What Ikea Can Teach Us About Software in AI Times

In 1956, a Swedish furniture designer solved a personal inconvenience... and accidentally upended the furniture industry.

Gillis Lundgren, Ikea's fourth-ever employee, was rushing to deliver a new "Lovet" table to a photo studio. The design was a beauty – clean Scandinavian lines, a leaf-shaped top, solid wood legs... the kind of piece you'd see in any high-end furniture store.

But there was one problem. The table wouldn't fit in his car.

Standing in a parking lot, sweating over a stubborn piece of wood, Lundgren had a simple idea: What if I just take the legs off?

A few moments later, the table (and its now detached legs) slid neatly into the trunk.

Ikea soon tried out shipping products in flat boxes. All folks had to do was put the pieces together at home. The new model worked...

Customers no longer had to wait months for a perfectly finished piece. They could take it home that day at a fraction of the cost – as long as they were willing to do a little work.

Nearly seven decades later, Ikea generates more than $50 billion in annual sales... all because one designer realized that, for most people, "good enough now" beats "perfect eventually."

Today, we're discovering the same thing about business software that Lundgren discovered about furniture.

That's because artificial intelligence ("AI") is creating a new do-it-yourself era in software. And it's challenging one of the biggest investment darlings of recent years – the Software as a Service ("SaaS") sector...

AI Is Giving Business Software Its Own Parking-Lot Moment

For decades, big-ticket software has worked like 1950s furniture... It was expensive and slow to deliver. And all the design power was in someone else's hands.

You'd select from a vendor's offerings. Then, you'd wait six to 12 months for engineers and consultants to "assemble" it.

These systems were sold as flawless, fully integrated platforms. But perfection took time.

Now, AI is blowing the legs off that model.

With AI-assisted coding, businesses can get results immediately by doing that fine-tuning in-house. Consider the numbers...

• AI-assisted coding has reduced development time by 55%.
• Prices have fallen from around $20 to less than $1 per million tokens (pieces of text) processed. So AI deployment is much more affordable.
• Projects that once took months to ship now take days.

This is a full-blown reset of how software is built. Why spend months – and millions of dollars – adapting a prebuilt platform? AI can assemble a custom solution in days, for cheap.

Today's businesses are discovering that they don't need the "custom Italian leather couch" version of software. They can get something that works today at a fraction of the price, then swap out parts as needed.

And there are big economic advantages to the do-it-yourself model...

The 'Good Enough' Revolution Is Underway

This isn't just about how software is built. It's also about how it's sold, priced, and delivered.

AI is the new disruptor. But before AI, SaaS smashed traditional software... where you bought a license and owned it forever.

Tech companies moved to selling subscriptions – offering software as a service, not a product. Instead of buying and owning software, you bought a subscription for the right to use it.

That's really good for sellers... It means smooth revenue and recurring customers as folks renew.

But it's not so good for customers. Businesses end up locked into pricey contracts, paying for each employee's subscription. Now, thanks to AI, they have other options.

This is already changing the market.

Last week, most traditional SaaS stocks sold off sharply... while AI-native company AppLovin (APP) surged 20%. AppLovin has rallied nearly 500% in the past year.

Meanwhile, favorites like Salesforce (CRM) and Adobe (ADBE) are trading at multiyear valuation lows. Take a look...

Wall Street is waking up to the fact that the "assembly model" is replacing the "finished platform" model.

The Tech Platforms Powering the Real Software Revolution

In 2025, you want to own the new winners... And I don't just mean the headline-grabbing AI application makers.

The "picks and shovels" of this trend will be just as successful. These are the supporting platforms that let every business do software on its own terms.

We can break the key players into three groups...

The Big Pieces:

These are the companies at the foundation of the next AI wave – think the "flat-pack parts" sold by Ikea. I'm talking about giants like Microsoft (MSFT), Alphabet (GOOGL), and Amazon (AMZN).

These leaders allow organizations to rapidly build, automate, and reconfigure their tools. And businesses can use their core cloud AI platforms to start small, experiment, and scale up as needed.

The Nuts and Bolts:

These are the "hardware stores" of the digital world. They supply the chips and network infrastructure that put AI in the hands of everyone, everywhere. Think of companies like Broadcom (AVGO) and Twilio (TWLO).

The Application Tools:

These companies sell the tools that put it all together. They offer everything from monetization and fintech to scalable, configurable AI solutions... giving companies specialized, out-of-the-box tools instead of one-size-fits-all options. Palantir Technologies (PLTR), C3.ai (AI), and AppLovin are great examples.

The businesses thriving in this new era are empowering customers with easy ways to do software on their own – just as Ikea turned every shopper into a home designer.

But the difference is, the furniture industry had decades to adapt to Ikea's disruption...

Software vendors may have only a few years – maybe quarters – before the "good enough now" model makes their businesses obsolete.

AI is killing SaaS. Make sure you're on the right side... and own the winners.

Good investing,

Josh Baylin