Daily Alpha · Substack
· Post-Market Alpha · by Buzzberg Research
Deep dives into enterprise AI ROI deficits and the disruptive potential of open-weight models.
Themes on this desk
AI OpEx
Enterprise AI token costs are outpacing productivity gains.
Enterprise AI token costs are outpacing productivity gains
Chamath Palihapitiya reports that internal AI token costs at his firm have been doubling every 45 days while yielding only 5-10% incremental productivity, suggesting that many enterprises face significant earnings risks from unmanaged AI spending.
Unchecked AI infrastructure spend may lead to missed EPS targets for companies that are 'all in' on AI without rigorous cost-routing controls.
Watch Monitor corporate earnings reports for margin compression attributed to rising AI-related OpEx.
Source →Open-weight models are challenging frontier model dominance
Open-weight models are increasingly matching the performance of closed-source frontier models at a fraction of the cost. Examples include Moonshot AI's Kimi K3 and Thinking Machines Lab's Inkling, which achieved higher accuracy on specific tasks than frontier models at significantly lower costs.
The commoditization of intelligence threatens the business models of closed-AI providers and shifts the economic burden of the $1.4T AI infrastructure buildout away from enterprises.
Watch Track adoption rates of open-weight models versus proprietary APIs in enterprise workflows.
Source →