Semiconductor-Focused Long Equity

Hardware
Before Software.

The Century of Material Science

A concentrated, long-only investment thesis on the semiconductor and AI-hardware stack powering artificial intelligence, robotics, and the next generation of quantum computing.

Explore the Thesis
AI / COMPUTE MEMORY QUANTUM WFE
Scroll
Thesis

A Pattern We've
Seen Before

Each technological revolution made capability dramatically cheaper and more abundant. Semiconductors are not a bubble — they are this era's steam engine: the hardware that powers everything built on top of it.

1800S Industrial Revolution Steam · Mechanization 1970S – 2000S Computing Revolution Moore's Law · Silicon NOW · ACTIVE AI Revolution Compute Scarcity · GPU Era ~2035+ · EMERGING Quantum Revolution Drug Discovery · Encryption
1800s

Industrial Revolution

Water and steam mills made goods cheaper and production far higher volume. Hardware defined the era.

Realized
1970s – 2000s

Computing Revolution

Moore's Law — twice the power, half the cost, every two years. Silicon became the foundation of modern commerce.

Realized
Now · Active

AI Revolution

Compute becomes the scarce input. Every model runs on physical hardware — GPUs, memory, and power silicon. Hardware leads software.

Active
~2035+ · Emerging

Quantum Revolution

A new class of chips changes science itself — beginning in drug discovery, encryption, and materials science. Too early to own, not too early to track.

Emerging
Strategy

Concentrated
Conviction

01

Long-Only, Concentrated

~15 high-conviction names. Deliberate weights sized to forward growth, not market cap.

02

US-Anchored Core

Compute, memory, and equipment leaders, plus ASML and TSMC — the irreplaceable chokepoints of global semiconductor supply.

03

Real-Asset Ballast

Gold, copper miners, rare earths, and short-term treasuries. ~15% allocation to compress tail risk while preserving upside.

04

Rules-Based Risk

Pre-committed drawdown rules. Single position down 15%: trim by 50%. Portfolio down 20%: apply index puts.

The portfolio is built around the irreplaceable nodes of the AI hardware stack. Every AI model, every data center, every autonomous system runs on chips. We invest where value is created first: in the silicon and the materials beneath it.

~85% EQUITY Logic / AI 15% Equip (WFE) 15% Analog / Auto 15% Memory 10% Foundry 10% Distribution 5% Real-Asset Ballast 15%
Logic / AI Accelerators
15%
Wafer Fab Equipment
15%
Analog / Embedded / Auto
15%
Memory (DRAM / HBM)
10%
Leading-Edge Foundry
10%
Distribution & Components
5%
Real-Asset Ballast
15%

Target weights. Illustrative only. Not investment advice.

Team

A True
Sector Specialist

Prabhu B. Iyer
Founder · CEO & CIO
PhD, Materials Engineering — RPI
BTech, Materials Engineering — NITK Surathkal
Process Support Engineer IV — Applied Materials
Executive MBA, Finance — Wharton, Class of 2027
Geographic Focus
UNITED STATES 9 of 11 names CORE POSITIONS PRIMARY ASML Netherlands TSMC Taiwan On watch: Korea · Japan

Prabhu B. Iyer founded Crystallum Capital to apply PhD-level materials science judgment to semiconductor equity investing — an edge that generalist analysts cannot replicate from a spreadsheet.

As a Process Support Engineer IV at Applied Materials, he has a ground-level view of semiconductor fabrication economics, supply chain dynamics, and which technology roadmaps are real versus aspirational. His work has contributed to over $140M in commercial impact in defect review for a major semiconductor client.

His scientific training provides structural analytical depth when evaluating the companies building the physical infrastructure of AI. The investment thesis at Crystallum is not a financial model imposed on an unfamiliar industry — it is built from the inside out.

Prabhu is completing his Executive MBA in Finance at the Wharton School, where he works with senior finance faculty on the investment framework underlying the strategy.

The Edge a Generalist Cannot Replicate
GENERALIST ANALYST Financial model Earnings calls Sell-side research Industry reports CRYSTALLUM EDGE Fab-floor variant perception PhD roadmap validation Supply chain ground truth Wharton faculty review $140M+ commercial insight

Faculty provide academic guidance; this is not an endorsement of any securities offering.

Contact

Get in
Touch

Crystallum Capital operates as a Separately Managed Account strategy. This website is for informational purposes only and does not constitute an offer to sell or solicitation of any offer to buy an interest in any account or fund.

StrategySeparately Managed Account (SMA)
FocusSemiconductor & AI Hardware
StructureLong-Only, Concentrated Equity

This website is provided for informational purposes only and does not constitute an offer to sell or the solicitation of any offer to buy an interest in any account, fund, or investment vehicle. Crystallum Capital LLC operates as a Separately Managed Account (SMA) strategy and does not engage in general solicitation or public advertising. Nothing on this site constitutes investment, legal, or tax advice. An investment in any strategy is speculative and involves a high degree of risk, including the possible loss of all or a substantial portion of capital. Any target returns or performance figures referenced are objectives only — not guarantees. Prospective investors should consult their own legal, tax, and financial advisers before making any investment decision.