Software: A Glass Half-Empty Analysis
Nilesh Jasani
·
March 18, 2025

When you hear a sudden tsunami warning, the right response is not to parse the warning for nuances. It is to start running for the hills.

The Instant Copyability Problem

We have been talking about the instant copyability issue in software for nearly two years. It is not a subtle shift. It is not an incremental evolution. It is something that has fundamentally altered the landscape. We have a section on our website dedicated to this, with articles going back to the early days of the argument. Our last note laid it out clearly (https://geninnov.ai/s/SLVPGv), with tables that made the case in an almost irrefutable way.

Yet, we often get the response: “Yes, but there are nuances.” There always are. There are reasons why some players will navigate the shift, why some others will even thrive or some who will not come under pressure at all. But at the aggregate level, if the overarching argument is right, then the software sector as a whole faces far more headwinds than tailwinds. It is a sector undergoing a radical overhaul. 

Before looking for nuances, the first step is recognizing whether one truly believes the overarching argument or not. And if one does, the next question is: Has anything been done in response before picking up on nuances to look for the brighter side or the reasons not to change?

When There’s a Tsunami Warning, the First Thing to Check Is Whether There’s Been an Earthquake

The most basic step in assessing a tsunami risk is to check whether there has been an earthquake. If there has, then it does not matter that there was no tsunami yesterday, or the day before, or for the past 50 years. The question is whether something has changed that forces us not to habitually use the history of no tsunami yesterday

In software, many professionals still cling to history. Analysts, especially, flinch from signals that indicate a quake. When valuations fall or earnings compress, the default response is to look backward. The simple analysis is that if AI spending rises, AI software spending also must rise, at least proportionately because historically, software companies’ revenues rose faster than technological spending growth.  

But this ignores the core issue: Has something fundamental changed that can be equated to an earthquake?

The tables in our last note made it clear: Software companies can no longer count on product or feature announcements staying technologically unique for more than a few weeks. That has not been happening at all in the AI era so far. Maybe some believe that this is temporary, but the burden of proof cannot be mere quoting of history. 

Alternatively, if few product or feature ideas can stay unique, the analysis must change. The historic trajectory of software spending on the back of new products or features is irrelevant. The AI era does not automatically mean that all software companies will flourish, even when they all have so much more to offer. The argument that software firms will make more money simply because they have more valuable features to offer ignores the fact that selling on the back of more features alone is no longer easy. This is not just a theory. It is already playing out in a competitive landscape that has shifted beneath many Microsofts, Adobes, and Salesforces.

The reluctance to acknowledge structural change is not just a habit of analysts. It is a reflex across the software sector. Most established companies seem to assume their distribution networks, installed bases, and incremental product improvements will keep them safe. But in an era where software features can be replicated instantly, the threat is not just a lack of growth.

There were times when software companies were bleeding because of copyability through piracy. Now, the copyability is more instantaneous and more vicious.

When You Want to Respond to a Tsunami Warning, The First Thing You Do Is Begin to Run

There is always a chance that a tsunami warning is a false alarm. Or that the wave is mild. Or that it will miss your specific location. All of that may be true. But the first reaction should not be endless debate about probabilities—it should be to move.

In software, the equivalent of “moving” is to reassess the fundamental processes by which companies and investors evaluate the sector. A venture investor, a public market investor, a corporate strategist—all must ask whether their frameworks today are any different from what they were before the instant copyability era. If their process remains unchanged from a few years ago, they are not preparing for the new era. They are either consumed debating whether there is any need to change or stuck undecided like deer in front of headlights.

For example, if one is still evaluating software companies based purely on the features and products they announce, then one has effectively changed nothing to factor in the lack of lasting uniqueness of any new idea. If a company is still judged on how many AI enhancements it adds rather than whether those enhancements provide a defensible competitive advantage, then the underlying framework is stuck in the past.

This applies not only to investors but to technology companies themselves. Many still assume that their buyer relationships, sales teams, and installed bases will provide insulation. They assume that layering AI onto existing software will be enough. But the competitive landscape is shifting too fast for those old defenses to hold, including relationship trusts, security, or data proprietariness. 

This is not about immediate portfolio overhauls. In venture capital and private markets, rapid change is not possible. But even in those spaces, like for investors in public markets, how one invests ahead must be different compared to how one invested before.

Once You Begin to Run, Then It’s Time to Think About How High You Need to Go

Once the movement starts, then it is time to consider nuances.

Machines doing coding are far from perfect. Coding automation is nowhere near as good as many had once feared—or hoped. There are still advantages in relationships, in distribution networks, in proprietary data. There are ways for software companies to survive and even thrive.

But recognizing nuances should not be an excuse for complacency. Take a typical software company that assumes it will retain customers because of existing relationships. It adds AI features, tweaks its product offerings, and assumes that its historical strengths will carry it through. This thinking ignores the core shift: The number of competitors and alternative solutions is exploding. The inertia that once protected incumbents is eroding fast.

The same mistake happens at an individual level. Many software developers, when confronted with the idea of AI-driven coding, focus on what machines cannot do. This is comforting but ultimately backward-looking. The real question is: What can machines already do, and how quickly is that improving?

The situation is worse for those impacted by the macro trends. At an aggregate level, software automation is already reshaping employment and value creation in the industry. The fact that some individuals will adapt and thrive does not change the broader or average reality. For example, at a country level, if a nation’s economy is deeply tied to software exports, the macro shift matters far more than any individual counterexample. Macro defenses cannot simply be everyone doing higher value-added, programming work. 

Once You See the Wave, You Must Decide How to Survive

As a long-only public market fund manager, this writer understands how difficult it is to accept when an industry faces secular pressure. When family, friends, or colleagues bring up the headwinds, the instinct is to turn defensive—to compare our industry to one that is struggling even more, to argue that these things are cyclical, to point to outliers who are still thriving.

At GenInnov, we surely assume that we can find ways to beat the odds stacked against long-only investing—whether from ETFs, higher-margin alternative products that distort intermediary incentives, or the ease with which investors now manage their own portfolios. All of this is natural. But we cannot wish away the challenges by pointing to a time when the industry was booming.

We have to attempt things differently. Whether in product design, portfolio construction, partnership arrangements, or value-addition, we must create better odds for ourselves. Doing what worked when times were good and expecting a different outcome is not a strategy.

For those in software, the changed landscape is not yet as obvious as it is in my industry. But the signs are already there. Instant copyability is a far deeper structural shift than the copyability issues of the pre-SaaS piracy era.

So yes, recognizing nuances is valuable. But it should never come before first accepting that the ground has shifted—and that the industry is facing a level of disruption that demands a clear-eyed risk assessment.

Related Articles on Innovation