TLDR
The podcast discusses the market fallout from the U.S. midterm elections, where widespread voter opposition to data centers triggered a sharp sell-off in AI stocks and raised doubts about Anthropic's upcoming IPO valuation. Hosts Alex and Ranjan debate whether political constraints will permanently slow AI infrastructure build-out, analyze a Meter report suggesting Anthropic's R&D pace has doubled due to internal AI use, and examine the role of prediction markets in election coverage, questioning if they add independent insight or merely reflect pollster models.
For you
PRO members can sign in and request more information — or their own version of this digest, tuned to what they care about.
Sign in to unlockSurprising Insights
Local voter referendums and gubernatorial races have become a more immediate bottleneck for AI infrastructure than federal policy or compute limits.
Anthropic's internal AI usage may be doubling its own research velocity, creating a recursive improvement loop that is not fully visible externally.
Prediction markets displayed prices (e.g., 97-99 cents) indicating near-certainty in toss-up races before official calls, yet may not represent novel information aggregation.
Key takeaways
Midterm election results showed strong anti-data center sentiment across 38 jurisdictions, causing AI stock volatility and challenging the assumed infrastructure build-out underpinning high valuations.
Political capacity at the local level has emerged as a major new constraint on data center expansion, potentially more impactful than federal regulation.
Anthropic's IPO valuation expectations have dropped from ~$2T to ~$1.2-1.5T amid market uncertainty, though the company may still proceed to avoid slowing growth narrative.
Meter research indicates Anthropic's model improvement rate doubled as of October 2024, with ~80% of acceleration attributed to internal AI use, though methodological transparency is limited.
Prediction markets called races before traditional networks but likely piggybacked on pollster models rather than discovering independent signals.
Notable quotes
“The data center build-out and the absolute scale that was required to make any of these numbers make sense almost certainly will not happen.”
▸ Notable quotes · 8:40“Everything depends on perfect execution. That's on their own internal businesses. That's on growth. That's on revenue. That's on unit economics. But that is also on the data center story to be the infrastructure for this gigantic story and all this promise.”
▸ Notable quotes · 16:08“The prediction markets almost in every race, I think it was all the toss-ups, effectively called them well before the networks did themselves.”
▸ Notable quotes · 52:55“Polling has always been kind of weird to me. Like, especially with advances in technology today, the idea, especially when it was still kind of, like, phoning people to ask them questions, was crazy.”
▸ Notable quotes · 56:25Chapters
0:00 — Intro & Sponsor Ads: Podcast introduction with sponsor reads for Cribl and Chargebee, followed by host Alex welcoming listeners from London and previewing topics: AI stock sell-off post-midterms, Anthropic IPO uncertainty, and prediction markets debate.
1:30 — Midterm Election Impact on AI Stocks: Discussion of how Democratic sweep and 38 jurisdictions restricting data centers triggered AI stock volatility. Market initially rose on hopes of spending discipline, then fell 4% as data center constraints threatened AI growth narrative.
5:00 — Data Center Restrictions & Political Constraints: Debate on whether local bans will permanently curb data center expansion or merely force negotiation. Ranjan argues political capacity is now the binding constraint; Alex suggests companies may relocate to willing communities and that restrictions may be overstated.
10:00 — Anthropic IPO Valuation Uncertainty: Analysis of Wall Street Journal reports that major investors doubt Anthropic can achieve $1.5T valuation. Leaked $80B run-rate figure criticized as improper pre-IPO communication. Kalshi odds for IPO announcement before December dropped from 90% to 57%.
20:00 — Meter Report: Anthropic R&D Acceleration: Review of Meter finding that Anthropic's model improvement rate doubled as of October 2024, with 80% attributed to internal AI use. Hosts debate significance given redactions and timing near IPO, while acknowledging visible acceleration in AI capabilities like video editing.
25:00 — Prediction Markets vs Pollsters Debate: Discussion of prediction markets (Kalshi) outperforming networks in calling races. Nate Silver and Chris Hayes argue markets merely tracked pollster models. Hosts note reflexivity concerns and ethical unease with gamblification of elections.
30:00 — Closing & Personal Banter: Lighthearted exchange about Alex's chair, Ranjan's marathon completion, and plans for next week's episode with M.G. Siegler.
Speakers
Alex — Host
Ranjan — Guest/Co-host