The infrastructure enabling AI compute at scale.

Concentrated long-only positions in the public-market beneficiaries of forced AI infrastructure spend — from compute silicon and interconnect to memory, networking, and power management.

The Technology Thesis

Why Technology?

The current technology cycle is not driven by consumer preference. It is driven by the physical requirements of AI compute, data transmission, and semiconductor supply. Capital is being forced into infrastructure at a scale that creates non-discretionary demand for a small group of companies — the ones that own, control, or disproportionately benefit from the bottlenecks.

GZC’s Technology pool targets these companies exclusively. We enter only when conviction clears our full research and technical framework: the Bottleneck-to-Ticker process, applied to every position. We hold for the duration of the thesis — not the duration of a trade.

The pool currently covers compute silicon, data center interconnect, memory and storage, networking, semi cap equipment, and power management — all through the BTT lens of forced spend, limited substitutes, and multi-year procurement visibility.

Bottleneck Categories

Six structural themes. One framework.

Compute Silicon

GPU and custom accelerator manufacturers at the center of AI training and inference demand.

NVDAAMDGOOGL

Data Center Interconnect

PCIe, optical retimer, and high-speed connectivity silicon — the chokepoint of AI cluster performance.

ALABMRVL

Memory & Storage

HBM and DRAM suppliers serving AI training and inference workloads with non-discretionary demand.

MUPSTG

Networking

High-speed switching and routing infrastructure for AI data centers and cloud scale.

ANETCRWV

Semi Cap Equipment

Lithography, deposition, and etch equipment — the upstream manufacturing bottleneck for every chip.

AMZNMSFT

Cybersecurity

Identity, endpoint, and network security infrastructure as AI attack surfaces expand.

VRTETN
Why This Pool

Three reasons the Technology pool matters.

01

AI infrastructure buildout represents a multi-year, non-discretionary capital expenditure cycle — hyperscalers have committed to $150B+ in capex annually. Supply chains cannot respond fast enough, creating structural scarcity.

02

The Technology pool targets companies that own critical supply chain bottlenecks. Demand is forced — data centers cannot operate without GPUs, HBM, interconnects, or power management. Substitution risk is minimal.

03

BTT analysis identifies when pricing power, backlog visibility, and procurement dynamics align with technical entry discipline. We size positions by conviction — not by diversification quota.

Pool Characteristics

How the Technology pool is managed.

Position Count Target
8–12 names
Typical Position Size
5%–12% of pool
Market Exposure
U.S. public equities, long only
Rebalancing
Thesis-driven. No calendar rebalancing.
Benchmark
S&P 500 Total Return (for comparison only)
Custody
Interactive Brokers, LLC — client-owned SMA
View Performance →← Back to Strategy