bitcoin it held. Sequans spent two years telling shareholders it was a Bitcoin treasury company. It exited when that framing stopped being accretive. That is not a crypto story. That is an equity arbitrage closing.
Sequans is a small-cap fabless semiconductor house. It designs baseband chips and modules for LTE-M, NB-IoT, and 5G IoT applications. Revenue is lumpy, gross margins are thin, and the balance sheet has carried real operating losses for years. It is not a company with discretionary capital.
It became a "Bitcoin treasury company" in the 2024โ2025 wave. The template came from MicroStrategy. Issue equity at a premium to net asset value, convert the proceeds into BTC, let NAV per share accrete, use the accretion to justify a higher multiple, repeat. The flywheel is elegant on paper. It has one hard dependency: a persistent premium. The moment the market prices a treasury company at or below the value of the coins it holds, issuance becomes dilutive. The machine runs backward.
Sequans was never MSTR. It lacked the convertible-debt access, the index inclusion, and the retail bid. Its premium was thinner and more fragile. A small-cap treasury strategy lives on the edge of that premium every single day.
By 2025 the template had spread to software firms, biotech shells, and a long tail of sub-$500 million caps. Most of them shared Sequans' problem: a premium built on narrative rather than access to capital. When the narrative cooled, the premium compressed. When it compressed, the model broke.
I spent early 2024 building a flow-attribution model for a Geneva fund, reconciling reported spot Bitcoin ETF inflows against exchange reserve deltas and whale wallet clustering. The lesson from that work applies here: when an entity sells, the useful information is never the size. It is the plumbing.
The sale of 314 BTC was almost certainly not executed as a market order. A position that size, from a company with no trading desk, typically routes through an OTC desk or a single negotiated block. OTC desks absorb the inventory and hedge elsewhere. The visible footprint on spot exchanges is close to nil. The real footprint lands days later in the desk's own delta management, spread across venues and time.
Sequans disclosed no transaction hash. No custody venue. No execution venue. No counterparty. From a chain-analytics standpoint, that is a data vacuum. I would need to cluster the company's known treasury addresses โ if they were even self-custodied โ and trace the outflow. Many small-cap treasury companies custody with a prime broker or a qualified custodian, which means the coins never touched an address the company controls. The chain shows movement. It does not show intent. That gap is where most retail analysis of treasury sales goes wrong.
Follow the gas, not the hype. In treasury-company analysis the equivalent is: follow the custody, not the press release. A disclosed sale is the last step of a decision that was made weeks earlier, in a boardroom, driven by a ratio nobody puts in the headline.
What I can reconstruct is the arithmetic. A treasury company's accretion depends on a single ratio. When mNAV is above 1, every share issued adds BTC per share. When it drops below 1, every share issued destroys BTC per share โ and the rational board response is to stop issuing, halt purchases, and monetize the reserve to fund operations. Sequans did not announce a pivot away from crypto as philosophy. It announced a liquidation. Those are different events, and the filing language blurs them.
Here is what I would track if I had the mandate. Exchange reserve deltas on the days bracketing September 25 โ a 314 BTC block that bypasses OTC and hits spot would produce a measurable, if small, uptick in inflow and a compression in bid depth. Then OTC desk inventory flows into and out of cold storage. And any subsequent 13F or equivalent disclosure from counterparties. Absent all three, the market is trading on a press release, not on data.
Then there is the cost basis, which was also not disclosed. Sequans likely accumulated through the 2021โ2022 institutional wave, when Bitcoin traded in the $30,000โ$60,000 band. If the disposal landed near cost, the sale is balance-sheet-neutral. If it landed below, it is a realized loss on a non-operating asset โ and in a bear tape, that is the kind of line item that forces a board's hand. Alpha hides in the margins. Here the margin is the gap between a company describing a sale as strategy and a company describing a sale as survival.
The reflexive read is that a corporate holder dumping BTC is bearish. I disagree with the framing, and the disagreement is structural rather than directional.
The ETF bid and the corporate-treasury bid are not the same animal. ETF flows are mandate-driven and price-insensitive. A model portfolio allocates 1% of an equity sleeve to a Bitcoin ETF and rebalances on schedule. That bid is sticky and roughly flow-elastic.
The corporate-treasury bid is reflexive. It exists only while the issuer's equity trades above the value of its coins. That makes it procyclical by construction: it accelerates into strength and evaporates into weakness. Sequans exiting is not a signal about Bitcoin's price. It is a demonstration of how the corporate bid behaves. Correlation is not causation โ and here, correlation is barely even correlation. 314 BTC does not move a $2 trillion asset. What moves is the behavior pattern it reveals across dozens of similar small-caps.
There is a second, less comfortable read. This may be good capital allocation. A chip company deploying shareholder capital into a volatile non-operating reserve, while its core R&D needs funding, is a governance question. Code does not lie; people do โ and people describe a forced sale as a "strategic realignment" every quarter.
Stop watching the price and start watching the mNAV board. Over the next several weeks, the signal is which other small-cap treasury companies print below-1 mNAV and whether their filings disclose disposals. Follow the filings, not the tweets. The next 314 BTC will not be announced. It will be inferred. The number is small. The pattern is not.
