Victory Capital Management INC

The biggest adds table is unambiguous: they’re betting that the AI build‑out and global demand cycle are still in the middle innings, not the ninth.

• KLA: A +917.6% position explosion (about +$1.05B) into a name that sits an ugly -71.4% versus their own average cost is an unusual move for a large, diversified manager. They’re willing to be early and wrong in the short term to secure exposure to process control at the heart of advanced nodes.
• Booking: A +2566.2% surge (about +$724.1M) says they see durable, high‑end travel and pricing power, not a one‑off post‑COVID normalization. That’s an explicit call that consumer discretionary spend at the top end can outrun macro worries.
• Tesla and UPS: A +56.2% add to Tesla (about +$484.0M) and +29.1% to UPS (about +$345.1M) extend the same logic: the AI and e‑commerce era will require more physical throughput — vehicles, logistics, last‑mile — not just cloud compute.
• State Street and Bank of America: A +128.5% add to State Street (about +$429.4M) and +23.8% to Bank of America (about +$344.2M) shows rising comfort with rate‑sensitive financials. They’re leaning into balance‑sheet leverage to a sticky‑higher‑rates world.
• NRG and utilities: NRG’s +32.9% increase (about +$344.6M) and strong adds to Constellation and Eversource say the fund wants regulated, cash‑flowing exposure to grid and power demand — exactly what AI data centers and electrification will stress.

Overlay that with continued, if more measured, adds to Nvidia (+$413.8M), Alphabet (both share classes), Meta, QCOM, and Micron: they’re not abandoning megacap AI; they’re extending the bet down the value chain and into the real‑asset infrastructure that supports it.

The trims list is effectively a map of where Victory thinks risk/reward has peaked or where capital is simply trapped in lower‑beta, lower‑optionality stories.

• Truist: A -41.7% cut (about -$559.7M) to Truist is the clearest rejection in large banks. While they lift State Street, JPMorgan, and BofA, Truist looks like the funding source — a regional‑tilted exposure with less fee leverage and more idiosyncratic risk.
• AMD and Amazon: Selling AMD by -15.1% (about -$446.3M) and Amazon by -7.2% (about -$289.4M) after very large gains is not an AI retreat; it’s profit‑taking in the most crowded longs. They’re reallocating from the obvious AI and e‑commerce winners into names where future earnings revisions might be steeper.
• Cisco and Visa: Cisco (-16.1%, about -$287.5M) and Visa (-13.3%, about -$223.7M) are classic “quality compounders” being bled to fund higher beta ideas. These are not thesis blow‑ups so much as acknowledgment that their structural stories are well understood and fully valued.
• IBM and Johnson & Johnson: A -26.7% reduction in IBM (about -$226.2M), which still sits below their own cost, and a -23.6% cut in J&J (about -$205.2M) show a willingness to exit slow‑growth defensives, even at unexciting marks, to make room for fresher growth vectors.
• SanDisk: Trimming SanDisk by -22.9% (about -$209.3M) after a huge gain vs cost suggests they see better semicondutor risk/return in Micron, KLA, Applied, and Broadcom than in this particular memory exposure.

Net, they’re cashing in liquidity where the upside is more linear and recycling it into areas with steeper operating leverage and less consensus positioning.

On the surface, sector weights barely budged: technology slid from 55.96% to 54.45%, consumer discretionary from 10.72% to 10.55%, health care from 7.80% to 7.72%, and finance from 6.62% to 6.54%. But the mix inside those buckets is where the thesis lives.

Industrials jumped from 4.37% to 5.32% as they ramped Tesla, UPS, and Martin Marietta. That’s a clean statement that the physical side of the digital economy — vehicles, delivery networks, aggregates for data centers and warehouses — is finally getting paid.

Utilities climbed from 3.14% to 3.83% on big adds to NRG, Constellation, and Eversource, while Energy rose from 1.25% to 1.86% via Devon and Exxon. Add Basic Materials moving from 1.03% to 1.40% on a +46.0% FCX add, and you get an explicit power‑and‑resources sleeve behind their AI exposure.

By contrast, software and legacy tech are being pruned even as semis and semicap are favored. Within tech, they’re shifting weight from mature hardware, old‑line IT, and some software (NOW, IBM, Cisco trims) into chipmakers, equipment, and design tools (KLA, Micron, QCOM, Synopsys, Nvidia, Broadcom), i.e., into the capital stock required to sustain AI workloads.

Real estate‑labeled names (Visa, Uber, Accenture in the dataset) collectively edged down from 4.54% to 4.18%, signaling less enthusiasm for asset‑light, fee‑based models at current prices relative to old‑economy cash flow and cyclical operating leverage.

Pulling the threads together, Victory is structuring its book around a simple but powerful view: AI isn’t a stock story; it’s a capex and energy story. The winners, in their eyes, will be the companies that sell the picks and shovels — fabs, tools, power, copper, logistics — not just the platforms that rent out GPU time.

The increased commitment to KLA, Micron, Nvidia, Alphabet, Meta, and Broadcom says they expect the AI compute cycle to drive another leg of earnings upgrades in semis and infrastructure software. The parallel build‑out in utilities, oil & gas, copper, trucks, and aggregates suggests they see that capex spilling into the physical economy over several years.

At the same time, the trims to obvious winners and low‑beta stalwarts show a manager actively re‑risking, not hiding in benchmarks after a strong run (their latest‑quarter performance sits at 17.95%). They’re comfortable trading some near‑term mark‑to‑market volatility for longer‑dated operating leverage.

For observers, the signal is clear: expect Victory’s future quarters to keep pressing this theme — marginally less megacap consumer internet and defensive pharma, slightly lower headline tech weight, and more capital parked in the infrastructure spine that powers data centers, electrification, and global mobility.