Crossover Investor

At the intersection of
AI infrastructure
and frontier capital.

I invest in AI infrastructure. My edge is having operated inside three layers of the stack before I started.

Before my current role, I co-founded an AI company backed by Sequoia-SourceCode Capital, served as acting CFO of an energy-backed AI infrastructure company developing power assets in Texas, and pioneered a new direct enterprise channel at HP Asia from zero to $200M in revenue.

Selected investments: Nebius · Databricks · Anthropic.

Read the Essays Get in touch →
Coverage · AI Infrastructure Stack
Physical Layer Power · Silicon · Packaging Focus
Semiconductor GPU · ASIC · HBM
Neocloud / Infra CoreWeave · Lambda Focus
Inference Serving Baseten · Together Focus
Foundation Models OpenAI · Anthropic
Applications Vertical AI · Agents
Investment Edge

Signals others don't see

Edge 01
Private compute economics

Access to unit economics from private infrastructure companies — GPU utilization rates, inference margins, and capex payback periods that never appear in public filings.

Edge 02
Asian supply chain dynamics

On-the-ground visibility into advanced packaging, memory, and optical component supply chains across Taiwan, Japan, and South Korea — the bottlenecks that determine the pace of AI infrastructure buildout.

Edge 03
Operator perspective

Having developed power assets for AI infrastructure in Texas and co-founded an AI company, I understand infrastructure economics from the inside — not just the balance sheet.

Research

A multi-essay framework for AI infrastructure value capture

Featured · July 2026 12 min read
After Talking to 15 Silicon Valley AI Companies, I See the Open vs. Closed Debate Differently

Two days in Menlo Park with Cursor, Fireworks, Cognition, Sierra and a dozen other AI-native companies. From Token Optimization to Model Routing — why AI's competition is shifting from Capability to Economics, and what it means across Frontier Labs, semis, cloud, and AI software.

Read the research
Field Research
12 min read
Where Is the Next Wave of AI Wealth?

A 100-year pattern has governed every major information industry. Wealth flows in three waves: infrastructure first, then compression, then gatekeepers. AI is running the same script.

Read the essay
8 min read
Why Traditional Value Frameworks Missed the AI Bottleneck Supercycle

2025 was a great year for performance. It was also the year I realized how much I had still missed — and why the framework shift AI forces on investors is more uncomfortable than most will admit.

Read the essay
7 min read
In the Age of AI, Collapse Isn't Gradual — It's Instant

SaaS companies lose market share. AI companies get erased. A note on paradigm velocity, taste, and what kind of VC (and LP) actually survives when the floor disappears overnight.

Read the essay
中文 · CN 12 min read
硅谷最有创业精神的人，开始不创业了

2026 年最有创业精神的动作，也许就是：加入那个可能让"创业"本身过时的东西。一篇关于范式切换、护城河，与那个最理性的"孤注一掷"的中文长文。

阅读全文 · Read
Topics

What I write about

AI Infrastructure Economics
Capital flows, margin structures, and unit economics across the compute stack
Supply Chain Dynamics
Advanced packaging, memory, optical — the physical constraints on AI scaling
Historical Patterns
What telegraph, telephone, and mobile data cycles tell us about where AI wealth concentrates
Private Market Signals
Reading neocloud and inference company economics before they reach public markets
Power & Physical Layer
Energy infrastructure, data center development, and the atoms behind the bits
Crossover Investing
How public and private market signals interact across the AI infrastructure cycle
Speaking

Selected talks

2025
Energy and Data Centers: The Biggest Opportunities?
Alignment Summit · Silicon Valley
Panel discussion on datacenter bottlenecks, energy infrastructure constraints, and emerging opportunities in space-based infrastructure for AI.
Panel at Alignment Summit 2025
2026
Scaling AI: The Infra, Energy and Capital Triangle
Silicon Valley
Panel on data center economics in the AI era — power, chips, and capital — and where new infrastructure opportunities open up as the stack scales.
Panel on Scaling AI 2026
Background

Operator before investor

Current
Crossover Investor
AI Infrastructure · Full Stack
Focused on the full AI infrastructure stack from physical layer through foundation models. Coverage spans public equities, private credit, and venture — with emphasis on the supply chain and infrastructure layers where private signals lead public markets by 6–12 months.
Prior
Acting CFO
Energy-Backed AI Infrastructure Company · Texas
Oversaw finance and strategy for a company developing power assets specifically for AI data center infrastructure in Texas. Hands-on exposure to the physical constraints, permitting dynamics, and capital structures of AI infrastructure development.
Prior
Co-Founder
AI Healthcare Company · Sequoia-SourceCode Capital
Co-founded an AI healthcare company backed by Sequoia-SourceCode Capital. Built from zero to product, giving me the operator's view of what AI applications actually pay for from the infrastructure stack, and where margins compress as models commoditize.
Prior
Division Head, Asia E-Commerce
HP · Asia Pacific
Built HP's Asia e-commerce division from zero to $200M in revenue. Developed a deep understanding of Asian manufacturing and supply chain dynamics that now informs investment analysis of semiconductor and hardware supply chains.
Connect
"I'm most interested in conversations with founders building in AI infrastructure, operators with supply chain visibility, and investors thinking about the crossover between public and private markets."
Email → LinkedIn → X / Twitter → Subscribe →