Open Circuit – "How to Spot an AI Bubble"

Podcast by Latitude Media
Date: October 10, 2025
Featured Guest: Azeem Azhar, Founder of Exponential View
Regular Hosts: Stephen Lacy, Kathryn Hamilton
Guest Host: Scott Clavenna (Latitude Media CEO)

Overview:

This episode dives deep into the question that’s looming over the tech and energy sectors: Are we entering or already inside an AI bubble? The discussion leverages experience from past tech booms and busts, examines current AI infrastructure investment trends, and explores the implications for the energy transition. Industry veteran and special guest Azeem Azhar brings a structured, data-driven perspective to differentiate between a “boom” and a “bubble,” while the hosts probe the downstream consequences for the energy grid, customer rates, and technological innovation.

Key Discussion Points & Insights

1. Opening Thoughts: Useless AI, “AI Slop,” and Facebook Feeds

• The show kicks off with a light anecdote about “dumbest uses of AI” to set the tone.
  • [00:36] Azeem Azhar (C): “I said to ChatGPT, I want you to talk back to me in cat language and cat language only. So 15 minutes of discussion where I would ask it questions and it would obligingly meow back to me. I think that's pretty dumb.”
• This launches a discussion of how AI has pervaded daily life and signals a broader societal fascination (or fatigue) with generative AI.

2. Are We in an AI Bubble?

• Thesis: AI infrastructure investment has exploded, potentially outpacing all US consumer spending in 2025, and data centers are “sucking up capital from other sectors.”
  • [01:19] Stephen Lacy (B): “Tech firms are building the equivalent of an Apollo program every 10 months...are we in the middle of an AI bubble?”
• The hosts draw parallels to prior bubbles (dot com, clean tech, telecom) and introduce guest Azeem Azhar for a historical, comparative take.

3. Lessons from Old Bubbles—And the Red Flags

• Scott Clavenna reflects on living through the dot-com and clean tech bubbles.
  • [04:48] Scott Clavenna (E): “One of the things I learned right away was that everyone lies to you...everyone has a story to tell, everyone is myth making, and everyone to some extent is lying.”
• Recognizing “bubble signals”:
  • Obscure or risky debt structures
  • Vendor financing echoing the telecom days
  • Companies IPO’ing on hype not product
  • High-profile, repeat bubble participants (e.g., Masayoshi Son)
  • [22:22, 27:03] E: Details his “six horsemen of the bubble apocalypse”

4. Azhar’s “Bubble Gauges” – Diagnosing the AI Economy

• Azeem Azhar shares a structured, data-driven framework for evaluating bubble risk:
  • Economic Strain: How much AI investment relative to GDP?
    • [13:56] “3% seems to be a really terrible number to reach unless you're in a war...Today with data centers and AI, we're roughly at a percent...green right now, but it's trending up towards amber.”
  • Industry Revenue Coverage: Is AI spend justified by current revenue?
    • [14:45] “We're covering about 16, maybe 20% of CapEx by the revenue that's being generated...definitely amber.”
  • Revenue Growth: Is revenue growth outpacing CapEx?
    • [15:32] “That 60, $80 billion this year was essentially nil in 2022...very, very healthy gauge right now.”
  • Market Valuations: Are valuations reasonable or into “modicum of insanity” territory?
    • [16:29] “They’re high...but they're not outrageous.”
  • Funding Quality: Is funding from strong balance sheets, or is debt/complexity creeping in?
    • [17:16] “Much of the capital is coming from company balance sheets...not a lot of debt...except for companies like OpenAI, Coreweave, XAI, where it's more stressed.”

Memorable Quote

[11:40] Azeem Azhar (C): “If you get three red lines, you're definitely in trouble. If you get two, you might want to tighten your seatbelt … Right now, a lot of these are green, a couple are amber… It’s a mixed picture.”

5. Revenue Realities vs. Hype

• Kathryn Hamilton challenges the “revenue model” narrative with MIT research showing most GenAI pilots aren’t generating revenue.
  • [18:29] D: “The most ability to get revenue is going to be really not on the super sexy stuff, but on...the incremental improvements, the specific problem-solving that IT can do...”
• Azhar’s counterpoint:
  • [19:39] “...for many companies in America and across the world today, that's the thing they can do with AI. It's genuinely difficult to make that move across to the Henry Ford operation, but maybe that will happen over the next three to six years.”

6. The Impact on Energy Infrastructure & the “Dark Fiber” Analogy

• Stephen Lacy queries the risk of stranded assets if the AI infrastructure boom stalls—will “half-built substations” be this decade’s “dark fiber”?
  • [31:17] B: “Are they stranded substations, half-built data campuses…if we do see a significant market correction?”
• Azhar:
  • [31:17] “They [data centers] are providing a much needed stimulus, a demand stimulus for innovation...Our welfare, our prosperity, our health has been connected to our energy capture...that’s a really, really good Thing.”
• Scott Clavenna:
  • [33:02] “The last thing the US power system is today is overbuilt. ... It's very much constrained. So it's hard to see… power infrastructure stranded assets. I could see data center digital infrastructure stranded assets ... but power, my God, we need a lot more.”

Notable Moment

[36:00] Kathryn Hamilton: “It's going to be hard to tell customers, homeowners, small businesses, whose rates are skyrocketing, 'don't worry, eventually you’ll be able to use this infrastructure.'...People are going to start rebelling against their rates going up.”

7. Cultural and Regulatory Roadblocks in Utilities

• The challenge isn’t just technical—utilities are slow to change, resistant to new orchestration platforms, and fixated on legacy systems.
  • [50:35] Kathryn Hamilton: “It's really hard for them to get out of their own way and try something new and different...you'll find much more disruption on the edge of grid side with folks who are working with companies like Nvidia to try to make a difference whether or not the utility wants them to.”
• Scott Clavenna:
  • [49:38] “AI could be sort of a middle layer that quickly and intelligently connects different systems, subsystems within a utility … That struck me as an excellent opportunity.”

8. Where Are the “AlphaFold Moments” in Energy?

• Stephen Lacy:
  • [52:17] “We really haven't seen yet like the AlphaFold moment in energy yet...do you think we're at that point?”
• Azhar:
  • [53:56] “Project Contrails would be able to reduce contrail-related global warming by...half a percent of our total emissions...we've got one, we just haven’t done anything with it.”
  • On materials: “It's much harder than we think … Turning that from press release to millions of tons...is going to take quite a lot of time.”

9. Do the Social/Environmental Benefits of AI Outweigh Its Costs?

• Big Picture Debate:
  • [57:34] C: “Do we believe in the idea that technological progress in general can be supportive of human welfare, given whatever environmental strain it causes?”
  • “The Degrowth agenda is...not sustainable...we have to learn from moments where we have been able to deal with this.”
• Scott Clavenna:
  • [59:59] “Right now...we need strong, strong leadership...AI has a tremendous amount of potential, but boy, it's very scary too.”

10. Is This a “Transformative Moment” Like Prior Industrial Shifts?

• Azeem Azhar:
  • [61:59] “It does feel similar to those transformations and it also is still open in terms of the direction that it takes...Without moderation, without policy intervention, that Gini coefficient will become higher and higher...Depending on the policy choices, I think historians will therefore write this about this differently.”

Notable Quotes & Timestamps

• [04:48] Scott Clavenna: “Everyone has a story to tell, everyone is myth making, and everyone to some extent is lying. And I wasn’t quite prepared for how pervasive that was.”
• [11:40] Azeem Azhar: “If you get three red lines, you’re definitely in trouble. If you get two, you might want to tighten your seatbelt ... right now, a lot of these are green, a couple are amber.”
• [31:17] Azhar: “They are providing a much needed stimulus, a demand stimulus for innovation, for investment into a sector that we all depend upon.”
• [53:56] Azhar: “We’ve probably learned, we’ve learned from what prevented us scaling up the first set of materials of this type … they may not be perfect, but they may get us to that answer more quickly than we might have otherwise done.”
• [61:59] Azhar: “Depending on the policy choices, I think historians will therefore write this about this differently ... they might talk about how some people got very wealthy but the baseline level of human experience improved dramatically. Which historian will write that story I think is going to depend on policy choices over the next five to ten years.”

Timestamps for Key Segments

• Opening and Intro: 00:02 - 03:40
• Bubble vs. Boom Discussion Begins: 04:22 - 10:20
• Azeem Azhar’s Bubble Gauges Explained: 13:32 - 18:29
• Energy Infrastructure/Stranded Asset risk: 31:17 - 36:00
• AI in Utility Culture and Practices: 49:10 - 52:17
• AlphaFold Moment/AI in R&D: 52:17 - 56:49
• AI’s Societal Benefit vs. Environmental Cost: 57:34 - 59:59
• Grand Historical Perspective: 61:32 - 63:29

Engaging Moments & Final Thoughts

• Hosts reflect on historic cycles: The group consistently compares the present AI boom to previous historic shifts, reminding listeners not to let their “bubble sensors” dictate every narrative.
• Energy optimism amid warnings: Despite concerns about speculative fervor and grid strain, all panelists agree on the need for greater investment and innovation in energy infrastructure—and see AI as a catalyst.
• Call for leadership: The guests and hosts agree that the outcome of the current boom—speculative bubble, transformative leap, or both—will rest heavily on policy and leadership as much as on technology itself.
• Cultural inertia is as great a challenge as technology: The slow adoption of AI in utilities is more about organizational inertia than technical limitation.

Conclusion

“How to Spot an AI Bubble” offers a nuanced, experience-rich perspective on the current AI wave. Rather than simply warning of a bubble or hyping a boom, the discussion provides listeners with a rigorous analytical framework for gauging market health, grounded in data and history. The greatest risks may yet be cultural and political, not technical. The episode delivers reassurance, caution, and a sense of anticipation: the outcome—much like previous tech revolutions—remains unwritten.