Henry Blodget on the Software Selloff Hysteria and the Problem for OpenAI

March 7, 2026

45 min episode · 2 min read

Henry Blodget

Episode

45 min

Read time

2 min

Topics

Artificial Intelligence, Software Development

AI-Generated Summary

Published Mar 8, 2026

Key Takeaways

✓AI Market Volatility: Software stocks declining in 2025 reflect elevated valuations meeting uncertainty, not fundamental collapse. The fear that Claude Code eliminates enterprise software demand ignores that businesses require accountability, support, and reliability that a junior employee vibe-coding replacements cannot provide. Treat sharp single-day selloffs driven by Substack newsletters as sentiment events, not structural signals.
✓OpenAI Competitive Moat Risk: When Google Gemini matched or surpassed GPT within two years of ChatGPT's launch, it signaled that OpenAI lacks the durable network effects Google had over Yahoo. Investors should note that of hundreds of 1990s dot-com IPOs, only Amazon generated sustained long-term returns — AI leadership is far from decided.
✓OpenAI Unit Economics Problem: OpenAI loses money on power users and is raising hundreds of billions in open markets while Google, Microsoft, and Meta generate tens of billions in annual free cash flow to self-fund compute. Bulls argue costs will fall and revenue lines will cross, but at an $800 billion valuation, the required revenue scale is enormous.
✓Private AI Stock Warning: When intermediaries cold-call offering Anthropic shares at a $350 billion valuation through SPVs with no financial disclosures, treat it as a late-cycle signal. Private secondary markets charge extra fees, offer limited liquidity, and provide less transparency than public companies at IPO — the opposite of what retail investors typically assume.
✓Media Survival Framework: Publishers with direct subscriber relationships, differentiated expertise, and brand trust are positioned to survive AI disruption. Serendipitous content discovery — readers finding stories they did not know they wanted — remains a human editorial skill AI does not replicate. Niche, expert-driven outlets and established brands with accountability hold structural advantages over generic content producers.

What It Covers

Henry Blodget, CEO of Regenerator and former Business Insider chief, joins Odd Lots at On Air Fest 2026 to assess AI market hysteria, OpenAI's $800 billion valuation problem, the collapse of software stocks, and what AI disruption actually means for media business models and journalism jobs.

Key Questions Answered

•AI Market Volatility: Software stocks declining in 2025 reflect elevated valuations meeting uncertainty, not fundamental collapse. The fear that Claude Code eliminates enterprise software demand ignores that businesses require accountability, support, and reliability that a junior employee vibe-coding replacements cannot provide. Treat sharp single-day selloffs driven by Substack newsletters as sentiment events, not structural signals.
•OpenAI Competitive Moat Risk: When Google Gemini matched or surpassed GPT within two years of ChatGPT's launch, it signaled that OpenAI lacks the durable network effects Google had over Yahoo. Investors should note that of hundreds of 1990s dot-com IPOs, only Amazon generated sustained long-term returns — AI leadership is far from decided.
•OpenAI Unit Economics Problem: OpenAI loses money on power users and is raising hundreds of billions in open markets while Google, Microsoft, and Meta generate tens of billions in annual free cash flow to self-fund compute. Bulls argue costs will fall and revenue lines will cross, but at an $800 billion valuation, the required revenue scale is enormous.
•Private AI Stock Warning: When intermediaries cold-call offering Anthropic shares at a $350 billion valuation through SPVs with no financial disclosures, treat it as a late-cycle signal. Private secondary markets charge extra fees, offer limited liquidity, and provide less transparency than public companies at IPO — the opposite of what retail investors typically assume.
•Media Survival Framework: Publishers with direct subscriber relationships, differentiated expertise, and brand trust are positioned to survive AI disruption. Serendipitous content discovery — readers finding stories they did not know they wanted — remains a human editorial skill AI does not replicate. Niche, expert-driven outlets and established brands with accountability hold structural advantages over generic content producers.

Notable Moment

Blodget revealed he spent considerable time writing a novel manually, then learned a friend had used Claude to generate a comparable 325-page manuscript in roughly twenty minutes. He called himself the world's biggest idiot from a business standpoint, yet argued the learning process itself retains value AI cannot replicate.

Know someone who'd find this useful?