Inside Hudson River Trading's Blistering Token Burn – Odd Lots

Inside Hudson River Trading's Blistering Token Burn – Odd Lots — Transcript

Market views

Bullish

NVDA— NVIDIA maintains dominant moat in AI training chips with no clear competitors, while also expanding into inference through Grok acquisition. Described as having an unassailable competitive position.
DATACENTER— Data center capacity and power are the primary bottlenecks in AI infrastructure, not chips themselves. Demand far exceeds supply with long-term contracts being signed at premium prices.
POWER— Power availability is identified as the true scarce resource constraining AI infrastructure buildout, with all major cloud providers competing for limited capacity.
AIMODELS— AI model capabilities are advancing exponentially with measurable productivity gains (50%+) and expanding use cases. Token spending growing rapidly as models prove their value in production environments.
BROADCOM— Broadcom positioned as key partner for custom inference chip development, with multiple firms pursuing partnerships for proprietary AI inference solutions.

Bearish

AILABS— Large AI labs face talent retention challenges and cultural shifts as they reach trillion-dollar valuations with limited remaining upside, creating opportunities for competitors.
HFT— High-frequency trading margins face structural compression as multiple firms simultaneously invest heavily in AI capabilities, creating a competitive race with diminishing returns.

Not investment advice.