Mitsubishi UFJ Asset Management Co., Ltd. doubles down on AI and US consumers

The biggest buys list makes it obvious where conviction is rising: scalable platforms with embedded operating leverage. Amazon led the dollar adds, with the position up 20.7% in shares and now at 3.61% of the book — a clear vote that e‑commerce, logistics, and cloud still have a long runway from here.

Netflix is the most aggressive single‑name bet in the quarter: shares are up 99.9%, almost a doubling of exposure, turning it into a $1.80B position. That move says they see durable pricing power and streaming consolidation outweighing near‑term churn or content risk.

On the AI side, they raised NVIDIA, Broadcom, and Alphabet (GOOG line up 9.9%) even after very large gains: Broadcom’s gain vs average buy is 206.4%, NVIDIA’s is 198.9%, Alphabet’s GOOG line is 152.7%. This is not dip‑buying; it is momentum in companies they believe are structurally under‑earning their AI potential.

The most intriguing outlier is AON, where they lifted shares by 38.8% to about $812.3M despite the position sitting slightly below their average cost. That looks like a deliberate build in a resilient, fee‑rich insurance franchise that benefits from higher complexity and global risk, not from rate cuts or loan growth.

Across these adds, a pattern emerges: they are paying up for platforms — cloud, chips, content, and fee businesses — where incremental revenue falls disproportionately to the bottom line.

The sells side of the ledger is small but telling: when this manager needs cash, it comes from defensives and mature cyclicals, not from their AI or consumer engines. The largest trim in dollars is Thermo Fisher Scientific, where they cut shares by 11.2% even though the position is already modest at 0.61% and sits below their average cost.

Thermo Fisher’s negative gain vs average buy (‑10.9%) and the decision to reduce anyway suggests an explicit de‑prioritization of tools and diagnostics relative to higher‑beta growth. It looks more like a thesis downgrade than a simple risk cut.

Procter & Gamble was also trimmed, with shares down 3.5% and a slightly negative gain vs cost. Using a defensive consumer staple as a funding source to buy Amazon and Netflix is a straightforward statement that the marginal dollar is better in growth than in safety.

Even in semis, the only notable trim is Texas Instruments, down 4.2% in shares. Given TXN’s 79.2% gain vs average buy, this looks less like a loss of faith in chips and more like recycling capital from lower‑growth analog exposure into higher‑octane AI names such as NVIDIA, AMD, Micron, Lam Research, and Applied Materials.

On the surface, sector weights hardly moved: technology slipped from 53.59% to 52.68%, while finance and health care are essentially flat. Under the hood, though, the quality of that tech exposure shifted further toward AI leaders and away from lower‑growth names like Texas Instruments.

Consumer discretionary is where the real rotation shows up, climbing from 12.63% to 14.03%. That jump is driven by the outsized adds to Amazon and Netflix plus meaningful lifts in Walmart, Costco, Home Depot, and AutoZone — a broad expression of confidence in the US consumer and in asset‑light, brand‑heavy retailers.

Industrials, energy, and other cyclical sectors are being managed more surgically. Industrials ticked down from 5.83% to 5.58%, even as they increased positions in Tesla, Caterpillar, and RTX, offset by the Thermo Fisher cut; it’s a rotation within industrials toward aerospace, autos, and heavy equipment.

Financials sit almost unchanged at 8.33% vs 8.34%, but composition is creeping toward diversified fee and market‑sensitive names: JPMorgan, Goldman Sachs, ICE, CME, SPGI, and AON were all increased. Classic bond‑proxy buckets — consumer staples, telecom equipment, and utilities mis‑tagged as real estate — are marginal funding sources, not growth engines.

Taken together, this book telegraphs a clear belief: the next leg of returns will come from AI infrastructure, mega‑platforms, and the US consumer’s willingness to pay for convenience and content. The manager is willing to tolerate volatility — and chase winners — in exchange for exposure to those secular curves.

They have not de‑risked after a strong three‑year run (weighted 3‑year annualized performance at 27.61%); instead, they are compounding into strength by adding to names with triple‑digit gains versus cost, like NVIDIA, Broadcom, Alphabet, AMD, Lam Research, and Micron. The modest reduction in headline tech weight masks a deeper concentration in AI hardware, software, and design tools.

For macro exposure, the portfolio is built to participate in both risk‑on and risk‑off regimes. Quality financials, large integrated oils, big pharma (Eli Lilly, J&J, Merck, AbbVie), and Berkshire Hathaway all provide ballast, but they are being grown gradually, not aggressively.

The key risk to this posture is obvious: a broad de‑rating in AI or a shock to US consumption would hit both the top tech platforms and the ramped‑up consumer discretionary sleeve at the same time. Conversely, if AI capex and consumer spending prove stickier than the market fears, this configuration is set up to continue outperforming.

In other words, Mitsubishi UFJ Asset Management is signaling it would rather be early and overweight the structural winners than perfectly timed but under‑exposed — and it is using every trim in defensives to back that call.