The Everything Feed - All Packet Pushers Pods

HS128: Planning for an AI Bubble Burst

·34 min·3 clips

Jonna Johnson and John Burke ask if AI is a bubble and what enterprise IT must do to prepare for a potential burst.

The episode opens with hosts Jonna Johnson and John Burke framing the central question: is AI a bubble, and what should enterprise IT do to prepare given the uncertainty. They reference a previous episode where setting an AI strategy was a top New Year's resolution for 2026 due to bubble concerns. John Burke kicks off by asking if recent months have reduced bubble worries, highlighting circular investing as a symptom, where NVIDIA invests in OpenAI to drive chip sales. Jonna Johnson elaborates on circular investing, noting it indicates a lack of viable market if companies must pay customers. She contrasts OpenAI's $1 trillion investment commitment with actual AI industry profits, which are orders of magnitude smaller, suggesting shaky economic underpinnings. The hosts discuss how hyperscalers like Microsoft and Google are embedding AI into existing offerings to generate profit, rather than selling it standalone. They debate whether AI is a product or a feature, comparing it to the dot-com boom where the internet became integrated into commerce. Jonna Johnson points out that many AI-focused vendors aren't profitable, relying on venture capital. The conversation shifts to bubble indicators, such as Super Bowl ads for hot tech, referencing pets.com and Bitcoin. They explore the impact of a bubble burst, including potential widespread unemployment and reduced consumer spending, affecting even non-AI businesses. John Burke emphasizes the need for business cases in AI deployment, while Jonna Johnson pushes for simpler use-case evaluations and cost sensitivity analyses. They discuss vendor lock-in risks, advising portability between AI platforms like OpenAI and Anthropic. The hosts touch on claims about AI automating white-collar jobs, citing Microsoft's Mustafa Suleiman's prediction, but note Wall Street skepticism. They highlight the importance of not assuming AI costs will decrease, using Enron's collapse as a cautionary tale. The episode concludes with advice on fire-hardening strategies, focusing on vendors with clear profitability paths, and considering internal AI hosting with smaller models. Jonna Johnson clarifies that LLMs are a subset of AI, not equivalent to AGI, and stresses understanding AI's capabilities and limitations for better strategic planning.

As heard by us

A grounded enterprise check on AI plans built to survive either a bust or continued expansion.

The AI bubble question lands here as an enterprise planning problem. Jonna Johnson, CEO of Nemertis, and John Burke, CTO of Nemertis, focus less on calling the top than on what IT teams can do when no one can prove whether a bubble exists or when it might burst.

Read the full review in PlayNext →

Why you'd press play

You want an AI strategy that can hold up if the bubble talk turns into a real burst.

Read the full recommendation in PlayNext →

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from The Everything Feed - All Packet Pushers Pods