Ubs Group 13F Portfolio

The biggest dollar adds are almost embarrassingly on-the-nose: Microsoft, Apple, Alphabet, and the semi equipment duo Lam Research and Applied Materials. This is an explicit statement that the next leg of AI returns will be captured by cloud-scale software and the manufacturing tooling behind the silicon, not by stretching further out the risk curve.

• Microsoft: An extra $1.06B and an 8.6% bump in shares is a strong endorsement of its position as the operating system of AI in the enterprise. With a gain vs. average cost of +87.2%, they’re not averaging down; they’re pressing a winner.
• Apple: A $516.9M add, with the position already +133.0% above their average cost, suggests they see more upside from ecosystem monetization and on-device AI rather than just hardware cycles.
• Alphabet (GOOGL/GOOG): Combined, the two lines saw over $660M in incremental capital. With both up roughly ~195% versus cost, Ubs is comfortable that AI infra and ad pricing power still have legs.
• Lam Research and Applied Materials: Shares up 17.8% and 15.6%, adding about $503.5M and $363.6M respectively. That’s a clear tilt toward the tools that enable high-bandwidth memory, advanced packaging, and leading-edge nodes — the durable layer of the AI stack.
• GE Vernova: A $357.9M step-up with shares up 19.2% reflects a parallel thesis around grid and energy infrastructure needed to power AI datacenters.

Outside pure tech, they quietly built Walm​art (+11.9% shares, +$270.7M) and Costco (+7.9%, +$183.0M). Those are not growth-chasing trades; they’re a bid for scale retailers with pricing power and traffic resilience if the cycle cools.

The funding side of the ledger is as telling as the buys. Ubs is not leaving semis; they’re rotating out of certain exposures that have already done the heavy lifting.

• Intel and AMD: Intel’s stake was cut by 28.5% (about -$1.08B) and AMD by 21.3% (-$965.3M). Both are deeply in the green (+214.4% and +243.5% vs. cost), which reads like profit-taking and a relative call that their upside is now less compelling than equipment, memory, or the hyperscalers.
• Taiwan Semi: Shares down 13.1% (roughly -$687.3M), despite a +128.9% gain vs. cost. That’s consistent with de-emphasizing foundry cyclicality in favor of more oligopolistic tooling providers.
• Palo Alto Networks: An 18.1% trim, freeing up about $510.6M, suggests some skepticism about how much incremental AI tailwind is left in high-multiple security names.
• Cisco and Texas Instruments: Double-digit percentage cuts (-13.2% and -13.1%), monetizing mature franchises that have rerated nicely while still keeping them in the book.

Importantly, they also used index ETFs as a cash machine. IVV was slashed by 22.0% (-$770.1M), QQQ by 7.0% (-$317.5M), and there were mild trims in SPY and IWF. Rather than de-risking the entire equity book, this looks like a pivot from anonymous market beta into hand-picked beneficiaries of the same macro themes.

On the surface, sector weights barely budged: technology inched from 47.7% to 47.8%, and the catch‑all “unclassified” bucket (ETFs, gold, Berkshire, GE Vernova) ticked from 28.2% to 28.3%. The story is in the mix beneath those steady headlines.

Tech exposure is subtly rebalanced: less in CPUs, foundries, and legacy networking, more in software platforms (Microsoft, Alphabet), diversified hardware (Apple), and semi equipment (Lam, Applied). That’s a shift toward higher-quality, more durable cash compounders within the same secular AI theme.

Consumer discretionary crept up from 6.75% to 6.84% on adds to Walmart, Costco, and a modest Tesla increase. That blend favors scale retailers and a controversial but still-core electric auto platform, suggesting Ubs wants cyclical upside but anchored in category killers rather than marginal players.

Health care slipped from 4.83% to 4.73%, mostly via trims in Eli Lilly and AbbVie despite sizable unrealized gains. They clearly still like the GLP‑1 and pharma cash‑flow stories, but are unwilling to let these become outsized single-name risks.

Elsewhere, they quietly upgraded the shock absorbers: real assets and income. Gold via GLD rose with a 7.8% share increase, and investment-grade credit via USIG saw a 4.8% add, even though that ETF sits slightly below their cost. Finance nudged down (7.07% to 6.97%) on trims in Morgan Stanley, partially offset by small adds in Goldman and their own UBS stock.

Taken together, this quarter’s moves sketch a pretty tight thesis. Ubs Group AG wants maximum participation in AI and digitization — but via dominant platforms, semi “picks and shovels,” and the infrastructure and power grid that underpins them, not just the headline chip stories that already re-rated.

The internal rotation within tech suggests they see the cycle moving from speculative enthusiasm to earnings delivery. By taking profits in Intel, AMD, Taiwan Semi, and Palo Alto Networks while adding to Microsoft, Apple, Alphabet, Lam, and Applied Materials, they’re effectively saying the value has migrated up the stack and into the supply chain.

Around that core, they are building a defensive shell. Adds to Walmart, Costco, Johnson & Johnson, and gold point toward a portfolio designed to survive a bumpier macro path, even as equity indices hover near highs. The extra allocation to US investment-grade credit, despite a small mark-to-market loss, reinforces this preference for carry with downside protection.

Expect future quarters to rhyme with this one unless the AI narrative truly cracks or policy shocks force a wholesale de-risking. As long as earnings from the mega-cap platforms and semi-cap complex keep confirming the story, Ubs is likely to keep redistributing capital within tech and infrastructure — not away from it — while using ETFs, credit, and gold as levers to fine‑tune overall risk.