Is software really dead, or just the way you used to sell it?

A 48-hour sell-off in February wiped roughly $285 billion off software valuations, and within days every founder on your timeline had the same hot take: software is dead. Indie hacking is dead. Product Hunt is dead. SaaS is dead. Building in public is dead.

It's a catchy narrative. It's also mostly wrong — or at least, wrong about the part people think is dying.

Here's the thing nobody wants to admit in a viral tweet: the tools got cheaper, the barrier to building got lower, and that's terrifying if your entire business was "I built a thing before anyone else could." But it's not terrifying if you actually know how to run a business. Let's break down what's really happening.

What's actually behind the "software is dead" panic?

The panic has a real trigger. Wall Street watched software stocks get hammered and coined a term for it. The software industry has been navigating a structural re-evaluation of the entire trillion-dollar software ecosystem, with over $1 trillion in market capitalization erased from software stocks in a single week in early 2026.

That's a real number. But market cap swings and "the category is dead" are two very different claims. Analysts at IDC have looked at the same disruption and come to a more nuanced conclusion: SaaS is not dead, but it is metamorphosing, entering a new chapter defined by AI, automation, and outcome-based economics. That's not a eulogy. That's a business model getting renovated.

Even the person who arguably started this whole conversation wasn't saying what people think he was saying. Microsoft CEO Satya Nadella declared on the BG2 podcast that SaaS is dead, a comment that set off a shockwave across the technology industry. But context matters — he was talking about the interface layer, not the underlying business of selling recurring software value.

Forbes summed up the more accurate framing pretty well: SaaS is not dead — the interface is, because anyone can now build anything. That's a completely different problem to solve than "shut it down and go get a job."

Why does it feel like indie hacking got harder overnight?

Because it did — but not for the reason people think. It's not that building got harder. Building got dramatically easier, and that's exactly the problem for anyone whose entire strategy was "build it first."

One indie hacker laid out the shift bluntly on Indie Hackers: AI copilots, agents, and modern infra haven't just made building faster — they've made taste and engineering judgment the bottleneck, which is exactly why half-baked micro-SaaS struggles. The follow-up point is the one that matters most: Micro SaaS isn't dead because demand disappeared — it's dying because the barriers collapsed, and when everyone can build, taste, distribution, and retention become the game.

Think about what that actually means. Five years ago, shipping a working product was a moat. You'd spend three months building, and that alone filtered out 90% of your would-be competitors who couldn't code. Now a competent solo founder with Cursor or Claude Code can clone your weekend project by Sunday night. The code was never the business. It just felt like it was, because it used to be the hardest part.

This is why you're seeing founders publicly torching their own category — the "I'm only making 22K a month, I'm basically broke, SaaS is dead" posts that get quote-tweeted into oblivion by people pointing out the obvious: a founder telling everyone else not to compete with them is not a neutral source. It's a competitive moat dressed up as public service.

Is SaaS spending actually shrinking?

No. This is the part that gets lost in the doom-posting. Look at the actual dollars moving through the industry rather than the stock charts.

Global SaaS spending is still climbing toward $315 billion — the market did not die, the interface changed, SaaS is not dead. Enterprise buyers haven't stopped writing checks for software. If anything, spending is accelerating. Businesses still need tools, and enterprise software spend is accelerating in 2026 at the highest rate in years — the highest absolute rate ever.

What's changing is how that money gets allocated and what buyers expect for it. IDC predicts the pricing model itself is on borrowed time: the traditional per-user, per-month licensing model becomes increasingly obsolete as digital labor replaces manual interaction, with 70% of software vendors expected to refactor pricing around consumption, outcomes, or organizational capability by 2028.

So if you're building a per-seat SaaS tool in 2026 with the exact same pricing logic as 2015, that's the actual thing that's dying. Not the category. The pricing model.

So what should you actually be doing instead of panicking?

Stop trying to win by being first to build something. You can't win that race anymore, and chasing it is a waste of the one advantage AI tooling actually gives you: speed to test ideas, not speed to lock in a moat.

1. Treat distribution as the product. If a competitor can rebuild your app in a weekend with an AI coding agent, your defensibility has to live somewhere else — your audience, your brand, your customer relationships, your ability to actually reach the people who need what you're building. Build the audience before or alongside the product, not after.

2. Pick a foundational AI tool and actually learn it deeply. This is where the "vibe coding" conversation gets practical. Tools like Claude Code have become foundational infrastructure for a lot of solo builders — not a toy, but the thing you're structuring your entire workflow around. Anthropic's official Claude Code documentation covers installation across terminal, IDE, desktop, and browser surfaces, and it's worth actually reading instead of just installing and improvising.

To get started with it:

1. Install via your platform's recommended method (Homebrew, WinGet, or native package managers — npm installation has been deprecated).
2. Run claude inside a project directory to start a session, pointing it at the folder you want it working in.
3. Log in with your Anthropic account, or set an ANTHROPIC_API_KEY environment variable if you want to skip the interactive login.
4. Use it to read your codebase, edit files, run commands, and integrate with your development tools — available in your terminal, IDE, desktop app, and browser.

Whichever tool you pick — Claude Code, Cursor, or something else — betting on one and going deep beats bouncing between five tools and mastering none of them.

3. Build something that solves a real, painful, specific problem — not a wrapper. The tools that survive the "AI made everything a commodity" wave are the ones solving genuine pain points with enough domain nuance that a generic AI agent can't just replicate the whole thing from a prompt. Thin wrappers around a single API call are the first casualties. Depth is the moat now.

4. Get comfortable being a marketer, not just a builder. If your product instinct got you this far, your next skill has to be distribution — audience building, positioning, retention, actually talking to the humans who'd pay you. The founders complaining loudest about SaaS being dead are often the ones who never built a distribution muscle in the first place, and now the thing that used to cover for that gap (a hard-to-build product) doesn't cover for it anymore.

Where can you actually go to build smarter, not just faster?

Beyond Claude Code's docs, a few resources are worth bookmarking if you're trying to build in this environment instead of complaining about it:

• Anthropic's Claude Platform Docs — useful if you're building on top of the Claude API rather than just using the CLI tool.
• Indie Hackers — still one of the better places to read real, unfiltered accounts of what's working and what isn't from people actually shipping.
• SaaStr — for the B2B and go-to-market side of things, especially their coverage of how pricing and positioning are shifting in the AI era.
• Claude Code on GitHub — if you want to see under the hood or contribute to the open ecosystem around it.

The barrier that used to separate "people who can build software" from "people who can't" basically evaporated in the last two years. That's not a death sentence for software businesses — it's a filter. It just moved the test from "can you ship code" to "can you find people, understand what they actually need, and keep them around." Everyone screaming that software is dead is really just discovering that the easy part got easy, and the hard part was always the hard part.

The founders quietly making money right now aren't the ones tweeting about the apocalypse. They're the ones who read the docs, picked a tool, and went and talked to a customer.