American Century Companies Inc rotates from megacap AI trades to AI infrastructure

The biggest add was Microsoft, up 6.8% in shares and about $380.3M in value, a straightforward statement that they see it as the durable AI platform with the best payout profile. That sits alongside a sharp tilt into the equipment and tooling powering AI demand: Lam Research was boosted 29.0% (about $157.6M added) and ASML 33.3% (about $118.9M added), tightening the fund’s grip on the semiconductor manufacturing bottlenecks.

The software stack underneath modern workloads also saw fresh enthusiasm. Cloudflare was lifted 14.3% (roughly $93.4M), Datadog 4.3% (about $18.4M), and Palantir 21.5% (about $110.7M), collectively betting that security, observability, and data platforms take a growing share of AI-related IT budgets.

Outside pure tech, they made an unambiguous statement in financials via Marsh & McLennan, raising the position 81.7% and roughly $330.2M despite the name sitting just below cost. In health care they leaned into innovation and operating leverage rather than pure defensives, adding Alnylam by 21.4% (around $105.0M), Medtronic by 13.4% (about $56.3M), Becton Dickinson by 8.8% (roughly $43.4M), and modestly lifting Merck and Regeneron — a pivot toward earnings growth in a sector that had been more value-heavy.

The funding list starts with the obvious winners. Nvidia, still 5.48% of the book and more than 1,100% above cost, was cut by 3.6% in shares (about $395.8M), while Apple, up more than 850% vs cost, was trimmed 3.1% (roughly $208.0M). Alphabet’s GOOGL and GOOG lines were reduced 7.4% and 4.4% respectively, taking out about $526.0M combined, and Amazon was cut 8.8% (roughly $485.6M) despite outsized gains.

This is less a repudiation of big tech than a recognition that risk-reward is better in the second wave of AI beneficiaries. Oracle stands out as more than just profit-taking: they slashed it 24.2% (about $152.4M) while sitting below cost, signaling waning conviction in its ability to keep up with newer cloud and data architectures. In health care and staples, they clipped Johnson & Johnson by 17.7% (around $150.9M), AbbVie, Gilead, and Walmart, rotating away from slow but steady franchises toward higher-growth therapeutics and devices.

Cyclical and yield exposures were also rationalized rather than abandoned. Taiwan Semi was pared 5.7% (about $130.5M) even as they doubled down on tools like Lam and ASML, and they cut Verizon by 15.4% and Baker Hughes, Chevron, Williams, and Howmet meaningfully. The message: keep exposure to cash-generative cyclicals, but use them as a buffer, not as the heart of the book.

On the surface, sector weights barely moved: technology dipped only from 60.37% to 60.27%, and consumer discretionary from 14.02% to 13.77%. The real story is inside tech, where capital is clearly sliding from end-user megacaps into semiconductors, semicap equipment, and infrastructure software.

Health care quietly rose from 9.47% to 9.54%, but the shift is meaningful: less in high-yield pharma giants like JNJ and ABBV, more in innovation-heavy names like ALNY, BDX, MDT, and REGN. Finance jumped from 1.92% to 2.38%, effectively a targeted bet on fee-based, capital-light risk management via Marsh & McLennan and a small incremental add to JPMorgan.

The portfolio also added ballast. Utilities edged up from 1.16% to 1.17% as they increased Duke Energy, and unclassified exposure — mainly the internal AVUS ETF and GE Vernova — crept from 1.16% to 1.26%, giving them diversified and energy-transition-adjacent sleeves. Industrials, energy, and telecom all ticked down modestly, suggesting those sectors are now viewed more as funding pools than active bet sizes.

American Century is effectively saying that the easy money in front-page AI trades has been made and that the next leg of returns lies in the picks-and-shovels and data plumbing behind it. The reallocation from Nvidia, Alphabet, and Amazon into Microsoft, Lam, ASML, Cloudflare, Datadog, and Palantir builds a book that is more levered to ongoing capex and software spend than to ad or consumer sentiment.

At the same time, a selective upgrade of health care risk and a beefed-up Marsh & McLennan stake indicate they’re comfortable carrying more factor volatility — growth, quality, and a bit of duration — so long as it is wrapped in resilient business models. Slight increases in utilities and their own AVUS ETF show an awareness that a -8.56% quarter can repeat and that some built-in shock absorbers are warranted.

Going forward, expect them to keep tech overweight but tilt increasingly toward capacity, bandwidth, and data layers rather than just AI front ends. If volatility in megacap tech persists, this book is already halfway repositioned: less reliant on one or two stocks and more wired into the broader capital-spending and innovation cycle that those giants are funding.