Every company is about to become a software company.
For fifteen years, people have said every company would become a software company. For most companies, it never happened. Now it will.
In 2011, Marc Andreessen wrote that software is eating the world.1 In 2015, Microsoft's CEO, Satya Nadella, said every business would become a software business.2
For tech companies, it came true. For almost everyone else, it didn't.
The reason was simple: custom software cost too much. A big bank could build its own systems. A 300-person distributor could not. So most companies did the next best thing. They rented software, one app at a time.
You can see the result in any office. A CRM here. An accounting tool there. A dozen spreadsheets in between, and the rest in someone's head. None of it was built for how that company works, so the company bends to fit the software. Companies use only about half of the software licenses they pay for.3
That is about to change.
Building software just got cheap
The cost of running AI fell more than 280-fold in under two years. Stanford's AI Index found that the price of a model as good as the first ChatGPT dropped from $20 per million tokens in late 2022 to 7 cents by late 2024.4
And AI now writes much of the code. By late 2024, more than a quarter of all new code at Google was written by AI, then checked by engineers.5 In Y Combinator's winter 2025 class, a quarter of the startups had codebases that were 95% written by AI.6
For most of software's history, the hard part was writing the code. Code is no longer the expensive part.
But buying AI is not the same as using it
Almost every company now uses AI somewhere. Very few are getting paid for it.
McKinsey's 2026 survey found that 80% of leaders say AI has made their people more productive, but only about 6% of companies get 5% or more of their profit from it.7 Boston Consulting Group found the same split: just 5% of companies get real value from AI at scale, and 60% get little or none.8
The 5% are pulling ahead. BCG found they grow revenue 1.7 times as fast as the laggards, with 1.6 times the profit margin.8
What do the winners do differently? They don't buy a better model. Everyone can buy the same models. They rebuild the work around it. In McKinsey's 2025 survey, the top performers were about three times as likely as everyone else to have redesigned their workflows around AI.9
That is the work most companies skip. They give their people a chat window and hope. Gartner expects more than 40% of AI agent projects to be canceled by the end of 2027, because of rising costs, unclear value or weak risk controls.10
Your data is the moat
Here is the part most people miss.
When anyone can build software, software stops being the edge. The edge is what the software is built from.
Your data. Your customers' history. The way your best people make decisions. The rules nobody wrote down. No vendor has those. No competitor can buy them.
Software built from those things is yours alone. A generic app makes you more like everyone else. Your own system makes you harder to copy.
Why the engineers have to come inside
This kind of software can't be built from a distance. Someone has to sit with your team, learn how the work really gets done, connect your systems, and test the AI on your real cases before anyone trusts it.
Palantir built its business this way, and coined the name: forward-deployed engineers, who work inside the customer until the system works. It is paying off. Palantir's U.S. commercial revenue grew 149% year over year in the quarter that ended June 30, 2026.11
Now the rest of the industry is copying it. Job postings for forward-deployed engineers rose more than 800% in the first nine months of 2025.12 OpenAI and Anthropic set up their own deployment teams. In June 2026, Amazon Web Services put $1 billion into one.13
The biggest AI companies have learned what their customers keep finding out: the model is not the product. The deployment is.
Why sooner beats later
A system like this compounds.
The first workflow teaches it how you work. The second builds on the first. Each month it holds more of your data, more of your decisions, and more of what your best people know.
A year in, it has a year of your company inside it. A competitor who starts then is not a few months behind. They are a year behind, and the gap keeps growing.
That is why the companies that start sooner will be ahead. Not because the technology is rare. Because time is.
- Marc Andreessen, "Why Software Is Eating the World," The Wall Street Journal, August 20, 2011. Link
- Computer Weekly, "Satya Nadella: Every business will be a software business," March 2015. Link
- Zylo, 2024 SaaS Management Index. Link
- Stanford HAI, AI Index Report 2025. Link
- Sundar Pichai, Alphabet Q3 2024 earnings remarks, October 29, 2024. Link
- TechCrunch, "A quarter of startups in YC's current cohort have codebases that are almost entirely AI-generated," March 6, 2025. Link
- McKinsey, The State of AI in 2026: On the Road to ROI, as reported by FM magazine, September 2026. Link
- Boston Consulting Group, The Widening AI Value Gap, September 30, 2025. Link
- McKinsey, The State of AI in 2025: Agents, innovation, and transformation, November 2025. Link
- Gartner prediction, June 2025, as reported by Outlook Business. Link
- Palantir Technologies, Q2 2026 results (SEC Form 8-K). Link
- Financial Times, as reported by Fast Company, 2025. Link
- The Next Web, June 30, 2026. Link