Transcription
For 2 years, the most profitable software companies in America have gone nowhere, while the NASDAQ has risen approximately 70%. Some of these companies have actually declined in value during a period when the broader technology index produced extraordinary gains.
And I want to explain to you why that gap, which looks like a problem, which looks like a sector being left behind, is in fact one of the most mechanically predictable setups I observe in the current market. The setup is not based on a belief that these companies are about to report better earnings. It is not based on a forecast that artificial intelligence will turn out to be less disruptive than feared. It is based on something more reliable than either of those. The mathematics of extreme positioning. And what happens when a trade that has worked for a long time becomes so crowded that the unwinding of it becomes more powerful than the trade itself.
Let me show you the structure of this because understanding the mechanism is what separates an investor who can act on this from a viewer who simply hears a stock tip and forgets it. Start with the gap itself. Why have software companies underperformed the NASDAQ by roughly 78 percentage points over 2 years? The answer is a narrative, a specific powerful narrative that has dominated institutional trading desks since early 2024. The narrative is that artificial intelligence will commoditize software. The reasoning runs as follows. If AI can generate code, automate workflows, and replicate the functions that software companies charge for, then the pricing power, the customer retention, and the high margins that made software one of the best businesses of the past two decades are all under existential threat.
It is a compelling story. And it has been profitable enough that hedge funds have made approximately $24 billion this year alone betting that software stocks will continue to decline. $24 billion concentrated in a single directional bet sustained over two years.
Now, here is the part that most investors never think through. And it is where the entire opportunity lives. The question is not whether the narrative is true. AI will change software. That much is genuine. The question is whether the people who have made $24 billion betting against software have already extracted most of what that bet has to offer. And whether the position has now become so one-sided that the reversal becomes the more powerful trade.
Consider what extreme positioning actually means mechanically. When a large amount of capital bets that a stock or sector will decline, it does so through short-selling. Borrowing shares, selling them, and planning to buy them back later at a lower price. The critical feature of a short position is its asymmetry. A long investor who buys a stock can lose at most the amount invested. A short-seller faces no such limit. As the shorted stock rises, the losses mount without a ceiling, and the pressure to close the position intensifies in real-time.
When a sector becomes as heavily shorted as software is now, it creates a specific and combustible condition. Everyone who wanted to bet against the sector has largely already done so. The selling pressure from new short positions is largely exhausted because the position is already crowded. And this means the marginal force acting on the price is no longer the weight of new shorts pressing it down. It is the latent pressure of existing shorts who will be forced to buy if the price begins to rise. The sector is in effect a compressed spring. The energy is stored. It requires only a trigger to release.
I have seen this pattern before. Not in software specifically, but in the energy sector before the recent conflict in the Middle East when capital was rotating into oil infrastructure while the consensus was still asking whether energy demand would recover. I positioned ahead of that rotation and generated approximately 60% returns in a period when the broad market produced single digits. The sector was different. The mechanism extreme positioning resolving into a violent reversal was identical.
Let me explain precisely how that release works because this is the part that determines the magnitude and the speed of the move. When a heavily shorted sector begins to rise, the short sellers face a deteriorating position. Their losses grow with each upward tick. Their prime brokers issue margin calls, demands to post additional capital against the growing loss. A short seller who cannot or will not meet the margin call is forced to close the position. And the only way to close a short is to buy back the shares that were sold. That forced buying pushes the price higher. The higher price triggers more margin calls on the remaining shorts. More forced buying. Higher prices again. The process feeds on itself, compressing what might have been months of gradual price recovery into days or weeks of violent upward movement.
This is what happened to Avis Budget Group in 2021. A heavily shorted, unglamorous stock that began rising triggered a cascade of short covering and produced returns of several hundred percent in a matter of weeks before the technical dynamic exhausted itself and the price retreated. The fundamental business of car rental did not transform during that period. What changed was the mechanical pressure of forced buying overwhelming the available supply of shares. The price was driven not by improved fundamentals, but by the structural necessity of short sellers buying to survive.
The software sector is not Avis. These companies are larger, more liquid, more institutionally owned. A squeeze, if it develops, will not be measured in days. It will unfold over weeks or months. But the mechanism is the same. Extreme short positioning creating stored pressure. A trigger releasing it. Forced buying amplifying the move into something far larger than the underlying fundamental change would justify.
Now I want to address the question that the surface level version of this analysis always skips, and it is the most important one. Why now? Why should this compressed spring release in the current environment rather than remaining compressed for another 2 years? The answer connects to the broader macro environment in a way that most commentary on short squeezes ignores entirely.
The narrative that AI would commoditize software was in part a story about growth, about which companies would capture the value created by artificial intelligence and which would be displaced by it. That story flourished in an environment where capital was willing to pay enormous multiples for the companies positioned at the center of the AI buildout and to punish the companies perceived as vulnerable to disruption. It was a momentum trade as much as a fundamental one, but the environment is changing in a way that undermines the conditions that sustained the trade.
Kevin Warsh was sworn in as Federal Reserve Chair in May, inheriting an economy with inflation elevated by the oil shock from the Middle East conflict and a fiscal situation nearly $39 trillion in debt with over a trillion dollars in annual interest that constrains the Fed's ability to support markets the way it has for 15 years. In this environment, the valuation multiples that the market is willing to pay for the most expensive, most crowded AI-adjacent stocks are under pressure because higher for longer interest rates compress the present value of distant future earnings most severely for the highest multiple stocks.
This creates the conditions for a rotation. When the most crowded, most expensive trades come under pressure from a changing rate environment, capital looks for the assets that have been left behind. The sectors where the negative narrative has driven valuations to levels that no longer reflect the underlying business reality. Software having underperformed for 2 years, having been shorted to extreme levels, having been priced as though AI would eliminate its business model, is precisely the kind of neglected sector that benefits when capital rotates away from the crowded trades.
And here is the fundamental flaw in the bearish software narrative that the market is beginning to recognize. The thesis that AI commoditizes software confuses two very different things. The cost of building software and the cost of distributing it. It is true that AI dramatically reduces the cost of writing code. It is not true that AI reduces the cost of acquiring millions of enterprise customers, integrating into the workflows of organizations with 100,000 employees, and maintaining the switching costs that make enterprise software so durable. A company does not abandon the software platform on which its entire operation runs because a cheaper alternative exists. The cost and disruption of migrating is prohibitive. The proof is already visible. Free alternatives to dominant enterprise software products have existed for years. Enterprises continue to pay for the incumbent products because the embedded systems, the trained workforces, and the integration dependencies make switching impractical. The disruption that the bearish narrative anticipated has in important respects already been available, and the incumbents have retained their customers regardless.
When the market begins to recognize this, when a piece of information emerges suggesting that software companies are not in fact disappearing, the short sellers face the deteriorating position I described, and the mechanical unwinding begins.
Let me now show you the specific signals I am watching to identify when the release is occurring because positioning around timing requires watching the right indicators rather than guessing.
The first signal is the price action on the software sector ETFs, the funds that track baskets of leading software companies. These funds have traded sideways and downward for an extended period, repeatedly failing at specific resistance levels where short sellers have been able to maintain their position. When these funds break decisively above their resistance levels on elevated volume, when they reclaim and hold above the moving averages that have capped them, it is the first confirmation that the spring is beginning to release. The funds finding support at higher lows, each successive low above the previous one, is the early structural signal that accumulation is occurring beneath the surface.
The second signal, is the short interest data itself. When short sellers begin closing their positions, not because the fundamental thesis has changed, but because the price movement is creating margin pressure, the short interest ratio declines. Falling short interest in a rising price environment is the clearest confirmation that mechanical forced buying is driving the move, rather than a fundamental re-rating. It tells you the move has its own momentum independent of any specific earnings catalyst.
The third signal, the one most retail investors never track, is the options market. Institutional investors positioning for a squeeze do so through call options, which provide leveraged upside exposure while limiting downside to the premium paid. Unusual activity in short-dated call options on individual software names or the sector funds indicates that participants with significant resources are positioning for a near-term directional move. This is often the earliest signal preceding the visible price action.
Now, let me give you the three specific situations within the software space where I see the most compelling individual setups with the same clarity I would apply to any rotation thesis and with the same insistence that these are positions sized for a specific mechanical move, not buy and hold investments to be married indefinitely.
The first is Commvault Systems, ticker CVLT. Commvault is a cyber resilience and data protection company and the thesis here contains an elegant irony relative to the bearish software narrative. The market has been shorting software on the fear that AI destroys software value. Commvault builds the tools that protect the data on which AI itself depends. As AI proliferates, the volume of data requiring backup and protection grows. And as AI makes cyber attacks more sophisticated and more frequent, the demand for cyber resilience increases rather than decreases. The company generates over a billion dollars in revenue growing approximately 19% annually with enterprise clients across industries and partnerships with leading cybersecurity firms. The stock has been punished alongside the broader software sector down approximately 60%. And is now in a recovery phase with each successive low higher than the last and institutional accumulation visible in the volume patterns. No one enterprise reduces cybersecurity spending because the consequence of a breach is existential. This is a structurally defensible business caught in a sector-wide sell-off.
The second is Expensify, ticker EXFY, and I want to be explicit about the risk because this is a small speculative name with a higher individual failure probability than Comvault. The stock peaked near $50 and has declined approximately 97%. It now trades near a dollar. Let me be clear about something. Cheapness is never a reason to buy. Cheap stocks become cheaper all the time. What makes Expensify interesting is not its low price, but the combination of factors around it. It is an expense management platform with millions of users still generating cash that recently executed a meaningful share buyback. A signal that management believes the stock undervalues the business. It has integrated AI into its platform rather than being displaced by it, building functionality that allows AI assistance to plug directly into expense data. And in a short squeeze environment, the smallest names with the highest short interest often move the fastest and the most violently because their limited share float means that forced buying has an outsized impact on price. This is a clearly speculative position appropriate only as a small allocation sized for its risk.
The third is Mara Holdings ticker M A R A formally Marathon Digital. And it is the purest short squeeze candidate of the three because over a quarter of its tradeable shares are currently sold short. The stock is down approximately 83% from its peak. It began as a Bitcoin mining company, but it is pivoting toward energy infrastructure and AI data center capacity, acquiring energy assets to position itself in the physical infrastructure layer of the AI buildout. The squeeze potential is structural. A stock down more than 80% with a quarter of its shares shorted that experiences forced buying if either Bitcoin rallies or the AI infrastructure pivot attracts attention. The chart shows the same pattern visible across these names. Successive higher lows and inverse structure suggesting a bottom gradual accumulation. It will not move in a straight line. The large number of investors holding losses will sell into every rally creating resistance. But the structural conditions for a violent upward move are present.
Now let me return to the broader principle because the three stocks matter less than the framework they illustrate. The era in which an investor could buy a quality company, hold it indefinitely, and ignore it has been undermined by the speed at which capital now rotates between sectors. This is not a criticism of long-term investing in the broad market. The S&P 500 held over decades remains a sound approach for most investors because the index is self-cleaning and the long-term trajectory is upward. But for the investor selecting individual stocks, individual sectors, or thematic funds, the assumption of permanence is dangerous. The leading companies of one era are routinely displaced by the next. The dominant phone manufacturer of 15 years ago, the most exciting networking company of two decades ago, these were not bad businesses. They were businesses whose moment passed as capital and technology moved on. Holding them indefinitely through the rotation away from them would have destroyed capital.
The software setup is the inverse of that dynamic. It is a sector that capital rotated away from that has been priced as though it's moment has passed and that is positioned for the rotation back. Driven not by a forecast of improved fundamentals, but by the mechanical reality of extreme positioning resolving into a violent reversal in a macro environment that is turning against the crowded trades that have dominated for two years. The unthinkable in this case is not a market crash. The unthinkable is that the most hated, most heavily shorted, most left behind sector in the market over the past two years becomes one of the best performing sectors over the next 12 months. Not because software suddenly became wonderful, but because $24 billion of short positioning has to unwind and the unwinding of an extreme position is more powerful than the position itself.
That outcome will not announce itself in advance. It will begin in the price action, in the short interest data, in the options flow, in the indicators that measure positioning rather than sentiment. And by the time it is visible in the headlines, the investors who understood the mechanism and positioned for the release will have already captured the most favorable entry points of the entire move. The spring is compressed. The pressure is stored. The only variable left is the trigger and the macro environment is supplying the conditions for it now.