Odd Lots

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

March 7, 2026

Key Takeaways Copied to clipboard!

  • Henry Blodget views the current AI disruption as analogous to the early 1990s internet boom, characterized by a wide range of predictions swinging between catastrophe and euphoria, leading to market twitchiness. 
  • The narrative that AI will destroy all legacy software companies is dismissed as hysteria, as Blodget argues that enterprise software needs accountability and customization that current generative AI cannot fully replace. 
  • OpenAI's massive spending costs, coupled with the incumbent Google catching up quickly with Gemini, suggest that the company is far from guaranteed to be the dominant 'Google of the AI era,' especially given the high capital expenditure required to maintain pace. 

Segments

AI Hysteria and Market Flip
Copied to clipboard!
(00:04:42)
  • Key Takeaway: The market narrative has violently flipped from AI overvaluation concerns to fears of AI destroying all software companies.
  • Summary: The discussion contrasts the previous year’s skepticism about high AI valuations with the current fear that AI will render legacy software businesses obsolete. Henry Blodget notes the OpenAI valuation climbed from $300 billion to $800 billion during this period. This rapid shift reflects uncertainty in predicting the scale of the AI opportunity.
Impact of Doom Piece
Copied to clipboard!
(00:06:27)
  • Key Takeaway: A single thought-piece predicting economic crisis due to AI caused a dramatic, albeit temporary, market reaction, highlighting current market sensitivity.
  • Summary: The market reacted strongly to a Substack piece predicting economic crisis by 2028 due to widespread AI adoption leading to joblessness. Blodget found the market’s reaction to a mere thought experiment ‘insane,’ emphasizing that even the ‘doom case’ predicted 90% employment. This volatility shows how small changes in assumptions drastically affect current valuations when looking into the uncertain future.
Software Doom Scenario Rebuttal
Copied to clipboard!
(00:08:45)
  • Key Takeaway: The bear case that generative AI will eliminate the need for traditional enterprise software purchases is considered hysteria.
  • Summary: Blodget strongly disagrees with the idea that companies will stop buying established business software because they can generate code via AI agents. Enterprise software requires accountability and support, which a junior person using an AI agent cannot guarantee when business operations depend on it. The market is currently overreacting to the potential for disruption rather than the reality of enterprise adoption.
AI vs. Dot-Com Comparison
Copied to clipboard!
(00:10:13)
  • Key Takeaway: The primary risk of AI is potentially catastrophic agentic systems, not mass unemployment, as historical technological transitions created more jobs overall.
  • Summary: Blodget argues that fears of total job elimination overlook historical precedent where technological shifts, like agriculture to industrial, ultimately led to job growth. While disruption causes pain, the economy historically creates new jobs to replace those lost. He suggests the greater, though less discussed, risk is the invention of an agentic system that causes significant real-world damage.
AI Stickiness and OpenAI’s Lead
Copied to clipboard!
(00:13:46)
  • Key Takeaway: Unlike the internet era’s strong network effects, current AI models lack inherent stickiness, evidenced by Google’s Gemini rapidly closing the gap with OpenAI.
  • Summary: The argument that AI lacks the strong network effects seen in Web 2.0 suggests switching costs between models like ChatGPT and Claude are low. The fact that incumbent Google caught up to OpenAI in less than two years challenges the thesis that OpenAI has an unassailable lead like early Amazon did over competitors. This rapid convergence suggests the AI landscape remains highly competitive.
OpenAI’s Business Model Viability
Copied to clipboard!
(00:19:24)
  • Key Takeaway: OpenAI faces brutal economics, currently losing money on power users, and must rely on future cost plummeting or massive revenue growth to justify its high valuation.
  • Summary: Despite high valuations, OpenAI is reportedly losing money on its most intensive users, making its current business model questionable. Bulls project that costs will drop significantly, allowing profitability once market share is secured. However, the current $800 billion valuation requires revenue and profit projections far exceeding those seen at the $300 billion mark.
AI Bubble Indicators and Private Markets
Copied to clipboard!
(00:21:59)
  • Key Takeaway: The proliferation of companies rebranding as ‘AI companies’ (like the karaoke machine firm) signals a euphoric bubble phase similar to the dot-com era.
  • Summary: The trend of non-AI companies pivoting to claim AI status indicates a euphoric phase where massive R&D experimentation is funded publicly. This contrasts with the dot-com era where innovation happened largely within established corporate labs. The secondary market for private AI stock, involving SPVs and high valuations without public financials, adds complexity and fees to accessing early-stage growth.
AI’s Effect on Media Industry
Copied to clipboard!
(00:28:08)
  • Key Takeaway: AI will benefit established, trusted media brands by increasing the value of quality, differentiated content in an environment saturated with generic ‘AI slop.’
  • Summary: The media industry faces intense competition as distribution shifts back to direct subscriber relationships, making brand trust paramount. While AI can assist with drafting and research, the core value remains in reporting scoops and providing analysis on what events mean, which AI cannot replicate. Trusted brands like Bloomberg will be positioned to charge for this differentiated, verified information as LLMs seek quality data sources.
CEO Pragmatism Over Free Speech
Copied to clipboard!
(00:42:14)
  • Key Takeaway: CEOs exhibit pragmatism, prioritizing the stability of their companies and shareholder value over crusading for personal free speech beliefs, especially concerning political figures.
  • Summary: The perceived shift among CEOs toward obsequiousness toward political figures stems from pragmatism, as public company executives are easily removed if they jeopardize the business. Their primary fiduciary duty is to the company and shareholders, not personal ideological stances. This cautious behavior may start to shift, but pragmatism has largely dictated corporate public stances.
Process vs. Output in Work
Copied to clipboard!
(00:47:45)
  • Key Takeaway: The learning derived from the difficult process of creation (like writing research reports) is lost when AI instantly generates the final output.
  • Summary: Blodget notes that while AI can produce high-quality research reports quickly, the month-long process of writing one as a young analyst provided invaluable learning. He hopes that humans will use AI to automate tedious tasks, freeing up time to focus on work they are truly passionate about, similar to how chess remains popular despite computers being superior players.