Milk Road AI OpenAI’s Revenue Ramp Kills the AI Bubble Narrative w/ Duncan & Patrick
Jan 23, 2026
A deep dive into OpenAI’s compute-to-revenue ramp and why soaring spending now looks like infrastructure, not a bubble. They debate where massive capital will come from and how global data centers will shape training versus inference. Geopolitics and chip export risks get scrutinized. Predictions on AGI timelines, new consumer AI devices, and who stands to benefit as AI scales round out the conversation.
AI Snips
Chapters
Transcript
Episode notes
Compute Drives Revenue Growth
- OpenAI's compute ramp has a matching revenue ramp, turning energy into monetizable intelligence.
- Revenue scaled from $2B (2023) to a projected $20B (2025) as wattage grew from MWs to GWs.
Energy Into Intelligence Is The Supertrend
- Converting compute and energy into intelligence is the central long-term trend powering AI value creation.
- New massive data centers coming online in 2026 will extend that conversion at scale.
Big Cash Reserves Calm Funding Fears
- OpenAI ended the year with over $40 billion in cash, easing short-term funding fears.
- That cash cushion reduces immediate risk despite massive future spending commitments.