‘AI is Eating Software!’ Is this Statement Correct, And What Does It Mean for Tech Investors If It is True?

As the AI boom unfurled, there has been a saying that AI is Eating Software. This is a bold statement to make because if it is true, it means a lot of investors who got rich off the back of the boom in many SaaS companies post-GFC could be about to lose it all.

But is this an exaggeration?

What are the Best ASX Stocks to invest in right now?

Check our buy/sell stock tips

‘AI is Eating Software’! Where did this term come from?

The original metaphor comes from Marc Andreessen, co-founder of Netscape and a prominent VC. In 2011, he wrote “Why Software Is Eating The World,” arguing that software was becoming the backbone of almost every industry. That phrase has since become foundational in how tech bros talk about digital transformation.

The “AI is eating software” line is often credited to Jensen Huang, CEO of Nvidia. He famously said,’ Software is eating the world … but AI is going to eat software’. Huang’s framing is that AI (especially deep learning) isn’t just a feature of software — it’s starting to subsume how software is built.

Ever since then, it has been used ever since. For instance, a more analytical piece in Forbes shortly after (by Tarry Singh, Martijn van Attekum & Jie Mei) took this idea and ran with it, positioning AI-powered tools (like code-generation) as a kind of existential threat to “traditional” software development.

There are skeptics too. For example, Workday’s CEO Carl Eschenbach recently explicitly debunked the “AI is eating software” narrative, calling it “overblown” and saying AI is more of a “tailwind” than a threat for their business.

What does it mean?

Well, it is self-explanatory on one level if you’re a layperson. But if you want to delve down into the nitty gritty, here you go (you’re welcome):

Increasingly, AI (especially generative AI) is being used to write, test, and maintain software. Rather than humans coding every line, AI assists or even produces boilerplate and logic.

Tasks like debugging, testing, and code review (which have historically required human effort) are becoming more automated via AI tools. So, instead of software being hand-coded in the traditional sense, future software systems are becoming “AI-native”.  If this continues, this could significantly reduce the manual labor of software development.

Now this is not to say software developers will be out of a job. But their jobs will evolve in the sense that developers might shift from writing every line to orchestrating AI: designing prompts, supervising AI-generated code, shaping AI output.

For existing software companies (especially SaaS stocks), there’s a risk: if they don’t embrace AI, they may be disrupted by newer, more AI-driven competitors. Conversely, those who embed AI deeply could strengthen their moats by offering more intelligent, adaptive software.

Is this true?

To some extent, it is. The idea that it is a universal apocalypse for software engineering is not true. But the idea that we are seeing real, structural changes in how software is developed, maintained, and scaled thanks to AI is true.

So what does it mean for investors?

AI doesn’t eliminate software companies — it reshapes who wins. The value shifts from companies selling “static” software to companies selling AI-enhanced, adaptive, or automation-heavy software. We recommend investors avoid “frozen in time” SaaS stocks (legacy workflows, rule-based, no machine learning integration). Instead, prioritise companies embedding AI deeply into their product and operations.

And those companies will win, not just in gaining market share but gaining margins. AI can automate costly tasks like support and onboarding, costs that incumbents who don’t update their architecture will need to continue to bear.

So who could win? And who could lose?

Microsoft (NYSE:MSFT) is well positioned. Already it has its own infrastructure and application software deeply integrated with AI. The key with this company is that because of its scale, Microsoft can absorb AI infrastructure costs and monetise via cloud and platform usage.

Snowflake could be another and is arguably one of the pure-play data‑platform names best placed: its Cortex AI is designed to build AI applications on top of customer data. ServiceNow could win too.

This company is very workflow‑oriented (IT, HR, operations), which makes it a natural fit for embedding AI agents or automation.As companies digitise more, the use of AI to automate service management and business processes puts ServiceNow in a potentially strong spot.

As for companies at risk, Salesforce could be one. It has a legacy business model that could be challenged by AI. Yes it is investing, but there is doubt about how much it structurally changes the economics of its business model.

Abobe has been cited as a company at risk. Even though it is developing its own tools, There’s risk that creative or design software becomes more commoditised by AI tools that generate images, design assets, or UX components. There are up and coming competitors like Canva and Figma that are more AI-first.

And one final one, Workday. Legacy HR/finance SaaS may face disruption if AI agents begin to replace parts of what customers currently use Workday for (reporting, planning, workflow automation). Some market skeptics argue that these older enterprise SaaS companies will need to re-architect to remain relevant in the AI era.

Conclusion

It may be a step too far to assert ‘AI is eating software’ because software is not going anywhere. But software will increasingly be AI-first and companies that have AI-centred models will win out ahead of those stocks that don’t. There is a lot of food for thought to investors to digest, but this doesn’t mean it should be a neglected task.