Tech Stocks: Oversold or Obsolete?

Software stocks might say the same today.

Europe. Nasdaq. Hang Seng. It doesn’t matter where; the market has thrown out anything remotely connected to software-as-a-service (SaaS) businesses.

The fear is that AI will make all software obsolete.

The ASX hasn’t been immune. Xero has seen falls of 15.9% in a single session. WiseTech dropped 10.7%. TechnologyOne fell 10.5%.

The S&P/ASX 200 Technology Index is now down nearly 40% since mid-September.

Source: TradingView

[Click to open in a new window]

Its RSI has hit 19, more oversold than the pandemic, more oversold than the GFC. Only the aftermath of the dot-com bubble is comparable.

This week, JPMorgan called it what it is: an overreaction.

The same market that bid these tech stocks into lofty valuations has now turned around and decided that AI can do everything SaaS can, and more.

In my view, both extremes failed to reflect fundamentals.

The reality has been that US earnings this quarter have been solid so far. Analysts still expect software sector earnings to grow 16.8% this year.

So what’s driving the panic?

The Vibe Coding Hype Machine

Part of the answer is social media. X.com is flooded with the narrative.

Every second post shows someone building an app with Claude Code, Clawbot or OpenAI in minutes.

The logic is basic enough: if AI can write code, why pay for software?

The narrative being sold is seductive, but like many stories, overly simplistic.

Traditional business applications, the market consensus might argue, are essentially databases with built-in business logic.

C.R.U.D. is the acronym that is thrown around to simplify their function. Create. Read. Update. Delete.

In this view, a banking app or HR software is just a bunch of CRUD actions. You read that data, update a balance, add sick days, etc.

All these actions will supposedly collapse into a new ‘agent era’.

In this world, AI becomes the organising layer. Excel isn’t a spreadsheet anymore; it’s a canvas that AI uses like a scratchpad.

Your calendar becomes another environment for AI to orchestrate through. The business logic moves out of individual apps and into a centralised AI that updates multiple backends simultaneously.

It sounds revolutionary. And parts of it are real.

But here’s what the hype misses.

First, writing the app (or orchestrating it) is the easy part. Maintenance, security patches, compliance updates, and integrations with other systems are a never-ending commitment.

Few companies want that burden when it’s not their core business.

Second, software companies aren’t selling code. They’re selling products.

Those products come with support, with compliance, with service-level agreements. This is why companies don’t just run open-source software, even when it’s free.

They want accountability, they want to focus on what they do best, and not some other problem.

Third, software companies will be the biggest beneficiaries of AI-written code in the short term. They’ll use it to ship features faster, cut development costs, and widen margins.

The agent era doesn’t eliminate software vendors. It empowers them.

Yes, over a longer timeline, companies will likely consolidate as AI completes more tasks.

But the expectation that all software will be replaced by AI in the coming years completely fails to account for today’s business needs.

And that brings me to my next major point.

Not All Software Is Equal

The market is treating software like a monolith. It’s not.

The key distinction is between horizontal and vertical software.

Horizontal software serves broad, general-purpose functions across many industries. Think spreadsheets, email clients, and generic customer relationship managers (CRMs).

Vertical software, by contrast, is built for specific industries with specialised workflows. Things like hospital patient management, freight logistics, and power grid controls.

Horizontal tools are more vulnerable. They’re easier to replicate, require less domain expertise, and face lower switching costs.

Vertical software has deeper moats. Things like regulatory requirements, industry-specific compliance, and workflows so embedded that ripping them out would be catastrophic to operations.

Mission-critical software costs typically account for less than 0.5% of a customer’s total revenue. But if it breaks, the knock-on effects are enormous.

Would a hospital really risk switching to an unproven AI system to save a few basis points?

This distinction matters for investors looking for opportunities.

The same point holds for finance software providers on the ASX, like Hub24 and Netwealth.

Yes, agentic AI coding tools could theoretically let financial advisers build their own systems or platforms. But these are highly regulated products with compliance and government obligations.

Yet they’re all facing the same selling pressure.

The Opportunity

Here’s what struck me most about this selloff so far. There’s been no fundamental deterioration.

No company has reported AI eating into their revenue. No earnings have been slashed due to customers cancelling contracts.

What we’ve seen is a sentiment shift, pure and simple.

That means there are many stocks within the carnage that should be on your watchlist today.

But let’s also be honest, some of these stocks deserved to come down.

Many were trading at over 100x earnings despite not delivering anything extraordinary.

Those kinds of multiple requires a long line of willing buyers happy to pay up. When sentiment shifts, those buyers vanish.

Just take Xero, Wisetech, and Technology One’s valuations. All three have seen their price-to-earnings ratios halve in a matter of months.

Source: Fiscal.ai

[Click to open in a new window]

But here’s the thing. Multiples expand and compress. That’s how markets work.

The good news is that most mid-to-large cap tech stocks have real intrinsic value. They’re profitable. They’re cash-flow positive. They have solid track records of growth.

This isn’t the post-pandemic tech exodus where loss-making concept stocks were obliterated.

At some point, the sector will bottom. Growth rates will suddenly look attractive again. And multiple expansion will return.

The question is timing. Of course, that’s always the hardest part. I don’t have an answer for you there.

For now, the path of least resistance may still be down.

But high-quality software names with multi-year contracts, regulatory moats, and deep customer integrations have been caught in the crossfire.

For patient investors, that’s an opportunity.

The fear is real. The selloff is overdone.

Time to build your watchlist.