IMC-Chicago LLC 13F Portfolio

The biggest adds read like a statement that AI is entering its platform phase. Capital is rotating from early infrastructure winners into diversified beneficiaries with stronger competitive moats and balance sheets.

• DIA: A +1,755.5% share increase and a 5.77% weight make the Dow ETF the single clearest tell. IMC is happy to let the mega‑cap industrial and financial complex carry a chunk of their risk, rather than continuing to scale single-stock cyclicals.
• EWY: The +1,758.2% add to Korea at 3.09% signals a deliberate bet that Korean hardware and memory names still have leverage to AI, but the stock‑picking edge has compressed; they want the country factor, not just a Micron clone.
• AMZN and AAPL: Amazon’s position explodes (+511.5% in shares) to 3.56% of the book, while Apple’s stake goes effectively from negligible to 2.24%. That’s a re‑rating of the hyperscale and device distribution layer of AI: you don’t have to guess which workload wins if you own the rails.
• TSM and ASML: Despite trimming other semis, they still increase TSM by +85.2% and ASML by +143.7%. The message is that leading‑edge capacity and lithography remain non‑negotiable chokepoints; they’re de‑risking the second tier (Micron, Western storage peers) while doubling down on the irreplaceable.
• MSFT, META, MRVL, KLAC: A new Microsoft stake at 1.04%, a near‑doubling in Meta, and surges in Marvell and KLA show a tilt toward AI’s networking and tooling layer. It’s a quiet but pointed upgrade from high‑beta HBM trades to the companies that set the roadmap.

The common thread across these adds is comfort with paying up for scale, breadth, and optionality rather than incremental operating leverage.

On the sell side, IMC looks like a trader ringing the bell on the first AI leg rather than an investor abandoning the theme. The biggest trims are all about taking gains where the curve has already gone parabolic or where thesis clarity has deteriorated.

• MU and AVGO: Micron is still a 5.23% anchor, but the fund cuts shares by -32.5% after a gain vs cost north of +400%. Broadcom is slashed -68.6%. They’re not exiting, they’re simply refusing to chase late‑stage multiple expansion in the obvious AI toll booths.
• SNDK, TSLA, ORCL: SanDisk is cut -76.3%, Tesla -45.5%, Oracle -36.1%. These feel like classic funding sources: mature winners with big embedded gains (SanDisk) or stories where volatility and execution risk have risen faster than edge (Tesla’s manufacturing and pricing saga, Oracle’s crowded AI narrative).
• CEF vs PSLV/PHYS/SPPP: They trim CEF by -48.6% while adding to PSLV and PHYS and leaving SPPP untouched. That’s not a metals call; it’s a structure call, rationalizing overlapping Sprott vehicles toward the preferred wrappers.

Viewed together, the sells show a manager that is allergic to hero trades overstaying their welcome. They cash in on the obvious AI winners, recycle into broader exposure, and selectively upgrade quality at roughly the same theme nodes.

Sector data confirms this is not a tech exodus so much as a reshaping. Technology still dominates at 48.08%, but that’s down from an estimated 60.97%, and the freed capacity is being redeployed into uncorrelated and cyclical sleeves.

Unclassified ETFs and structures jump from 16.29% to 26.44% as DIA, EWY, TQQQ, KRE, and JEPQ all scale up. That’s a conscious move from single‑name risk toward factor and income products, especially in financials (KRE) and covered‑call tech (JEPQ).

Consumer Discretionary climbs from 4.19% to 11.16% on the back of Amazon, Costco, Walmart, Nike, and gaming name Take‑Two. IMC is effectively saying that if AI sustains a growth cycle, consumer platforms and brands will be one of the cleaner downstream expressions.

Industrials fall sharply from 16.18% to 6.13%, despite a new Boeing stake, because of the Tesla cut and the shift of cyclicality into the index bucket via DIA. Meanwhile, Energy goes from 0.77% to 2.7% on adds to Bloom and new Chevron and EXE, and Real‑Estate‑labeled exposures (which here are really China tech via Alibaba, consulting via Accenture, and Uber’s mobility platform) rise from 0.9% to 2.68%.

Finance and Consumer Staples also appear for the first time, via JPMorgan and Pepsi at 1.45% and 0.66%. The pattern is classic late‑stage bull behavior from a disciplined trader: keep tech as the core, but surround it with hard assets, banks, and staples that can survive a rate and volatility shock.

Taken together, IMC-Chicago’s Q2 2026 book says they expect AI to remain the macro center of gravity, but they no longer trust narrow semis and story stocks as the only way to play it. They’ve upgraded into platform names, diversified geographically through Korea, and wrapped a sizable chunk of risk inside Dow and tech‑tilted ETF sleeves.

The build‑out of precious metals (PSLV, PHYS, SPPP), copper (FCX), and energy (CVX, BE, EXE) suggests they are also gaming for a world where AI’s demand for power and infrastructure collides with supply constraints. That hard‑asset ballast would look prescient if inflation proves sticky or if the market finally prices in grid and materials bottlenecks.

On the risk side, the book is now more sensitive to broad equity drawdowns and less to single‑stock blowups. If indexes crack, DIA, EWY, TQQQ and consumer platforms will all get hit in unison; the metals and energy sleeve plus banks and staples are there to soften that punch, not avoid it.

And the contrarian adds — from Nike despite drawdown, to heavily scaled China tech exposure via BABA and BIDU, to a fresh UBER position — show an appetite for asymmetric upside where sentiment is still fragile. If the AI trade broadens into a multi‑year capital cycle and global growth muddles through, this barbell is designed to harvest that upside while keeping enough hedges on the table to survive the next volatility spike.