The Bucket
~15 % of the Q4 2025 13F, almost entirely Intel calls. The bucket is the cleanest expression of one of Aschenbrenner’s most-emphasized claims: leading-edge logic must be onshored. Concentration of cutting-edge fabs at TSMC in Taiwan is a national-security and AGI-program risk, and the fund expresses the bet on US fab capacity through Intel Foundry (with Tower Semiconductor as a smaller specialty-foundry adjacency).
Thesis
The two names in the bucket are very different:
- Intel (INTC) — the only US-headquartered company with leading-edge logic ambitions. The Q4 2025 call position is structured for the upside scenario in which Intel Foundry wins external customers (rumored: Microsoft, Apple, NVIDIA) and 18A enters volume production with respectable yields. The cash-at-risk is a fraction of the $747 M notional — call options are convex, low-delta exposure to a re-rating event.
- Tower Semiconductor (TSEM) — Israeli specialty foundry that runs analog, RF, sensor, and mixed-signal nodes nobody else wants to operate. Already partly consumed by Intel for US capacity (the failed Intel acquisition in 2024 was abandoned, but the operational coupling remains). Plays the same onshoring theme one node behind the bleeding edge.
What the fund explicitly does not own: TSMC (briefly held puts in Q3 2025), Samsung Foundry, Global Foundries, ASML, AMAT, KLAC, Lam, AVGO, AMD, NVDA. The semiconductor exposure is concentrated almost entirely on the US-fab onshoring axis rather than the AI-accelerator design axis or the wafer-fab equipment axis.
Constituents
| Company | Ticker | Q4 2025 line(s) | Value | % of 13F |
|---|---|---|---|---|
| Intel | INTC | Calls + 1 sh com | $746.8 M | 13.54% |
| Tower Semiconductor | TSEM | Common | $84.9 M | 1.54% |
| Bucket total | $831.7 M | 15.08% |
Position History
INTC calls were established in Q1 2025 at smaller notional and grown each quarter through Q4 2025. The Q4 2025 step-up to $747 M notional was the biggest single-quarter add to any individual position in fund history. Tower was added in Q3 2025 at roughly current size.
Read-Through
The Intel calls position is a high-convexity bet on a political-industrial re-rating: the CHIPS Act, the rumored Trump-administration push to get hyperscalers and frontier labs to commit to US-fabbed silicon, and the closing of the 18A vs. TSMC N2 yield gap. None of these need to materialize for the fund to break even (calls cost premium, not notional), but if any does, the upside is multiples of premium. The lack of any AMD or NVDA long underlines that the fund is not betting on accelerator share — it is betting on where leading-edge logic gets fabricated.