A concentrated book, led by the AI build-out.
In the public markets we run a concentrated long/short book that spans several theses. The biggest today is AI. The model layer is only the visible tip. Behind it is a much larger reorganization of the physical economy: power, data centers, memory, networking. We go long the businesses that compound through that shift and short the ones it leaves behind. The book stays long-biased and diversified.
Our theses
These are a few of the separate bets we run, not all of them. The ones we're most confident in get the most money, and the rest get less. But we never let any single one get too big.
The physical layer of AI
The memory, leading-edge chips, and sensors AI can't scale without.
Software-native finance
Digital banks and AI-driven lenders taking share from incumbents they undercut on cost.
What AI makes cheap
Businesses whose core product is the routine knowledge work AI now does for almost nothing.
Overvalued consumer
Premium consumer and restaurant brands still priced for growth that's already fading.
Companies we follow
We follow a wide set of public companies and write about how we think in our research. A few of them:
- MicronMU
One of three major makers of the world's DRAM, and a key supplier of the high-bandwidth memory AI accelerators run on. Memory is now a consolidated business with real pricing power when supply is tight. The market still prices it like a boom-bust commodity cyclical, which understates how structural AI demand has become.
- SanDiskSNDK
The NAND flash and storage business spun out of Western Digital in 2025, now standing on its own. Flash sits in a consolidated industry where AI-driven data growth keeps tightening supply. Coming out of a cyclical trough as a fresh spin-off, it's under-covered and priced below what normalized flash earnings would support.
- GoogleGOOGL
Search, YouTube, and advertising fund a fast-growing cloud business and DeepMind, a frontier AI lab. Few companies own the whole AI stack: the chips, the models, the data, and the distribution to reach billions of people. Fears that AI will eat search keep the multiple modest, so the AI upside comes close to free.
- DaveDAVE
A consumer fintech app offering banking and small cash advances to Americans the big banks underserve. Its AI underwriting keeps loss rates low, and the business has scaled into real profitability. The market still lumps it in with subprime lenders, which misreads both its growth and its improving credit.