Alphabet and OpenAI can both have massive ad businesses within AI, says Deepwater's Gene Munster
Speakers
Gene Munster
— Managing Partner, Deepwater Asset Management
Summary
Gene Munster, managing partner at Deepwater Asset Management, joins Closing Bell to discuss ChatGPT's advertising plans and the 2026 AI trade. He argues Google and OpenAI can both win in AI advertising, sees hyperscaler capex guidance as the next positive catalyst, and expects the AI infrastructure buildout to continue due to inference demand. He also stands by his call for small-cap outperformance and mentions the LOUP ETF.
- Gene Munster reacts to ChatGPT's advertising plans.
- He believes Google and OpenAI can both have massive AI advertising businesses.
- He sees hyperscaler capex guidance as the third hurdle for the AI trade in 2026.
- He expects capex growth around 40% vs. 34% consensus.
- He argues inference demand is much larger than training and will extend the AI infrastructure buildout.
- He reiterates his December call for small-cap outperformance.
- He cites the LOUP ETF as a way to play small caps.
- He notes recent TSMC news, Nvidia CES commentary, and AMD's rally as market context.
Ideas
Google and OpenAI both win AI ads.
Gene's reaction to ChatGPT's ad plans is that Google is not necessarily the loser. He thinks Google will do well in 2026 and that both Alphabet/Google and OpenAI can have massive advertising businesses within AI agentic commerce. The reason is that consumer behavior is changing: people are spending more time searching for information, whether via traditional Google search or chatbots, and Google's 800 million monthly active users and 2.5 billion search users show strong engagement. He calls it a rare case where both companies can win.
AI infrastructure buildout continues on inference.
Gene sees the AI trade's third 2026 hurdle as hyperscaler capex guidance at the end of the month. He expects capex growth of about 40%, above the recently raised 34% consensus, because investors are making a fundamental mistake by focusing on training infrastructure while the inference market is far larger. In 2025, inference was five times the size of training, and over time it could be orders of magnitude bigger, so the AI infrastructure buildout party can continue longer.
Small caps can outperform on AI revisions.
Gene stands by his December prediction that small caps can outperform. If investors believe the AI infrastructure buildout and inference opportunity are real, smaller companies should mathematically see higher and greater earnings revisions. He points to the firm's small-cap-focused ETF LOUP as an expression of this view.
