AI Is Changing the Software Business... For the Better

Editor's note: Earlier this year, the software sector suffered one of its biggest sell-offs in recent memory. Investors got spooked by AI and rushed for the exits. But as our colleague Whitney Tilson explains today, that was an overreaction. While AI will indeed shake up the software industry, the changes it brings will benefit the companies that adapt...


Earlier this year, software stocks lost more than $1 trillion in market cap in one week.

Investors had become convinced that AI would make legacy software companies obsolete. The logic was simple: AI would be able to cheaply write programs that would replace expensive software tools.

Not only that, but if AI agents replace human workers, surely businesses would need to buy fewer software licenses.

So investors dumped their legacy software stocks.

I understand these fears. Agentic AI is impressive. It can complete multiple complex tasks with little human direction. It's only going to get better.

And it's true, some software companies will become obsolete. Companies with simple tools that are easy to replicate are toast.

But the death of software has been greatly exaggerated. Here's why the story is more complicated...

You see, firms that own trusted programs, databases, and workflows are still in a strong position.

I'm talking about software that captures and stores data on things like sales, customer relationships, human resources, and company finances. You simply can't run a business without these systems.

What most folks are missing is that the "brains" of agentic AI – the large language models ("LLMs") like Anthropic's Claude and OpenAI's ChatGPT – won't replace these systems. On the contrary, LLMs use them.

Salesforce (CRM) CEO Marc Benioff agrees. His software is designed to track customer interactions, data, and more. He says the real value of tech is in the "trusted data and workflows" that LLMs sit on top of.

These are the "systems of record" that contain corporate data. And corporate data is like gold: It's valuable, and it needs to be protected at all times.

The companies that own this layer will not only survive the AI revolution... They'll thrive in the years ahead. And AI is part of the reason...

The Software User Base Is Changing

The number of software users is about to explode.

But many of those new users won't be people... They'll be AI agents.

You don't have to take my word for it. Just listen to the man who became one of the richest people on the planet thanks to the rise of AI...

Nvidia (NVDA) CEO Jensen Huang called the software sell-off "the most illogical thing in the world." He doesn't view agentic AI as software. He views it as labor.

In a February interview with CNBC, Huang asked folks to imagine a day when we have robots in every home. The robots won't build their own kitchen appliances, he said. They'll simply read the manual and use what's sitting on the counter.

As Huang explained, "Why rewrite Excel when Excel exists? Just use it."

Because AI agents are much cheaper than human employees, the near future will bring exponentially more software users.

Say you're delivering a presentation on customer retention at a business conference. One AI agent could find the data you need from your Salesforce account and analyze it in an Excel spreadsheet. Another AI agent could create graphics using Adobe Firefly and arrange them on PowerPoint slides.

Another agent could book your flight, hotel, and rental car... order your dinner... and then submit the invoices for reimbursement – probably for review by yet another AI agent.

With AI agents doing the bulk of repetitive, clerical tasks, we should see an explosion in software use. And saving on labor costs will allow companies to afford even more software.

Of course, software companies will have to adjust.

Instead of building software for humans, they'll need to build software for AI agents. They'll also have to rethink how they price their software.

Most software companies charge by the "seat" today. That's the number of humans using the software. In the near future, they'll charge by the number of AI agents using it... or by the number of actions taken, tasks completed, or other outcomes.

Some software companies, like ServiceNow and Salesforce, are already doing this.

Over time, the usage-based revenue from AI agents will overtake the seat-based revenue from licenses for human users. And the software pie will grow much larger.

That's what happened during the last big technological shift.

In the early 2000s, Salesforce introduced the Software as a Service ("SaaS") model. Customers used software in the cloud rather than on their own computers... And companies charged a small recurring subscription fee rather than a large one-time cost.

This made software programs affordable for a lot more users. Legacy software giants adjusted, and software made more money, not less.

The rise of AI agents will have an even bigger impact.

Market-research firm Grand View Research expects the agentic AI market for global enterprise to grow 46% per year through 2030.

Every company is racing to build agentic AI into its business. Software companies shouldn't fear this trend... And neither should investors. It's time to embrace it – and profit.

Good investing,

Whitney Tilson

Editor's note: Whitney believes the AI revolution will be a disruption unlike anything he has seen in his career. We're already seeing it reshape entire industries. But while the upheaval is real, it will also create massive investment opportunities for those who stay ahead of it... And Whitney has developed a brand-new system to help everyday Americans do just that.

Further Reading

Just about every business is talking about AI. But talk doesn't drive results. That's why investors need to learn how to find the real disruptors amid all the chatter.

Between the software sell-off and other market shocks, this has been a stressful year for some investors. But we're in one of the best bull markets of our lifetimes. And history shows we should expect it to continue.

Back to Top