Technology and software stocks have been sold down on concern that artificial intelligence will undercut their business models.
On the Switzer Show last night, Netwealth chief executive Matt Heine argued that his company benefits from AI rather than being threatened by it. Meanwhile, macro strategist Jacqui Fernley made a similar case for the wider market.
So how can investors spot stocks caught up in an AI sell-off?
What’s driving the AI sell-off?
The worry driving the sell-off is twofold.
As AI makes work cheaper and faster, customers may need fewer software licences, which lowers revenue for companies that charge per user. And a smaller competitor, using AI to write code, may be able to build a rival product with far fewer staff and a faster time to market.
Netwealth, the ASX-listed investment platform, was caught in the sell-off. On the Switzer Show, Heine argued that instead of being existentially threatened by it, AI works in his company’s favour. Here’s what he says to look out for, using Netwealth as an example, as Matt Heine laid out.
How to spot an AI sell-off stock
1. Does it charge by the seat, or by the transaction?
Many software companies charge per user, a model known as per-seat pricing: revenue rises and falls with the number of people using the product. For these companies the AI concern is direct. If a customer can do the same work with fewer people, it buys fewer seats, and the software company’s revenue falls.
Netwealth charges on a different basis. Heine said its revenue is tied to the volume of money and the number of transactions on the platform, not the headcount of the firms that use it.
“Rather than being a seat-based business, our business is all based on volume and transactions,” he said. “So, as AI becomes more efficient… our firms and the people that we work with are actually more efficient and putting through more volume and more transactions. We’re actually a net benefactor of that.”
That difference reverses the direction of the AI effect. If AI makes financial advisers more productive, they process more business, and more business on the platform raises Netwealth’s revenue. A per-seat business loses revenue when AI lets its customers do the same work with fewer licences. Heine said the alignment was not deliberate, describing the company as “by chance, as opposed to any particular foresight, incredibly well set up for an AI world.”
The test for an investor is to work out whether a company earns per user or per unit of activity. The two respond to AI in opposite directions.
2. Can it be copied, or is it protected by more than software?
The second concern is that AI lets a small, well-funded team rebuild a product quickly. Heine’s response was that software is only one part of what Netwealth sells. He named four barriers a new entrant would have to clear.
“It’s going to be difficult for one person vibe coding in their bedroom to actually come in and do what we do because there’s a huge amount of regulation and licenses that are required just to operate,” he said. “You then need to be able to build out very broad capability to match what we’ve built over 25 years at the same time as building very deep relationships with in excess of 4,000 advisers and do it at a price point that we’re able to whilst investing heavily back into the product.”
Of those four barriers, only one is code. The regulatory licences, the capability built over 25 years, and the adviser relationships are not things AI reproduces. That’s people-stuff.
Heine said AI widens that gap rather than closing it: “We see the moat that we’ve built over a very long period of time actually getting wider and AI and technology really supporting it in many ways, supercharging what we’re already doing.”
The test for an investor is to identify what actually protects a business from competition, and work out how much of it is software that a rival could rebuild with AI.
Heine put some numbers behind the description. As at the 20 July program, he said Netwealth held more than $135 billion on its platform, with about $30 billion more in its own products, served by a technology and product team of 350 to 400 people and used by more than 4,000 advisers.
A generative-AI chatbot introduced about a month earlier was answering around 19 per cent of incoming queries, he said, and he expects the development team to be at least 30 per cent more efficient over the next 18 to 24 months.
3. Why is the share price really falling?
A falling share price does not by itself show whether a business is deteriorating. According to Jacqui Fernley who also spoke on last night’s show, it could just be sentiment-driven.
She’s been watching US earnings season as well as the local market. She says that US bank earnings, for example, are holding strong despite the fact that gloomy economic stories are a dime-a-dozen at the moment. She also pointed to the ETF SEMI which invests in the global semi-conductor craze, and its recent decline:
“If I was in the semi-ETF, I’m up 112%, so I’m not surprised it’s dropped 15% in four or five days or something,” she said. “So that’s profit-taking.” A share price that falls after a large gain can reflect investors banking that gain rather than a change in the underlying business.
The test for an investor is to check whether a decline is matched by weaker fundamentals, such as rising bad debts or falling demand, or whether it follows a strong run and looks like profit-taking.
