Which Tech Stocks Are Worth Buying Cautiously in Q4 2026

I’ve spent most of this year telling readers to be selective about tech, and I’ll be honest, that advice has occasionally looked wrong for a quarter at a time before looking right again. That’s the nature of investing through a capex supercycle nobody has a clean precedent for. Heading into the fourth quarter, I don’t think the answer is to avoid tech. I think the answer is to buy it the way you’d cross a busy intersection: eyes open, moving with purpose, not sprinting.

Big Tech is on pace to spend roughly $725 billion on AI infrastructure this year, up from about $410 billion in 2025, with Wall Street projecting north of $1 trillion in 2027. Numbers that large stop being abstract and start being a genuine test of whether the revenue shows up to justify them. Sequoia’s David Cahn has been tracking what he calls the AI revenue gap, the difference between what hyperscalers are spending and what AI-related revenue currently supports, and that gap has been widening rather than closing through 2026. That’s the backdrop for every tech stock decision I’m making this quarter, and it should be the backdrop for yours too.

The Market Has Started Pricing Conviction, Not Just Capex

The single most useful shift I’ve watched this year is how differently the market now reacts to similar headlines. Microsoft posted around $90 billion in quarterly revenue, up 18% year over year, and the stock rose about 8% on the report. Meta posted $60.8 billion, growing faster at 28%, and the stock fell roughly 10%. Same basic story, capex-heavy hyperscaler beating estimates, completely opposite stock reaction. The variable investors are pricing now isn’t how much a company spends on AI infrastructure. It’s whether they believe management can show that spending converting into durable, monetizable demand.

That distinction matters enormously for how I’m approaching purchases this quarter. I’m not screening for “exposure to AI.” Everything in mega-cap tech has exposure to AI at this point. I’m screening for companies whose most recent earnings call gave me a specific, credible answer to the question of where the revenue is actually coming from, not a restated capex number dressed up as good news.

Where the Underlying Demand Actually Looks Real

Cloud infrastructure remains the clearest place I can point to actual, measurable monetization rather than promise. AWS is running at roughly $150 billion annualized with growth around 28%. Google Cloud is near $80 billion annualized and growing a remarkable 63%. Azure’s AI-specific run rate sits around $37 billion, up over 120% year over year. These aren’t speculative bookings or pilot programs, they’re revenue lines showing up in quarterly filings with real growth rates attached.

Contrast that with the enterprise software layer sitting on top of the infrastructure, where the picture is messier. A widely circulated MIT study from last year found that roughly 95% of enterprise generative AI pilots were producing no measurable impact on the bottom line, out of tens of billions in corporate spending. I’ve also been tracking product-level failure rates across the AI tool ecosystem more broadly, and as of this quarter, close to one in ten tracked AI products has already shut down or been acquired out of distress. That’s not catastrophic, but it’s a useful reminder that infrastructure demand and application-layer demand are not the same trade, and I’d rather own the picks-and-shovels layer right now than bet on which application survives.

Why “Cautious” Still Means Buying, Not Waiting

I want to be clear about something, because I get pushback on this every time I write a piece like this one. Cautious does not mean sitting in cash waiting for a correction that may or may not come on any particular timeline. Prediction markets tracking the odds of an AI-related market pullback by year end have actually drifted lower over the summer, from around 26% in June to somewhere in the high teens by late July. Markets aren’t pricing an imminent unwind. They’re pricing genuine uncertainty, which is a different thing entirely, and genuine uncertainty is exactly the environment where disciplined buying, rather than all-in conviction or full retreat, tends to be rewarded over a multi-year horizon.

The concentration risk is real and worth naming directly. The ten largest stocks in the S&P 500 now represent more than a third of the index, and the cyclically adjusted price to earnings ratio for the broader market sits close to 40, well above the long run average in the high teens. That’s not a reason to avoid the sector. It’s a reason to size positions with concentration risk explicitly in mind rather than assuming an index fund alone gives you adequate diversification the way it might have a decade ago.

What I’m Actually Doing With New Money This Quarter

My own approach for the fourth quarter leans toward the infrastructure layer I described above, sized deliberately smaller than I’d want if I were purely chasing the growth numbers, with cash reserved specifically for volatility around the next earnings cycle rather than deployed all at once. I’m treating any pullback tied to capex-conversion skepticism as a buying opportunity in the names with actual revenue evidence behind them, and treating pullbacks in names still running purely on narrative as a signal to wait, not to average down.

Margins matter more to me right now than growth headlines. The Magnificent Seven collectively carry net margins above 25%, nearly double the S&P 500 average, and that cash-generating capacity is the single biggest difference between this cycle and the dot-com era, when plenty of companies were burning cash to chase growth that never arrived. As long as that margin profile holds even as capex climbs, I’m comfortable staying invested. If free cash flow at the hyperscaler level starts turning meaningfully negative for a sustained stretch rather than just compressing, that’s the signal that would actually change my mind, and I’ll say so here when it happens.

For readers who want to track capex-to-revenue conversion and margin trends across these names systematically rather than reading a new headline every week, there are Quant Rating Tool specifically to fold factor data like this into one score you can check before every earnings season rather than reacting after the fact.

The Bottom Line for Q4

Buy tech this quarter, but buy the layer of the stack where the revenue is already showing up in the numbers, not the layer still running on the promise that it eventually will. Keep position sizes honest about how concentrated the index already is. Watch free cash flow, not just capex headlines, as the variable that would actually change the picture. That’s a more boring answer than either “back up the truck” or “get out now,” and I’ve come to believe boring is usually the correct answer in a market this expensive.

Gavin Thorne writes on technology sector positioning and macro-driven equity strategy. This article reflects his personal research process and is intended for informational purposes only. It does not constitute investment advice.