Thirty-four coins.
That's all that was left on the tape. Thirty-four bitcoin โ call it $2 million and change at recent prints โ sitting inside the treasury of a NYSE-listed semiconductor company. And then, last week, not even that. Sequans Communications filed the quietest bomb of the third quarter: its bitcoin treasury balance is now zero. Not trimmed. Not hedged. Not rotated into some freshly branded "strategic reserve." Zero.
The headline machines did what they always do. "Dumping." "Abandoning." "BTC strategy scrapped." Three words of panic dressed up as analysis. And look โ I get the reflex. When a company that spent the spring stacking 658 coins walks that number all the way down to nothing, the crowd reads it as a confession.
But I've been on the desk long enough to know the exit is almost always more informative than the entrance. This one has a structure. Let's open it up.
Sequans is not a crypto company. It never was. It's a French-American fabless chip designer โ IoT, 5G, narrowband LTE, the unglamorous silicon that lives inside smart meters and industrial sensors. It lists on the NYSE. It files 10-Qs. Its CEO, Georges Karam, answers to a board, not a Discord.
Which is exactly why it joined the DAT trade.
Digital Asset Treasury. That's the label the sell-side invented for the Strategy template โ take a public balance sheet, issue cheap convertible debt, convert the proceeds into bitcoin, and let the market re-rate you as a leveraged crypto proxy. It worked spectacularly for one company. It worked adequately for a dozen. And then, through this year, it metastasized into a cottage industry of small caps all running the same spreadsheet with different tickers.
Where the yield is sweet, the risk is steep. Sequans walked straight into the steep part.
Here's the timeline as I've reconstructed it from filings and press statements. The treasury stood up in the spring of 2025, accumulating a peak of roughly 658 coins. In July 2025, the company issued convertible notes โ the standard financing rail of the playbook โ and used part of that capital to add to the position. By late June, the stack had already been cut to 314 coins. By late September, it was gone. Full unwind. Convertible notes redeemed in full. No default. No residual debt. No margin call. No drama.
Four months from peak to flat.
Now here's where the coverage gets it wrong, and where the real information lives.
Look at what Sequans actually did and you'll notice it was never a bitcoin trade. It was a financing arbitrage that happened to use bitcoin as the intermediate asset. Read the loop in order: issue convertible notes โ convert proceeds to BTC โ use BTC to redeem the notes โ exit with a cleaner balance sheet. That is not a treasury strategy. That's a repo leg with volatility attached.
Chasing the alpha before the liquidity dries up is one thing. Chasing it with borrowed money, collateralized against a semiconductor company's credit, is another. Sequans ran the second version, executed it, and closed the door behind it. I've picked apart this structure in a dozen other contexts โ the same skeleton shows up in every "borrow fiat, buy crypto, pray for mNAV" deck I've sat through since 2021 โ and the tell is always identical. The coins get treated as collateral, not as conviction.
The coins were never the point. The spread was.
I want to be precise about mechanics here, because "dumping all of its bitcoin" implies panic, and panic implies a forced seller. Sequans was not a forced seller. A forced seller shows up in the tape as a single ugly print โ one block, one counterparty, one crater and a lot of Telegram screenshots. What we saw instead was a scheduled glide path: 658 โ 314 โ 0, spread across roughly five months, with debt redemption completed on the company's own clock. That's an orderly deleveraging, run by someone who knew where the exit was before they ever walked in.
I've seen the moon, now I'm looking for the exit. Most DATs never write that sentence down. Sequans wrote it, signed it, and filed it.
And here's the piece the bears keep missing: the underlying business was improving the entire time. Q2 product revenue came in up roughly 80% year over year. The six-month order backlog more than tripled. When your actual operating engine starts coughing up real cash, the opportunity cost of parking a non-yielding, 60%-drawdown-prone asset on the balance sheet stops being theoretical. It becomes a line item a CFO has to defend in front of a board that answers to shareholders, not to a coin community.
Hype is the fuel, but fundamentals are the engine. Sequans just remembered which one was under the hood.
One more mechanical detail that nobody has flagged, and it matters. There's a timeline inconsistency buried in the disclosure. The company began reducing its position in May, yet the convertible notes that funded the position were issued in July. You can't shrink a stack before the money that built it arrives โ unless there were multiple tranches, or the reporting is sloppier than the press release implies. Either way, it tells me the unwind decision may have predated the market's awareness of it by months. Boards don't move that fast. Boards move that fast when they already decided.
Which brings us to the metric that actually runs this sector, and that almost nobody outside the sell-side desks discusses plainly: mNAV. Market-to-net-asset-value. The ratio of a company's market cap to the market value of the coins it holds.
When mNAV trades above 1, the flywheel spins beautifully. The company issues equity or converts at a premium, buys more coins, lifts the per-share coin count, and the market rewards it with an even higher mNAV. Profitable. Reflexive. For a while, self-justifying. That's the Soros loop translated into a 10-K โ price creating fundamentals, fundamentals creating price.
When mNAV flips below 1, the same flywheel becomes a grinder. Issuing equity to buy coins now destroys value per share. The rational move is to stop buying. The next rational move is to sell and retire shares. And when several DATs reach for that exit at once, you get a reflexive unwind: financing freezes, coins get liquidated, price drops, mNAV compresses further, more sellers appear.
The crowd moves fast, but the ledger moves faster. Sequans just showed the market exactly what the exit ramp looks like โ and it was clean.
Compare and contrast, because the sector is now a tale of two cities. Empery Digital liquidated 1,400 coins for roughly $87 million โ call it $62,000 a coin โ a print that, depending on its basis, smells like a realized loss wearing a strategic-decision costume. Satsuma wound down its treasury vehicle outright. And Strategy itself โ the original, the bellwether, the company whose name became shorthand for the trade โ sold through the summer before resuming accumulation in late August as the price recovered.
That last datapoint is the one I'd tattoo on every DAT bull's forearm. The biggest holder on earth stopped being a pure accumulator and started timing the market. Whether you call it risk management or a tell, it broke the "permanent holder" mythology in public, on a filing, where anyone could read it.
Meanwhile Strive kept stacking. Twenty-five thousand coins, another 1,355 added inside the last week. So no, this is not a sector collapse. It's a sorting. The strong balance sheets are concentrating. The weak ones are unwinding. Matthew effect, crypto edition: to everyone who has, more will be given; and from everyone staring down a maturing convertible, even the last thirty-four coins will be taken away.
Here's where I part ways with the consensus read.
The prevailing interpretation is that Sequans "abandoned" its bitcoin strategy โ a small-cap waving the white flag, a crack in the institutional narrative. I don't buy it. Not because the narrative is strong. It isn't. I don't buy it because the frame is lazy, and it's hiding the actual signal.
Sequans didn't abandon anything. It closed a trade.
The CEO's own language gives it away. Georges Karam described the approach as "prudent and opportunistic." Read those two words carefully, because they're carrying load. "Prudent" is what you say to debt holders. "Opportunistic" is what you say to equity holders. Neither is what you say if you believe bitcoin is a generational reserve asset that belongs on a corporate balance sheet forever.
Opportunistic cuts both ways. It means they'll come back if the math works. It also means they never had conviction to begin with.
Which raises the question nobody on the conference circuit wants to ask: how many of the other hundred-plus DATs are running the same opportunistic script while marketing it as a religion? Because I've been on the feed when these companies announce. The language is always liturgy. "Long-term treasury strategy." "Bitcoin as a reserve asset." "Commitment to shareholder value through digital asset accumulation." Then the mNAV compresses, the convertible comes due, and suddenly everyone's a pragmatist with a spreadsheet and a redemption schedule.
I covered the ICO boom in 2017. I watched projects with white papers made of pure hopium reach nine-figure valuations because they said the right words about decentralization. The DAT era is that same movie with better suits and a real legal entity. Same script, same flywheel, same third act. The ones with actual cash flows survive the rotation. The ones without become next year's cautionary slides.
Here's the genuinely contrarian take, and I'll plant a flag on it: the Sequans exit is bearish for DAT narratives and bullish for DAT discipline. Every clean, orderly unwind that gets filed โ no default, no fire sale, no forced liquidation โ lowers the psychological cost of the next exit. That's the bad news. But it also forces the sector to compete on something other than reflexivity. The DATs that survive this sorting will be the ones whose underlying businesses don't need the coins to be interesting in the first place.
Speed kills, but slow kills too in this game. Sequans chose slow โ five months, three tranches, debt retired โ and walked away intact. That's not failure. That's the trade working exactly as designed.
So what do I watch from here?
One number: Strategy's next 8-K. If the summer selling was tactical and the August buying is real, the narrative survives the quarter. If the "permanent holder" resumes selling into strength, the entire complex reprices within days.
One ratio: mNAV across the small-cap DATs. The first company to disclose a sustained sub-1 print is the first domino, and I want to see how it's framed long before I see how it trades.
And one question, which nobody has answered. If a chipmaker with an 80% year-over-year revenue print decided bitcoin wasn't worth the balance sheet space โ what exactly is the pitch to the ones without the revenue?
The exit ramp is lit. Sequans just drove down it. I'd start counting how many of the others even know where it is.