Bullish and bearish opinions expressed in this episode, paired with supporting transcript quotes. The quote confirms what was said—not whether the opinion is correct.
Bullish
$LOVABLE— Lovable is seeing explosive growth with 770,000 applications created per week, 80% non-US usage, and 80% non-engineer users. Cuban is an investor and sees it as a key enabler for global entrepreneurship through AI-powered app development.
$AIMODELS— World models (video-based AI) are seen as the future beyond text/image LLMs. Cuban invested in Yan LeCun's world model company (AMI) and Matter.com for satellite-based world model data, viewing video processing as critical next frontier.
$OPENEVIDENCE— Cuban invested in Open Evidence, an AI tool for medical optimization that he personally uses. Sees major opportunity in AI-powered healthcare combining wearables data with blood testing for personalized health insights.
$APPLOVIN— AppLovin highlighted as success story that started with $8 domain and no VC funding, became one of largest ad platforms. Now powers ecommerce ads with strong performance metrics (one brand went from $4M to $16M and turned profitable).
Bearish
$DATACENTER— Cuban warns data center buildout is 'planning for perfection' with massive debt. Expects technological breakthroughs in AI efficiency will make many data centers obsolete, comparing to fiber optic overcapacity. If wrong on data centers, it will be because of video/world models.
$VC— Cuban warns current AI bubble 'could just destroy a lot of VCs and a lot of funds and a lot of PE' who are going all-in at high valuations. Entry price matters and many deployed at peak valuations, with companies spending cash flow on capex then borrowing more.
$GOOGL— Google and Meta are borrowing hundreds of millions to billions for AI capex despite having cash flow, indicating financial strain. This debt-fueled AI buildout is seen as risky 'planning for perfection' that may not generate expected returns.
$META— Meta borrowing billions for AI infrastructure despite having cash flow, spending all cash flow on capex then borrowing more. Part of risky debt-fueled AI buildout that Cuban sees as 'planning for perfection.'
$PALANTIR— Cuban suggests Palantir's forward-deployed engineer model (which requires 4 engineers per client) reveals AI's limitations and that competitors like Microsoft, Anthropic, and OpenAI are doing the same thing, undermining Palantir's supposed differentiation.