Arrowstreet Capital, Limited Partnership doubles down on megacap AI and energy

Arrowstreet’s biggest dollar adds cluster tightly around the dominant AI platforms. The fund boosted Microsoft by 64.3% (a $3.51B add), Broadcom by 118.1% ($3.14B), and Nvidia by 19.2%, effectively taking the view that the AI stack’s profit pool will consolidate in a handful of hyperscalers and chip suppliers, and that recent volatility is an entry point rather than a warning.

Alphabet is treated as a co-core: the GOOGL line was increased by 90.9% and GOOG by 76.1%, together adding over $4.14B of exposure to the same franchise. Arrowstreet also grew Amazon by 32.1% (a $1.65B add), reinforcing a belief that cloud plus commerce still has a long runway, with current gains vs average cost (for example +26.8% on Amazon) offering a margin of safety rather than a reason to harvest.

Under the hood, the fund is also buying into cyclical AI beneficiaries and controversial software rather than taking profits. Salesforce was lifted by 130.2% and Palantir by 277.4% despite both sitting below Arrowstreet’s average buy levels. Intuit, Adobe, and Fortinet all saw double‑digit or near‑double‑digit share growth. In Energy, a new $1.41B position in Canadian Natural Resources, a fresh $766.3M in Exxon, and a 121.4% increase in Suncor define a clear bet on size, integrated value chains, and oil sands leverage to higher-for-longer prices.

Funding these outsized adds required real cuts, and Arrowstreet went first to legacy winners and hedges. SanDisk, up an eye‑popping 1451.6% vs average cost, was trimmed by 18.8%, pulling out roughly $415.5M while leaving a still‑sizable 0.97% position. Micron, up 298.7% vs cost, was cut by 12.0%; this looks like disciplined profit‑taking in the most cyclical part of the AI memory trade while leaning harder into Broadcom and Nvidia.

The portfolio’s long-standing insurance complex in precious metals was a clear source of cash. Agnico Eagle was reduced by 21.8% (about -$382.6M) and Wheaton Precious Metals by 9.9%, while Barrick was nudged down. Given their triple‑digit percentage gains versus average buy prices, these look like conscious de‑gearing trades: less gold optionality, more exposure to cash‑yielding hydrocarbons and software.

Arrowstreet also eased back on older‑guard infrastructure and intermediaries. Cisco was cut by 22.1% (roughly -$387.9M) and Arista by 10.0%, even as Fortinet and other security names were increased, signaling a rotation from network hardware to software‑centric security. CME Group, Nutrien, Philip Morris, and O’Reilly all saw mid-teens or low‑teens cuts, suggesting that stable, cash‑generative but slower‑growth holdings are being tapped to finance higher‑conviction growth and energy trades.

The sector bar chart shows a clean re‑rating of the book toward growth and oil. Technology’s jump to 57.87% is not a broad tech beta move; the adds are concentrated in a narrow set of platforms (Microsoft, Alphabet, Amazon, Nvidia, Broadcom) and selected software names, while some hardware and memory (Cisco, Micron, Arista) are trimmed. Arrowstreet is trading within tech, away from commoditized infrastructure and toward software and core silicon rent‑takers.

On the other side of the barbell, Energy’s move from 1.94% to 4.99% is striking. The basket — Suncor, Canadian Natural Resources, Exxon, and Valero — is skewed to integrated models and oil sands producers that can throw off significant free cash flow in a resilient price environment. This is funded in part by shrinking Basic Materials from 8.16% to 6.04%, with cuts in gold miners and metal names reallocating from stores of value to producers of cash.

Defensives and rate‑sensitives were gently pared back. Health Care slid from 8.61% to 7.9% despite new AstraZeneca exposure, as UnitedHealth, Cigna, HCA and McKesson were trimmed. Finance dropped from 2.24% to 1.62% on CME selling, and Telecom from 4.11% to 2.6% on networking cuts. Arrowstreet is accepting more growth and commodity risk, and a bit less defensive ballast, in exchange for concentrated bets where it believes the next leg of returns will emerge.

Taken together, this 13F says Arrowstreet sees the current drawdown as an opportunity, not a regime break. They are adding hard into megacap AI and platform software while most allocators fret about crowding, and they are pairing that with a decisive move into large, cash‑rich energy producers. The through‑line is simple: own the scarce assets with durable pricing power and real cash generation, and fund them by lightening up on hedges and middlemen.

The willingness to average down in Salesforce, Intuit, Adobe, Palantir, Intuitive Surgical, and Boston Scientific — all showing double‑digit losses versus Arrowstreet’s average buy — underscores a time‑horizon edge. They are not momentum‑chasing; they are underwriting multi‑year compounders and accepting short‑term pain to increase exposure at more attractive entry points.

On the risk side, the book is now even more levered to a continued AI capex boom and a firm commodity tape. A reversal in either would hurt: more than half the portfolio rides on tech multiples and data‑center spending, while a third leg now depends on Energy and cyclicals. But for investors trying to read Arrowstreet’s view of the world, the message is unambiguous: the next phase of returns, in their eyes, will be driven by software oligopolies, AI infrastructure, and oil‑linked cash flow — not by gold, rate defensives, or low‑beta financial plumbing.