This week I did something I don't do very often: I bought three individual shares in one go, and all three went into my SIPP — the pension pot I won't be touching for years. That's deliberate. These are speculative, high-conviction plays with very long time horizons. If they work, they've got decades to compound inside a tax wrapper. If they don't, the positions are small enough that I'll be annoyed but not hurt. Here's exactly what I bought and my reasoning for each one.
First, some context. My SIPP is the 'adventurous' corner of my portfolio. The bulk of my money sits in broad ETFs — S&P 500, global all-cap, a bit of FTSE 100. Boring, reliable, does the heavy lifting. The SIPP is where I let myself take a few more calculated risks with individual shares I believe have genuine long-term potential. The key word is 'calculated'. Every position is small — we're talking tiny slices of the pie — and collectively these individual picks make up less than 10% of the SIPP. The rest is ETFs. I'm not betting the pension on anything.
With that out of the way, here's what I bought this week.
First: a small position in Constellation Energy (CEG). This is the largest producer of carbon-free energy in the United States, operating 21 nuclear reactors that generate about 10% of America's clean electricity. Nuclear is having a moment — not because of environmental policy, but because of AI. Data centres are insatiable. A single AI training run can consume as much electricity as 100 homes use in a year. And unlike solar or wind, nuclear delivers 24/7 baseload power. The AI boom needs power that doesn't stop when the sun goes down. Constellation sits right at the centre of that need.
They acquired Calpine recently, expanding to 55 GW across nuclear, natural gas, and geothermal — a genuinely formidable energy portfolio. Recent earnings were strong, and they reaffirmed full-year guidance. The stock is down from its highs and for a company with this kind of strategic position — nuclear leader, AI demand tailwind, strong cash flows — a pullback interests me. I'm not buying for the next quarter. I'm buying because I think nuclear power is going to be a defining energy story of the next 20 years, and Constellation is the biggest player in the space. The position is tiny relative to my portfolio. If I'm wrong, it's a rounding error. If I'm right, I've got a seat at the table.
Second: a small position in Nebius Group (NBIS). This is easily the most speculative thing I own. Nebius is an AI cloud infrastructure company — they provide the GPU clusters, cloud services, and developer tools that AI companies use to build and run models. They were formerly Yandex's international businesses, restructured and re-listed. Nvidia directly invested into Nebius. Let me say that again: Nvidia — the company that makes the GPUs everyone is fighting over — put its own money into this company. That got my attention.
Nebius is building massive Nvidia supercomputing capacity through the end of the decade. They've signed contracts with Microsoft and Meta. They're being added to the Nasdaq-100, which means index funds will have to buy it. The stock has rallied strongly year-to-date and AI cloud makes up almost all of their revenue — this is a pure-play AI infrastructure company growing at an extraordinary rate.
Now, the risks. This thing trades at a steep multiple. It's burning through cash on infrastructure build-out with enormous capex plans. If AI demand slows, or if there's an overcapacity problem in a few years, this stock could get absolutely hammered. That's why my position is small. I wouldn't put serious money into something this richly valued. But as a small, speculative bet on AI infrastructure with Nvidia's backing and Nasdaq-100 inclusion? I'm comfortable taking that risk in my SIPP where it's got decades to play out. If it goes to zero, I lose a tiny amount. If it keeps executing, I own a piece of the picks-and-shovels play of the AI revolution.
Third: a small position in Cipher Digital (CIFR). This one is the wildcard. Cipher was a Bitcoin mining company — and still mines Bitcoin profitably at their Odessa, Texas facility. But they've been quietly transforming into something much more interesting: an AI data centre infrastructure company. They've secured long-term leases with Amazon Web Services and Google/Fluidstack, with contracted revenue tied to massive planned capacity in Texas.
The first site with AWS is being energised soon. If they hit that timeline, the revenue transformation from speculative Bitcoin miner to contracted infrastructure provider starts to become real. The stock has performed strongly and Bernstein calls Bitcoin miners 'the power landlords of AI' — because they already control the power infrastructure, the land, and the grid connections that AI data centres desperately need. That's the thesis in one sentence.
The risks here are execution risk — delays in site energisation could trigger sharp sell-offs — and the fact that Cipher is still not profitable on an annual basis. This is a bet on execution, not current fundamentals. Small position. Long time horizon. I'm watching whether they can actually get those data centres online on schedule. If they do, I'll probably add more. If they don't, I'll probably sell and move on. For now, it's an option on the AI infrastructure build-out.
So that's the week. Three shares split equally: Constellation Energy (nuclear power for AI), Nebius Group (AI cloud infrastructure, Nvidia-backed), and Cipher Digital (Bitcoin miner turning into AI data centre landlord). All three are connected by a single theme: the massive, underappreciated infrastructure build-out that AI requires. Everyone's focused on the models and the chatbots. I'm interested in the power plants, the data centres, and the GPU clouds that make it all possible.
Are these sensible, conservative investments for a 66-year-old? Probably not by conventional standards. But they're small, they're in my pension (longest possible time horizon), and they're backed by genuine structural trends rather than hype. I'm not recommending anyone follow me into these. They're speculative — all three could lose money, maybe a lot. The majority of my money remains in broad ETFs because that's what I believe in most. These are the satellite holdings around the core — interesting, exciting, and kept firmly in their place.
As always: this is what I did with my own money. It's not a recommendation. Do your own research. All investing carries risk — you can lose money, and past performance doesn't predict future returns. These are particularly speculative positions — concentrated, volatile, and not suitable for most investors. I'm comfortable with the risk because the positions are tiny relative to my overall portfolio and sit inside a pension I won't access for years. Your circumstances, goals, and risk tolerance are different from mine.