No, AI is not about to kill the software industry

Hello again, and thank you, as always, for spending time with Fast Company’s Plugged In.

In a remarkably influential 2011 Wall Street Journal op-ed, Netscape and Andreessen Horowitz cofounder Marc Andreessen declared that software was “eating the world.” From entertainment to commerce to transportation, he argued, startups that were about code at their core were disrupting many of the world’s most deeply entrenched businesses. That was just the beginning, he warned: “Companies in every industry need to assume that a software revolution is coming.“

Fifteen years later, we know that some of the disruptors Andreessen cited—such as Zynga, Groupon, and Skype (RIP)—did not, in fact, eat the world. His larger point, however, played out much as he predicted. Software really does run everything these days. And many of its purveyors are among the most successful companies in the world.

Recently, however, Wall Street has been spooked by the possibility of another sea change in the making: AI might be on the verge of eating software. The sudden leap forward in the capability of software-writing LLM tools such as Anthropic’s Claude Code has investors worried that the corporate behemoths presently making tidy profits by selling subscription-based software—particularly for enterprise customers—might find themselves unable to compete with apps coded by AI for very little cost.

This theoretical collapse of the software industry is known as “The SaaSpocalypse,” a name I hate but can’t quite avoid acknowledging. (I promise not to bring it up again.) It’s reflected in the stock performance of such seemingly robust companies as Workday (down 35% year to date), Adobe (-26%), Salesforce (-25%), Autodesk (-21%), and Figma (-19%). On February 23, after Anthropic published a blog post touting Claude’s ability to modernize software written in the 66-year-old COBOL programming language, IBM—COBOL’s kingpin for most of that time—saw its biggest one-day stock drop in more than a quarter century.

Investors are right to expect that AI will radically change software as a business in the coming years. The evidence is already here, in the form of developments such as Block—the parent company of Square—announcing on February 26 that it’s terminating 40% of its 10,000 employees. Explaining the brutal reduction, CEO Jack Dorsey contended that AI will allow a smaller team to accomplish more and do it faster, and said he was getting ahead of an inexorable industry-wide trend. What happens next remains to be seen, but Block will surely never be the same.

Still, Wall Street’s apparent belief that AI spells bad news for today’s software titans is premature, and possibly just misguided, period. It’s certainly heavy on vibes rather than hard data: Monday’s dip in the S&P 500 apparently stemmed in part from a dystopian imaginary June 2028 memo published by Citrini Research. Laying out a sweeping nightmare involving AI crushing the U.S. economy, it name-checked specific companies such as DoorDash and Zendesk as being incapable of competing with AI-infused apps and agents. Well, maybe, though even the document’s authors admitted they were “certain some of these scenarios won’t materialize.”

In a little over two years, it will be possible to assess what Citrini got wrong and right. For now, it remains equally possible to imagine futures in which 2026’s software-based kingpins aren’t mowed down by AI, even if the technology’s coding chops will continue to improve indefinitely rather than hitting a wall.

For one thing, the software business isn’t solely about writing software. It requires selling it—sometimes in the form of hefty annual contracts—and supporting it when things go wrong. It will be difficult for AI (or even most AI-savvy startups) to take on these tasks outside of the human-powered infrastructure that major software companies have built, often over decades.

In Sun Microsystems cofounder Scott McNealy’s memorable phrase, enterprise customers like having “one throat to choke”—someone with the bottom-line responsibility of making them happy. They wouldn’t get that by vibe-coding their own in-house replacements for major apps, or buying them from a tiny company offering look-alike equivalents. Instead, they have a powerful incentive to keep doing business with companies that have already shown an ability to deliver.

People who use AI to write their own apps might even develop a newfound appreciation for all the ways software suppliers make their lives easier. For instance, last April I wrote about the note-taking app I’d vibe-coded for my own use, and said I’d put it together in a week. What I didn’t know at the time was that I’d spend the next 11 months fiddling around with new features, squashing bugs, and stressing over the fact that I—not Apple, Google, or Notion—bear responsibility for the app’s security and data integrity. I’d do it all over again, but because it’s been great, mind-expanding fun, not because it’s saved me money or time.

It’s far too early to conclude that existing software giants won’t use AI to grow even more dominant. After all, they have considerable resources to throw at that challenge, and deep knowledge of the industries they serve. AI could be a potent accelerant to their growth, or just a way to slash costs by reducing human headcount. But there’s little evidence it’s on the cusp of figuring out how to build and market products humans will find compelling without plenty of guidance.

Even as the technology puts pressure on software companies—say, by introducing enough competition that it’s tougher to endlessly raise prices—they might be intrepid enough to find a new path forward. IBM, for example, isn’t short on AI savvy of its own; if the company can’t find a way to make money from customers wanting to modernize COBOL-based platforms, it’s IBM’s own fault, not Anthropic’s.

Yes, history is full of sobering case studies of once-mighty software companies that got overwhelmed by technological change. In the 1990s, for example, the PC’s shift from the text-based DOS to the graphical interface of Windows was ruinous to big names such as Lotus, WordPerfect, and Ashton-Tate, none of which bet big enough on Windows early enough. Their miscalculation was unquestionably Microsoft Office’s gain.

But it doesn’t always pan out that way. In the following decade, Office faced a similar threat as productivity migrated to internet-based tools. When Google launched products such as Docs and Sheets, stuffed them with innovative features, and offered them for free, observers thought that might be terrible news for Microsoft. Not so: The company reacted skillfully enough that Microsoft 365, as it calls Office in its current form, is bigger than ever, to the tune of $95 billion in revenue last year.

In Silicon Valley, it has become fashionable to tell workers that the only way to remain relevant is to embrace AI rather than fear it. As Nvidia CEO Jensen Huang puts it, “You’re not going to lose your job to an AI, but you’re going to lose your job to someone who uses AI.” The same principle applies to today’s software companies. They’re not going to be killed by AI—only by other companies that are better at seizing the opportunities it offers than they are.

You’ve been reading Plugged In, Fast Company’s weekly tech newsletter from me, global technology editor Harry McCracken. If a friend or colleague forwarded this edition to you—or if you’re reading it on fastcompany.com—you can check out previous issues and sign up to get it yourself every Friday morning. I love hearing from you: Ping me at hmccracken@fastcompany.com with your feedback and ideas for future newsletters. I’m also on Bluesky, Mastodon, and Threads, and you can follow Plugged In on Flipboard.

More top tech stories from Fast Company

If technology could bring traffic fatalities down to nearly zero, why not embrace it?
What the elevator can teach us about self-driving cars. Read More →

Anthropic’s autonomous weapons stance could prove out of step with modern war
The Pentagon is demanding that the AI company remove the safety guardrails from its AI models to allow all lawful uses. Read More →

Is Apple about to debut a new iPhone camera feature?
What is ‘variable aperture’ and why you should care. Read More → 

 AI can write now. What happens to reporters?
If bots can reliably draft copy, ‘something big’ might be happening to the job of a journalist. Read More → 

Apple killed Dark Sky. Now its creators are trying again with a new weather app
Acme Weather brings back the team behind the cult-favorite forecast app, with new features designed to show uncertainty. Read More → 

15 incredibly useful things you didn’t know NotebookLM could do
From managing meetings to maintaining your car, Google’s Gemini-powered research tool can provide all sorts of eye-opening revelations. Read More →

View the full article