Optiver Holding B V 13F Portfolio

The biggest dollar adds show Optiver taking AI from a clever theme to a capital-intensive core. Nvidia, Micron, and TSMC now sit at the heart of the book, with Nvidia up 3403.6% in shares and Micron and TSMC both materially increased – they’re explicitly backing the full AI compute stack, not just the poster child.

At the same time, they are rebuilding the megacap platform layer that monetizes that compute. Apple, Microsoft, and the twin Alphabet lines all saw large increases; Alphabet via a fresh GOOGL position and a 591.6% add to GOOG, Microsoft up 640.3%, and Apple expanded by 62317.0% from a token stub to a real line. Amazon’s +1861.8% move completes that bet on cloud, ads, and consumer rails.

The new software names tell you they expect AI to manifest as networked, usage-based SaaS, not just chips. Datadog, Cloudflare, and Zscaler all saw aggressive initiation or scaling, with Zscaler’s shares up 2416.8%. Netflix and Costco debut as high-quality consumer demand sinks, but sized more like satellite convictions; the real size is squarely in AI infrastructure and the platforms that control attention and spend.

Finally, the huge build in SPY, DIA, and IVV (with a new QQQM line) suggests they prefer to own a lot of the cycle via indices and then express sharper views by overweighting AI and software winners on top, rather than trying to be right stock-by-stock across the entire S&P.

The sells are few but loud: they are not de-risking; they are reallocating. The biggest trim is XLF, slashed by 72.8%, a clear rejection of generic U.S. financial beta in favor of more surgical bets on specific banks, brokers, and asset managers.

Broadcom is the only notable AI infrastructure name moving the other way, with a 40.1% reduction despite being up 98.2% versus cost. That looks like harvesting a big winner to fund higher-conviction semis (Micron, Nvidia, TSMC, Analog Devices) where they see more upside per unit of risk.

American Express is down 34.8%, a classic funding source: modest gain, mature brand, and far less operational torque than the high-growth software and AI stack they’re building. The trims cluster around "blunt instruments" – sector ETFs and slower-growth incumbents – to free capital for sharper expressions in the same macro themes.

On paper, technology’s weight drops from an estimated 83.31% to 58.91%, but that’s misleading. Much of what used to sit in single-name tech risk has been reconstituted as index exposure plus a more focused, higher-quality tech core.

The real move is into unclassified ETFs (27.6% vs 11.37%), which here are broad U.S. indices (SPY, IVV, DIA, QQQM) and country funds (FXI, EWZ, EWY) – this is index beta masquerading as a sector. Finance steps up to 5.79% with concentrated plays in Coinbase, banks like Lloyds, Barclays, BBVA, and capital-markets infrastructure via CME, Blackstone, and BlackRock.

Consumer discretionary climbs to 3.88% via Amazon, Netflix, Costco, and Walmart, essentially a bet that both premium and mass-market U.S. consumers stay resilient. Energy edges up to 1.42% as Shell, Chevron, and ConocoPhillips become a modest but clear hedge against higher-for-longer inflation and commodity shocks.

Seen together, sector rotation is less about abandoning tech and more about building shock absorbers around it: indices to smooth idiosyncratic risk, financials and energy to hedge the macro, and select consumer names to capture demand-side strength.

Optiver is positioning for an environment where AI and megacap platforms keep compounding, but factor and macro noise get louder. The index-heavy spine tells you they want to stay fully engaged with U.S. risk assets, while the semis-and-software wedge says they still believe alpha lives in the compute and data stack.

The migration from XLF into named banks and market-structure plays shows a preference for balance sheets and fee streams they can underwrite, not just a sector ETF that trades with rates headlines. Adding high-yield via USHY and emerging and Asian exposure through EWZ, FXI, and EWY says they’re willing to take measured spread and EM risk, but only in small, liquid slices.

Energy majors and defensive-but-growing retailers like Costco and Walmart are the quiet tell: they are not buying a Goldilocks soft landing; they are buying a world where inflation, geopolitics, and real-economy volatility can coexist with AI-driven profit pools. Expect them to keep rotating inside this framework – tweaking which chips, which clouds, and which banks they own – rather than abandoning the core thesis of AI-intensive growth carried on top of broad, tradeable beta.