BNP Paribas Asset Management Holding S A 13F Portfolio

The biggest incremental dollars went straight into the core AI software stack. Microsoft (up 8.4% in shares, +$294.1M) and Alphabet’s A class (up 8.3%, +$236.9M) were the top two adds, with Nvidia also getting another $109.7M despite already being a 5.90% position and sitting roughly 97.4% above the fund’s average cost.

Beneath the headline platforms, they are leaning into the semiconductor “picks and shovels” that enable AI fabs. KLA (shares up 37.7%, +$222.3M) and Applied Materials (up 12.7%, +$147.4M) are clear winners; both trade well above the fund’s average buy levels, yet they are buying more, which reads as a strong structural view on wafer inspection and equipment demand.

Conviction is not limited to tech. Deutsche Bank is up 29.2% in shares (+$131.6M), Bank of America and American Express are also being built, and Visa/Mastercard both see healthy adds. Meanwhile Walmart’s stake jumps 33.0% in shares (+$111.6M) and Costco nudges higher, suggesting a deliberate build‑out of scale consumer and payments rails that benefit from nominal growth and AI‑driven efficiency without needing a perfect macro.

The funding side of the ledger is dominated by realized winners in semis. AMD is the standout: the position is cut by 31.3% (‑$762.0M) even though it sits roughly 337.2% above the fund’s average buy price. Micron (‑19.7%, ‑$296.2M), Intel (‑28.4%, ‑$296.7M), Marvell (‑25.8%, ‑$209.9M), and Lam Research (‑27.7%, ‑$412.8M) round out a clear theme of banking AI‑driven cyclical gains.

They are also dialing back some high‑multiple growth and momentum names that had become crowded trades. Palo Alto Networks is down 10.8% in shares (‑$120.6M), while ServiceNow, Ciena, FROG, and Cisco all see reductions, even though most are well in the green versus the fund’s cost. The message is that cybersecurity and networking remain core, but position sizes are being shaped to reflect more mature growth and full valuations.

Outside of tech, there is quiet profit‑taking in health care leaders and select defensives. Eli Lilly is trimmed by 13.5% (‑$187.0M), with smaller cuts to AstraZeneca, Gilead, Intuitive Surgical, and Linde. In consumer, TJX is reduced by 21.4% (‑$138.3M), suggesting a preference for scale big‑box and e‑commerce (Walmart, Amazon, Costco) over off‑price apparel this late in the cycle.

On the surface, sector weights look almost unchanged: technology inches down from an estimated 65.01% to 64.2%, health care edges from 10.28% to 9.93%, and industrials are flat near 4.21%. The real action is underneath — a migration within tech from cyclical chip makers toward equipment, bandwidth, and storage, and a modest build in financials and consumer staples‑like retail.

Finance rises from an estimated 3.2% to 3.6% as Deutsche Bank, Bank of America, and American Express grow, giving the portfolio more exposure to rate‑sensitive earnings and credit spreads. Real “financial infrastructure” in payments (Visa and Mastercard) also creeps up, mis‑tagged as real estate but economically very much part of the transaction rails theme.

Consumer discretionary edges up from 8.43% to 8.69%, driven by Walmart and Costco, even as TJX is cut. Telecommunications hardware and networking — Arista and Cisco, plus Ciena (filed as Utilities) — are effectively a growing AI plumbing sleeve, and the unclassified bucket rises as they scale GE Vernova and Everpure (PSTG), both tied to the physical and data‑center infrastructure that will have to catch up to AI demand.

Put together, this is the behavior of a manager that still believes the AI‑led cycle has legs but is no longer willing to ride the noisiest parts of the trade. They are migrating exposure from commodity‑like chip names toward software, tools, inspection, and the bandwidth/storage providers that stand to benefit from sustained, not just speculative, AI usage.

Building stakes in large banks, payments networks, and mass retailers says they want a portfolio that can live with higher nominal GDP, higher rates, and policy uncertainty. These are businesses that can reprice, capture operating leverage from AI internally, and still throw off cash if AI equity multiples compress.

For observers, the takeaway is that future 13F snapshots are likely to show incremental adds to dominant platforms and infrastructure enablers rather than fresh forays into second‑tier AI hardware. Expect further pruning of stretched winners in health care and cyclical tech as they continue to fund a barbell: high‑quality AI and cloud on one side, durable, cash‑generative franchises in finance and retail on the other.