Big Tech’s AI bill comes due - has the sell-off gone too far?

Folks, if you’re holding tech ETFs, or you’ve made a few cheeky bets on a hyperscaler, I get you. It’s uncomfortable out there.

In October, the tech sector was trading at record highs. Barely three months later, it has gone from euphoria to a warzone. Investors haven’t just lost patience with Big Tech’s ballooning AI spend - they’ve developed seemingly negative patience. As JPMorgan analyst Toby Ogg put it:

“The sector isn’t just guilty until proven innocent - it’s being sentenced before trial.”

What makes this even more jarring is where capital has rotated. Walmart (NASDAQ: WMT) now trades near 45 times earnings, while much of the Magnificent Seven sits closer to 25–40 times. Staples are being priced like growth, while growth is being treated like excess.

With Amazon capping a bruising earnings season (Nvidia still reports at the end of February), we spoke to ETF Shares’ David Tuckwell to put the tech wreck in perspective and get his thoughts on how some of the biggest names performed.

This sell-off is less about AI - and more about interest rates

Tuckwell frames the pullback in AI stocks as primarily a macro event, not a referendum on AI’s long-term value.

Major banks, including JPMorgan and Macquarie, have reversed expectations for multiple rate cuts, now flagging the possibility of a hike in 2026 or 2027, a shift reinforced by the appointment of perceived inflation hawk Kevin Warsh as Fed Chair.

Markets, Tuckwell argues, don’t react to where rates are, but to how expectations change. When those expectations rise or become more volatile, long-duration cash flows are discounted first.

“In a hangover market, investors rotate out of growth and toward quality, stability and value-ish names. That explains why Apple and Walmart have acted as ballast while others were punished," he says.

AI capex looks scary - it’s meant to

Across earnings season, one constant was rising AI spend. Microsoft, Google, Amazon and Meta are all leaning into capex at a scale that’s uncomfortable and well ahead of consensus.

But Tuckwell urges investors to look at the big picture and consider moments in time when such scale of spending was necessary to achieve market dominance, such as when railroads laid tracks or car manufacturers built factories before the returns showed up.

"We think AI will follow the same path. Yes, it’s going to be costly. Yes, you need to build the data factories first. But the rewards will ultimately flow to those who are prepared to wait," he says.

Remember, Jeff Bezos once described Amazon as “famously unprofitable” as it sacrificed profits in the 2000s and 2010s to dominate e-commerce. Apple and Netflix followed similar paths - investing billions in designs, devices and content upfront before profits eventually caught up.

Who’s executing - and who’s slipping?

Tuckwell critiqued the performance several popular tech names held by Australian investors directly or via ETFs.

Apple (NASDAQ: AAPL) is the clear standout for Tuckwell in this reporting season.

Earnings crushed expectations, with iPhone revenue of $85 billion versus $79 billion forecast. Margins are approaching 50%, and CEO Tim Cook described demand as “simply staggering.”

But Tuckwell’s real enthusiasm lies in Apple’s differentiated positioning within AI. Instead of building its own models, it's being tactical. For example, by embedding Google's Gemini into Siri.

“And while everyone obsesses over massive, energy-hungry models, Apple wins if small language models take off. By controlling hardware and distribution, Apple controls the most valuable real estate in AI - without needing to build the models itself," he says.

Microsoft (NASDAQ: MSFT), by contrast, was punished despite beating expectations. Earnings came in at $4.14 per share versus $3.97 expected, revenue topped US$81 billion, and Azure growth remained robust at 37–38%. The issue comes back to nearly US$40 billion of quarterly capex rather than performance.

At one point, Microsoft shares fell as much as 12% in what Tuckwell describes as "a harsh sell-off that’s unjustified." “The market was spooked by high capex while revenue guidance lagged very slightly," he says.

But what investors are missing is CEO Satya Nadella’s “AI portfolio strategy.”

“Instead of betting on a single model, they’ve positioned themselves to win across every path - Copilot, internal AI, OpenAI, data centres and Anthropic. It’s like building a portfolio of stocks where one of them must win," explains Tuckwell.

Palantir (NASDAQ: PLTR) delivered a beat-and-raise quarter, but Tuckwell has more sympathy for the sell-off. Its growth increasingly resembles outsourced IT and consulting displacement, particularly within the U.S. government, which has slashed more than 300,000 employees since Trump came into power and effectively redirected that money to Palantir.

It's a great business model... as long as the government is willing to give you more and more contracts, and that's the key risk as well.

“They plug their tech in and leaving them is like changing a tyre on a moving car,” he says. 

“But a lot of recent growth has come from Trump-era government restructuring - and that tailwind won’t last forever."

Amazon (NASDAQ: AMZN) arguably deserved its bruises. Earnings narrowly missed expectations, net income came in almost 20% below forecasts, and capex guidance of up to $200 billion shocked the market. The stock is down over 10% in after-market trading at the time of writing.

“They’re facing a pincer movement,” says Tuckwell. “Walmart is winning on e-commerce distribution, and Google is competing aggressively for AWS.”

Still, he believes Amazon’s dominance in robotics provides long-term optionality that markets may be undervaluing.

Alphabet (NASDAQ: GOOG) once again beat estimates, topping US$400 billion in annual revenue for the first time. Cloud revenue surged 48%, backlog jumped to US$240 billion, and AI demand continues to accelerate. Capex is rising sharply - but Tuckwell believes it’s justified.

But what the market is missing about Google is the opportunity for its autonomous driving business, Waymo.

“If Waymo captures just 1% of the global fleet over the next decade, it becomes a trillion-dollar business - and the market isn't pricing this in," says Tuckwell.

He also notes Alphabet continues to grow operating cash flow by 10–20% annually, underpinning his long-term view that Google could one day become a US$100 trillion company.

Equities

Why Google could go to US$100 trillion (not hyperbole)

View

Meta (NASDAQ: META) delivered what investors asked for: AI spending directly tied to advertising growth.

“The only lingering concern is whether Meta can build its way to the next platform,” says Tuckwell, noting Zuckerberg’s history as an acquirer rather than a builder.

Tesla (NASDAQ: TSLA) is the stock Tuckwell is most cautious on - not because the products are weak, but because of intensifying competition from China.

“We’re seeing BYDs overtake Teslas in many markets,” he says. “That’s because China has subsidised EVs in a way Western governments haven’t been willing to.”

Big Tech versus software: where the real risk sits

The other big story this week was the sell-off in local and international software-as-a-service (SaaS) stocks after Amazon-backed Anthropic's Claude AI platform introduced a new feature that can help legal departments automate processes like document reviews and templates.

Investors are worried about what AI can do and what it will disrupt, and this has drawn a fault line between hyperscalers and software. Names like Intuit, Atlassian, Xero, and Salesforce are down 30–50% from their highs, with concerns about AI agents driving indiscriminate selling.

Tuckwell likens the psychology to the global financial crisis, where one concern in one part of the market was indiscriminately causing sell-offs in another, even if the two weren't related. "In 2008, investors didn’t know where the risk was. Today, rapidly evolving tech creates the same uncertainty," he says.

But he rejects the idea that software is structurally broken.

“Proprietary data with IP protections cannot be disintermediated by agentic AI. That’s not opinion - it’s a legal fact," says Tuckwell.

In a high-cost, high-risk build phase, scale matters. And right now, scale favours the giants, and thus investors may be better off sticking to the relative safety of mega-caps with strong balance sheets and at the top of the AI apex, where attractive buying opportunities are emerging.

"As Warren Buffett once said 'The stock market is a device for transferring money from the impatient to the patient,' says Tuckwell.

"In AI, we have a once in a generation, perhaps once in a lifetime opportunity. And the last thing you want is to look back on this in 10 years, in 20 years time and say 'I got scared by a few headlines.'"

How are you feeling about technology stocks right now?

Let us know your thoughts below!

ETF

ETFS Magnificent 7+ ETF (HUGE)

Global Shares

View