Nordea Investment Management AB 13F Portfolio

The biggest adds tell a clear story: Nordea believes the next leg of AI returns will accrue to laggier hardware, power infrastructure, and factory automation rather than the headline chip winners.

• Intel: A 327.6% position size jump to 0.53% of the book, adding about $501.2M, is a statement that Intel’s turnaround and foundry ambitions are finally investable. The position is only 25.9% above average cost, so Nordea is leaning into a still-early rerating rather than chasing a parabolic chart.
• GE Vernova: A 29.4% add (about $210.8M) to $928.2M signals a strong view that grid, turbine, and transmission upgrades are a structural winner of AI’s power hunger. At a 62.5% gain vs cost, they’re pressing a winner, not bottom-fishing.
• Exelon: Boosting the utility stake by 28.6% (roughly $137.7M) with only a 3.3% gain vs cost looks like a valuation-driven re-rate call on regulated power as datacenter demand ramps.
• Rockwell Automation: A 26.8% increase (about $128.5M) shows conviction that industrial automation will monetize AI at the factory floor, not just in the cloud.
• AbbVie, Johnson & Johnson, and Meta: Meaningful dollar adds in these three show Nordea layering in high-cash-flow defensives and profitable digital platforms as the growth core, not as afterthoughts.

Taken together, the “buy” list is infrastructure-heavy: fabs, grids, and robots that benefit from AI, plus large-cap drugmakers that can weather any derating in the growth complex.

On the sell side, Nordea is not abandoning AI; it is cashing in where multiples and gains are most extended to fund the infrastructure build-out and more defensive ballast.

• AMD and Lam Research: Cuts of 25.6% and 19.8%, respectively, freeing an estimated $613.5M and $528.2M, look like classic profit-taking. With AMD up 213.0% vs cost and Lam up 448.8%, the risk-reward has simply shifted to names like Intel and power utilities.
• Marvell and Broadcom: Trims of 23.2% in Marvell (about $235.1M) and 5.2% in Broadcom (about $196.3M) suggest Nordea sees better incremental upside in earlier-stage beneficiaries than in the semis already priced as AI royalty.
• Cisco and Arista: Selling 12.2% of Cisco (about $239.2M) and 15.3% of Arista (about $104.5M) shows some skepticism that datacenter networking will keep pace with the chip cycle or that current expectations leave much upside.
• Linde and Trane: Large reductions in Linde (down 42.5%, about $415.6M) and Trane (down 23.2%, about $300.5M) indicate that “green industrial” winners have moved from under-owned to fully valued in Nordea’s eyes.

Even mid-cap defensives are on the chopping block where gains are fat and growth modest: Wabtec, Emerson, Waste Management, and others were clipped to help pay for the grid, automation, and health care adds.

Despite all the trading around the edges, sector weights barely budged in headline terms — but the mix inside them did, and that’s where Nordea’s thesis shows up.

Technology remains the engine at 59.61% vs 59.71% last quarter, yet within that slice Nordea rotated from high-multiple semis (AMD, Marvell, Lam, Broadcom) toward more cyclical or under-appreciated plays (Intel, Apple, Meta). This is AI exposure with less crowding and, in their view, more asymmetry.

Health care crept up to 9.60% from 9.19%, driven by higher stakes in AbbVie, Johnson & Johnson, Eli Lilly, and McKesson. That looks like a conscious choice to balance volatile AI enthusiasm with durable patent and distribution cash flows.

Financials moved to 6.70% from 6.41% as Nordea modestly increased JPMorgan, Citigroup, American Express, Hartford, and Wells Fargo. They’re effectively rebuilding classic rate and credit exposure alongside the growth book.

Outside these, industrials, real estate, and telecom-related names ticked down slightly as individual trims outweighed adds. Utilities nudged up to 1.74%, but the character of that exposure — more Exelon, less bond-proxy — underscores that Nordea is treating power as AI infrastructure, not as a sleepy dividend sleeve.

Read across the book, Nordea is betting that the easy money in front-line AI winners has been made and that the second derivative of the theme now lives in infrastructure, power, and process automation.

By sharply increasing Intel and adding to GE Vernova, Exelon, and Rockwell, they are positioning for a capital-expenditure supercycle: new fabs, rewired grids, and automated plants needed to make AI real in the physical world. Those moves are financed by trimming high-flyer semis, networking equipment, and select industrial champions whose multiples already reflect years of good news.

Parallel adds in AbbVie, Johnson & Johnson, Eli Lilly, McKesson, and major banks show an insistence on keeping ballast as the AI story matures. Nordea isn’t walking away from mega-cap tech — NVIDIA, Apple, Microsoft, Alphabet, and Meta remain core — but they are clearly upgrading the quality and breadth of the ecosystem around them.

For observers, the message is straightforward: if the AI cycle continues, Nordea expects the marginal dollar to flow to power, plumbing, and automation rather than to the same handful of poster children. If it stumbles, they want enough health care and financials in the book that the drawdown is survivable.