Yesterday I appeared on the Saudi channel Ansharq, which covers business news in a partnership with Bloomberg. It’s a different experience answering questions with real-time human translation going in both directions — this must be what it feels like for diplomats at the U.N. — but it was an interesting conversation about the sustainability of current AI spending and the potential impacts of AI on job markets worldwide. You can watch an edited clip of the conversation (primarily in Arabic) via the link below, or read on for an AI-generated, human-edited overview and summary of the full discussion.

Link to edited and translated segment.

Overview

Patrick Moorhead assesses the sustainability of high AI capital expenditures and addresses market parallels to the dot-com bubble. Unlike the dot-com era, which was plagued by unmonetized investments and massive unused dark fiber, current datacenter capacity is near full utilization, and major players like Anthropic are demonstrating strong profitability and cash flows alongside ongoing downstream demand. While debt financing and high buildout costs through 2028 remain monitored risks, Moorhead emphasizes that AI adoption is still in its infancy, with only a small fraction of business workloads and consumer use initiated. He expects emerging labor market impacts to drive job shifting and continued political scrutiny.

Outline

AI Capex Sustainability and Downstream Demand

• The Asharq interviewer introduces Patrick Moorhead and asks how AI companies can sustain their gains amid massive capital expenditure announcements and when the market can test their true capabilities.
• Moorhead identifies downstream demand as the primary question and notes that firms including Anthropic, Microsoft, and Alphabet are answering it through earnings and software integration.
• He highlights credit and financing as the second critical issue, raising the question of whether industry buildout can continue funding at current levels into 2027 and 2028.

Comparisons to the Dot-Com Crash and Datacenter Utilization

• The interviewer notes market anxiety surrounding the rising cost of debt and insurance for hyperscalers, drawing comparisons to the dot-com era and the Great Financial Crisis.
• Moorhead contrasts the dot-com bubble’s 97% unused dark fiber capacity with today’s datacenter capacity running near 100% utilization.
• He points out that while VC-backed firms consume significant capacity today, early dot-com firms like Amazon were wildly unprofitable for years.
• Moorhead cites Anthropic’s reported operating profits and billions in operating cash flow as evidence of fundamentally healthier financials.
• He concludes that these structural differences make him very bullish on the current AI buildout compared to past speculative cycles.

Labor Market Disruption and Future Adoption Trajectory

• The interviewer raises the issue of labor market consequences, asking why potential job losses caused by AI productivity are seemingly being ignored by the market.
• Moorhead states that widespread job reductions have not materialized over the preceding nine months, although Microsoft’s recent buyout offers to 60,000 workers represent a notable development.
• He observes that AI adoption is still early, estimating that only 2% of enterprise workloads currently use AI.
• Moorhead predicts significant job shifting toward roles that require human intuition and creativity rather than pure job loss.
• He also predicts ongoing political scrutiny worldwide, noting that political issues could slow AI growth.

Companies Mentioned

• Anthropic
• Microsoft
• Alphabet
• CoreWeave
• Amazon

Keyword Summary

AI, capex, Anthropic, Microsoft, hyperscalers, dot-com era, data centers, labor market, job loss, productivity