Inside Jane Street Group, Llc’s Q1 2026 Playbook: From AI Darlings To Factor Beta

The biggest buys table makes clear they’re pressing into three themes: broad beta, second-derivative AI infrastructure, and targeted cyclicals that benefit from both.

On the beta side, they take SPY up over +103.5%, QQQ another +17.9%, and IWM +55.3%, while turning tiny foreign ETF stubs into real positions: SPDW explodes by +12238.3%, VT +498.8%, IEMG +45.4%, IEFA +54384.1%, and EWY +50567.2%. That’s a deliberate choice to express macro views and factor tilts (large-cap US, growth, small-cap, and ex-US) via liquid wrappers instead of stock picking.

Underneath, they’re doubling down on the AI stack, but away from the most crowded nodes:

• AVGO, MRVL, MU, and AMAT all get meaningful adds, signaling a preference for networking, memory, and tools over just the most obvious GPU flagships.
• ASML is ramped +319.8%, underlining conviction in lithography as a non-negotiable choke point in capacity.
• On the systems and manufacturing side, DELL is boosted +654.4% and CLS +131.3%, classic AI data center and electronics beneficiaries.

In software, they lean into platforms that can monetize AI workflows: PLTR is up +79.7%, NOW +82.7%, SHOP +1027.1%, CRWD +2016.7%, NET +209.9%, and SNOW +266.7%, even where they’re sitting on unrealized drawdowns (PLTR, NOW, SHOP, CRWD). That is not P&L optimization; it’s a thesis bet that these names are strategic rails rather than tactical trades.

Outside tech, there’s a quieter but clear build in cyclicals tied to industrial and defense demand: JCI climbs +274.8%, KEYS +310.8%, GLW +78.9%, HWM +2875.9%, and GM +37.0%. Together with more C (+318.1%) and SOFI (+86.7%), this reads as a view that rate-sensitive and industrially levered names still have room if growth and capex persist.

Funding this rotation required real sales, and Jane Street is unambiguous about where they’re taking chips off the table.

The headline trim is Taiwan Semiconductor: TSM’s stake is cut by -77.4%, freeing an estimated -$837.8M from a position that’s deeply in the money at +175.7% versus average cost. That is classic risk management: rotate from the most obvious AI foundry beneficiary into a broader basket of semis, tools, and systems where the risk/reward may look cleaner.

Alphabet is another major source of cash. Both GOOGL and GOOG are reduced, down -30.8% and -44.4% respectively, even though they still sit on very large gains of +76.0% and +116.1%. Rather than abandoning the theme, they’re right-sizing a mega-cap that has become a default AI proxy and using the proceeds to seed higher-beta, more operationally levered software names.

They also unwind a large portion of their precious metals and defensive growth exposure. SLV is cut by -57.3%, releasing around -$806.5M of silver exposure and pivoting that capital into more idiosyncratic Pan American Silver (PAAS), which is simultaneously ramped +3488.2%. UNH is halved (-52.5%), and NFLX is trimmed -53.3%, shrinking a legacy streaming winner that now carries a negative mark-to-cost. The pattern is clear: harvest liquidity from crowded, mature winners and broad hedges to finance more targeted, thematic risk elsewhere.

The sector bars show technology still dominates, but its share of the pie actually falls from 35.45% to 30.97%, even as they add to a long list of AI and software names. The reason is the simultaneous build in unclassified ETF exposure: wrappers jump from 40.12% to 44.94% as SPY, QQQ, IWM, SPDW, VT, IEMG, IEFA, and EWY scale up.

Health care is the main loser in the rotation, collapsing from 4.88% to 1.78% as UNH is sharply reduced and BSX only modestly increased. Consumer discretionary also steps down from 6.97% to 4.51%, with AMZN still growing but NFLX and exposure around media (WBD up, but not enough to offset) netting out to a smaller slice.

Meanwhile, they quietly thicken up real-asset and rate-sensitive buckets. Energy inches from 2.99% to 3.29% via big adds in FANG and EQT, while Utilities move from 1.99% to 2.29% with VST and NRG both significantly higher. Basic materials jumps from 0.04% to 0.88% on PAAS, and Finance rises from 1.08% to 1.87% courtesy of C and SOFI. Industrials lift from 2.63% to 4.54% on JCI, GLW, KEYS, HWM, and GM.

The result: a portfolio that still leans hard into tech and AI, but is now flanked by global beta, industrials, energy, utilities, and precious metals—effectively hedging a world in which AI drives power demand, capex, and commodity tightness as much as it drives software multiples.

Taken together, the moves sketch a fund that expects AI to keep compounding, but wants to own it more through the ecosystem than through the single most crowded names. They’re trading TSM and a chunk of Alphabet strength for a web of semis, tools, data center OEMs, and software platforms that could capture a broader range of AI-driven spend.

The surge in SPY, QQQ, IWM, and international ETFs says they’re comfortable expressing a lot of their macro view through index beta. That can serve both as a directional bet and as inventory for relative-value and options activity, but in 13F space it reads as intent: if you think the soft-landing-plus-AI narrative still has legs, this is how you scale it quickly.

The incremental pushes into energy (FANG, EQT, BE), utilities (VST, NRG), and PAAS point to a second-order thesis: AI datacenters and re-shored manufacturing will have real-world power and commodity footprints. Industrials like JCI, KEYS, GLW, HWM, and GM round out that view as beneficiaries of higher capex and defense/aero demand.

Going forward, expect Jane Street to continue this barbell: AI infrastructure and software at the growth end, balanced by real assets, utilities, and global beta, funded by trimming oversized winners when they run. The 2026-Q1 book says they’re not backing away from risk after a -4.94% quarter; they’re upgrading how, and where, they take it.