BNP Paribas Financial Markets 13F Portfolio

The biggest dollar add is Micron, where they lifted exposure by 73.5% and pushed it to 4.88% of the book. Pair that with a 36.8% add to AMD and a 59.6% add to TSM, plus incremental NVDA, and the message is blunt: AI demand is nowhere near peaking, and they want the more cyclical, supply-sensitive legs of that trade.

They’re not just buying chips; they’re buying the system-level beneficiaries. Tesla is up 27.8%, Caterpillar nudged higher, and GE jumped 53.4%, tying AI and automation to a broader industrial upcycle. This reads as a bet that AI-driven productivity and capex support both data centers and real-world capital goods.

At the portfolio-structure level, they hammered macro beta:

• ACWI was ramped by 197.2% to $776.3M, signaling a call on global equities rather than just US tech.
• EWY jumped 45.4%, and SMH rose 48.1%, adding a Korea-and-semis regional tilt that lines up neatly with the memory/foundry thesis.
• IWM and SPY both saw mid-teens percentage adds, showing a willingness to ride a broader risk-on tape rather than hiding in a handful of AI darlings.

They also leaned into defensive growth: UnitedHealth was lifted 20.4% even though it sits slightly below cost, and Eli Lilly, AbbVie, and J&J were all added to. That mix says they want upside to a hot economy but are not comfortable running a one-way tech momentum book.

On the funding side, the most telling move is the 43.0% cut to Microsoft, shrinking it to 1.43% of the book. They also trimmed Apple by 7.4% and the GOOG share class by 8.3%, even as they added aggressively to GOOGL, effectively reshaping but not abandoning Alphabet exposure.

More structurally, they’re siphoning gains from the highest-quality, best-loved parts of the semi and equipment complex:

• ASML was cut by 32.7%, AMAT by 18.9%, AVGO by 18.8%, ADI by 38.6%, and TXN by 43.2%. These are all up strongly versus cost, making them ideal cash registers.
• SNDK, with an enormous gain vs cost, was reduced by 20.7%, again monetizing a legacy home run.

Outside tech, they bled risk from biotech beta via a 20.7% trim in XBI, even as they boosted broader health-care stalwarts. SOXX was nudged down 7.5% while SMH was boosted, suggesting a preference for a different factor composition or constituent mix rather than a retreat from semis. The pattern is consistent: take money off in crowded, fully rerated AI and biotech proxies, then rotate into more cyclical, capacity-sensitive names and broader market and EM risk.

Technology’s share of the book actually dipped, from 59.69% to 57.17%, despite the huge adds in Micron, AMD, TSM, Lam Research, KLAC, Intel, and WDC. That’s because the trims in Microsoft, Apple, ASML, AVGO, AMAT, TXN, ADI, SNDK, and Meta more than offset the adds at the margin. Within tech, exposure is clearly migrating from mega-cap software/platforms and ultra-premium equipment into memory, logic, storage, and manufacturing gear.

Consumer and industrial cyclicals are quietly moving up the ranks. Consumer Discretionary climbed from 7.09% to 7.96% as they added to Amazon, Walmart, Home Depot, Netflix, and Costco — a vote for resilient US consumption. Industrials rose from 4.54% to 5.11% on Tesla, GE, Caterpillar, and the 46.7% increase in the XLI ETF, tying the AI and infrastructure story to an on-the-ground capex and transport cycle.

Health Care edged up to 4.69%, with adds to UnitedHealth, Eli Lilly, AbbVie, and J&J, balancing growth with defensiveness. Finance moved from 3.54% to 3.88% on higher JPMorgan and Morgan Stanley stakes, echoing the risk-on macro message. Energy nearly doubled from 0.64% to 1.12% on a 91.5% add to Exxon, offering an inflation and commodity hedge.

Unclassified ETF exposure (broad equity, factor, EM debt, high yield) is steady around the mid-teens but more assertively pro-growth: ACWI, EWY, SMH, IWM, EMB, HYG, SPY, and XLI were all increased. The sum of the parts is a portfolio that still lives and dies with tech, but is now flanked by global, cyclical, and defensive sleeves that make the AI bet more macro-integrated and less single-theme fragile.

Taken together, the 2026-Q2 13F says Bnp Paribas Financial Markets believes we are in an AI-led, growthy but late-cycle environment. They are willing to own more cyclicality and global beta, yet they are consciously redistributing risk within tech away from the most fully priced winners.

The aggressive build in Micron, AMD, TSM, Lam Research, KLAC, Intel, WDC, and SMH suggests they see a multi-year AI capex wave that still has room to surprise to the upside. These are the businesses whose earnings swing hardest when demand and pricing inflect — exactly what you buy if you think the industry is moving into a sustained, capital-intensive build-out rather than a short hype cycle.

At the same time, they’ve layered in more consumer, industrial, financial, health-care, energy, and EM exposure. That combination reads as a view that AI and productivity will coexist with stronger nominal growth, higher capital spending, and some inflation risk, rather than a disinflationary tech-only boom.

Going forward, investors should expect this book to keep operating as a barbell: high-beta, AI-linked semis and cyclicals on one side, and a thick cushion of ETFs, health care, and credit on the other. The key tell will be whether they continue to reduce mega-cap software and premium equipment to feed memory, foundry, and global beta — if that pattern persists, their thesis is that the second derivative of AI, in industrials, EM, and commodity demand, is where the next leg of returns lives.