
Aug 14, 2026 · 54 min
AI’s growth story collides with compute costs and bubble fears
OpenAI’s Explosive Growth Continues
The episode tests whether extraordinary AI revenue and investment can produce durable returns while reshaping labor, media, and technology markets.
- 1OpenAI’s accelerating revenue depends on scarce, extraordinarily expensive computing power, complicating comparisons with Anthropic and other AI firms.
- 2Investors debate whether semiconductor and hyperscaler valuations reflect durable demand, circular financing, or concentration risk among major AI customers.
- 3AI’s expansion reaches beyond models into robotics, delivery work, venture capital, and podcast media businesses facing difficult scaling economics.
Don't miss
The robotics discussion asks whether machines are approaching a ChatGPT moment while confronting the practical limits of humanoid designs.
The brief
OpenAI’s consumer and business revenue is accelerating, with Codex contributing to growth, but the discussion stresses that computing power remains scarce and enormously expensive.
The central investment question is whether semiconductor and hyperscaler valuations anticipate durable returns or a bubble built on long horizons, circular financing, and concentrated customers.
Google DeepMind’s Carolina Parada examines whether robotics is nearing a ChatGPT moment, weighing robots’ environmental understanding against the practical advantages of specialized machines.
Kate Bronfenbrenner explains how New York’s proposed Delivery Protection Act could affect Amazon’s delivery partners, covering pay, safety, liability, and possible operational workarounds.
The episode widens the lens to AI venture capital, where a new $100 million fund is targeting agents, robotics, infrastructure, and applications that perform real jobs.
Its final business lesson comes from Alex Cooper’s Unwell: a $500 million valuation and beverage ambitions show how difficult it is to scale a podcast-led media brand.
Listen to the full episode and explore every guest, topic, and moment on PodLume.

Google DeepMind
OpenAI
Kalshi