A conviction-plus-income kind of top-up day in the SIPP today. Two individual shares I've been adding to for a while — Meta and Palantir — and one dividend ETF that quietly does its job in the background. Different holdings, different rationales, but they all share one thing: I already own them, I already believe in them, and today I wanted just a little bit more of each.
Let me walk through each one in turn, starting with the biggest name and working down to the income generator.
First: more Meta Platforms (META). I've written about Meta before — the advertising cash machine behind Facebook, Instagram, WhatsApp, and Threads, with over 3 billion daily active users across its family of apps and a digital advertising duopoly alongside Google that shows no sign of cracking. But the reason I keep adding isn't the advertising business — it's the AI infrastructure build-out happening underneath it. Meta has been pouring tens of billions into AI data centres, custom silicon, and open-source large language models (Llama). They're building infrastructure that doesn't just power better ad targeting — it builds a moat. When your AI models are trained on the engagement data of 3 billion people and your recommendation engine gets better every time someone scrolls, the competitive advantage compounds. The advertising business funds it all — a cash flow machine generating tens of billions in free cash flow annually that can be reinvested into AI without breaking a sweat. At the same time, the stock has had a strong run recently and some might say it's expensive. I'm not arguing it's cheap. I'm arguing that the long-term value of the AI infrastructure they're building — and the data moat that feeds it — is probably underappreciated by a market still fixated on quarterly ad revenue numbers. This top-up was modest. I already own Meta across the SIPP and the ISA. Today was just about nudging the SIPP position a little higher. Not because I think I know where the stock is going next week — I absolutely don't — but because my conviction on the five-to-ten-year story hasn't changed.
Second: more Palantir Technologies (PLTR). Palantir is one of the most polarising stocks I own. Some people see a data analytics company with deep government contracts, an expanding commercial business, and a genuinely differentiated technology platform. Others see a consulting firm dressed up as a software company with a valuation that assumes perfection. I see both, and I've sized my position accordingly — small. But I keep adding, and here's why. Palantir's core product — Foundry for commercial customers, Gotham for government and defence — is not easy to describe in a sentence, which is part of why the market has such a hard time valuing it. It's an operating system for data: it ingests messy, siloed information from across an organisation, builds a unified data model, and lets decision-makers ask questions and run scenarios in real time. It's not a dashboard. It's not a BI tool. It's closer to digital infrastructure for complex organisations — the NHS during COVID, the US Army for logistics, BP for energy operations, Airbus for supply chain optimisation. Once embedded, it's incredibly sticky. The US government alone represents a deep, multi-decade revenue stream that most software companies would kill for. On the commercial side, Palantir has been growing rapidly, and their Artificial Intelligence Platform (AIP) — launched in 2023 — has been a genuine catalyst. AIP lets organisations deploy large language models on top of their own proprietary data inside Foundry's secure environment, which solves the biggest enterprise objection to AI: data privacy. You get the power of LLMs without shipping your sensitive corporate data off to a third-party API. The bear case is real — the stock trades at a premium valuation, stock-based compensation has historically been high, and the government business, while sticky, doesn't grow as fast as commercial. I'm comfortable with those risks because the position is small and my time horizon is long. If Palantir executes on the commercial opportunity and AIP becomes a standard piece of enterprise AI infrastructure, this could be a genuinely transformational business. If it doesn't, I lose a small amount of money in a position I sized to be survivable. Today's top-up reflects modestly increased conviction after watching the AIP rollout and commercial customer growth over the past quarter. Nothing dramatic — just a nudge.
Third: more VHYL — the Vanguard FTSE All-World High Dividend Yield UCITS ETF. I wrote about this one in detail recently when I topped it up alongside VGOV and VAGS, so I won't repeat the full thesis. But in short: VHYL is the steady income generator in my SIPP, holding around 1,500-2,000 stocks globally screened for above-average dividend yields. Lots of financials, energy, utilities, consumer staples, healthcare — sectors that tend to pay reliable dividends rather than chase hypergrowth. The ongoing charge is 0.29%, the yield is north of 3%, and inside a SIPP wrapper those dividends are free of UK tax. Why more today? Because dividends inside a pension compound tax-free, and every unit of VHYL I accumulate now represents a stream of future income that will keep paying out quarter after quarter regardless of what the growth side of the portfolio is doing. It's not flashy. It won't double in a year. But it quietly builds income-generating capacity in the background, and over a multi-decade horizon, that capacity makes a material difference to what the pension can eventually provide. Today's top-up was small and routine — just keeping the allocation where I want it.
Stepping back, the combination of Meta, Palantir, and VHYL in today's top-up session tells you something about how I think about the SIPP. Meta is conviction in a dominant business with an AI tailwind — a bet on continued digital advertising dominance and AI infrastructure compounding. Palantir is conviction in a controversial but potentially transformative data platform — a smaller bet with higher upside and higher risk. VHYL is none of those things — it's just the quiet income engine, slowly building a stream of tax-free dividends that will keep flowing for decades. Growth, speculation, and income. Three different jobs. Three different risk profiles. All within a SIPP that remains overwhelmingly indexed to broad market ETFs. That's the balance I'm aiming for — enough conviction to stay engaged, enough income to stay grounded, and enough indexing that I don't have to be right about any of it for the pension to do its job.
As always: this is what I did with my own money. Not a recommendation. All investing carries risk — you can lose money, and past performance doesn't predict future returns. Individual shares carry more risk than diversified funds. Palantir in particular is a volatile stock with a premium valuation, significant stock-based compensation, and a business model that is not straightforward to evaluate. Meta faces regulatory risk across multiple jurisdictions, potential changes to digital advertising tracking, and the possibility that its enormous AI spending doesn't deliver the expected returns. Dividend-paying stocks can cut or suspend their dividends, and high-dividend strategies can underperform the broader market for extended periods. A SIPP is a pension — you cannot access the money until you reach the normal minimum pension age (currently 55, rising to 57 in 2028). Tax rules can change. Do your own research and consider seeking professional advice where appropriate.