There’s a particular kind of discomfort in watching QQQ climb week after week while your own portfolio quietly drifts the other way. That’s exactly where I find myself writing this update. The index has been up — meaningfully — over the past three months. And yet here I am, in the red.

It would be easy to read that as a failure. I’ve learned to read it differently.

The rally I wasn’t invited to

This quarter’s strength was narrow, and it was almost entirely a hardware story. With the semiconductor shortage biting, the names that ran were the ones holding the picks and shovels — Micron, ARM Holdings, Marvell. If you owned the silicon, you had a wonderful three months. If, like me, you’ve been tilted toward software and the hyperscalers, the market gave you a much colder shoulder.

Two fears are driving that.

The first is software disruption. After Anthropic’s latest breakthrough, a lot of investors turned pessimistic on software-as-a-service — the worry being that if an AI agent can do what a SaaS tool used to do, the moat starts to look thinner. We’re already seeing it in how some of these names trade: sold down not because the fundamentals broke overnight, but because people are quietly asking how relevant the business will be five years out. And markets, as always, hate that kind of uncertainty.

The second is the hyperscaler question. There’s real doubt about how the cloud giants scale their AI when the hardware itself is in short supply. Training costs are enormous, tokens aren’t cheap, and nobody has shown a clean, obvious model for turning all of that spend into durable profit yet. When the path to monetisation is fuzzy, the multiple compresses. That’s the weight my side of the portfolio has been carrying.

So I did the unglamorous thing — I sat still

I’ll be honest: there isn’t a dramatic table of trades to report this quarter. I spent these months travelling, and reading. A lot of reading — books and long-form pieces on where technology is actually heading. I watched Jensen Huang’s Computex keynote more than once. I went through Satya Nadella’s interviews and speeches. Some quarters the best position is the one you take in your own thinking, not in your brokerage account. This was one of them.

A few convictions did firm up enough for me to act on them.

Why I’m starting a position in Microsoft

I added a little Microsoft, and the reason is a thesis I’ve been circling for a while.

The way I read it, Microsoft is quietly trying to own a layer. Their push into their own hardware — Surface-class laptops and desktops running on ARM, paired with Nvidia’s system-on-chip — isn’t really about competing on devices. It’s about sitting above the silicon. The vision, as I understand it, is an orchestration layer powered by AI agents: when an app like Adobe issues a command, Microsoft’s layer decides whether the work runs on the CPU, the GPU, or the NPU. The developer no longer has to reason about which chip to reach for. Microsoft handles that — and, conveniently, the security layer too.

What makes this more than a power grab, to me, is the philosophy behind it. In Hit Refresh, Satya keeps returning to empathy — not just toward people, but toward developers, toward hardware partners, even toward old rivals. The Microsoft that once built fences now works comfortably with Linux. His view is that in this AI era there’s simply too much business up for grabs to spend energy walling everyone out. Grow the pie, partner widely, take your share of a rising market.

That same philosophy is why I don’t think Nvidia ends up as the sole SoC inside the Microsoft PC. If it did, Nvidia would hold an enormous say over the platform — and that’s precisely the kind of single-vendor dominance Microsoft’s posture seems designed to avoid. I’d expect Intel and Qualcomm in that mix too. A few players, kept in healthy balance.

Coming back to Amazon

I also added a bit of Amazon — a name I’d actually sold in the past because I felt it had a long road ahead of it. Going back through the numbers changed my mind.

The North American business has genuinely turned profitable. AWS is still growing. And once you fold in the advertising arm, the picture looks quite different. Judging on just those two engines — AWS and advertising — I think Amazon looks undervalued on a three-to-four-year view. Sometimes the thesis you walked away from is worth revisiting once the fundamentals catch up to it.

Housekeeping: what I trimmed and what I leaned into

A few other moves rounded out the quarter:

• Sold Shopify on valuation. The story is fine; the price was asking too much of the future.
• Exited Lemonade at a good profit. Less a verdict on the company, more that I’d rather not dwell deeply in the insurance business.
• Added more Nvidia. Conviction here is unchanged.
• Added a little more Palantir into the recent tech drop. I’d rather buy these names when fear is doing the pricing.
• Traded Micron — bought and sold for a gain. A nice reminder that the hardware rally was worth participating in, even if only tactically.

Portfolio Update

• Stock Portfolio: ~$1,100,000
• Crypto: ~$43,530
• SRS: $91,900
• Cash: ~$50k
• Unlisted Securities: ~$700,000
• Property Value (Equity Value): ~$493,133

Total net worth= $2.48 mil

The lesson this quarter

A red quarter against a green index stings a little. But I keep coming back to the same idea: the milestone is never the finish line, and a down quarter is rarely the verdict. The market spent these three months rewarding hardware and punishing the software and cloud names for sins they haven’t necessarily committed yet. My job isn’t to chase the part of the move that already happened — it’s to keep sharpening the map and stay disciplined about where the next few years actually lead.

I’ll share more as the thinking develops.

As always, nothing here is investment advice — just one investor’s journey to FIRE by Forty, shared with full transparency. Do your own homework.