Tech Wipeout: Google's $500 Billion Crash, Anthropic's Losses Exposed

Anthropic

Across the shows: 80 speakers in other episodes, 95 bullish and 61 bearish calls, past 30 days

• John Clark Levin · Kurzweil TechnologiesBullish

  Claude currently leads rivals on a plurality of tasks, though the three frontier labs remain close.

  Levin favors Claude because it better translates user intent into finished work with less handholding. Given the same prompt, Claude more often picks up what the user actually wants, even when it is not stated precisely, which matters for the broad economy. He qualifies this by saying Anthropic, OpenAI and Google DeepMind are within spitting distance and that Gemini beats the other two on a significant number of tasks, so no one is running away with the race.

  “That means Claude on at least a plurality of tasks compared to Gemini and ChatGPT will be able to accurately translate the user's intent into the work they want done without a whole lot of extra handholding.”at 26:01

• John Clark Levin · Kurzweil TechnologiesBearish

  Anthropic's pricing power is at risk as rivals match its capability more cheaply.

  Levin argues that because the top three labs are close and no one leads at all times, competition limits Anthropic's pricing power. If another company can deliver a similarly capable model at a lower cost, Anthropic may have to cut prices. That risk comes while the company is still meeting substantial computing obligations, which raises the stakes of a price cut.

  “So no one is running away with the race yet. That's important for pricing power. If another company can deliver a similarly capable model at a lower cost, then Anthropic may have to cut prices even while meeting substantial computing obligations.”at 26:56

AI bubble & tech valuations

• David Woo · David Woo Unbound

  Nasdaq's valuation assumes a winner-take-all AI outcome; commoditization is more likely.

  Wu argues the Nasdaq's valuation is priced for AI to produce a winner-take-all outcome, while he believes commoditization is the most likely result. He contrasts the US setup, where Nvidia and Jensen Huang have acted as a lender of last resort for troubled AI companies and cushion trouble with money, with China, which lacks that backstop. He concludes China will ultimately set the way for the US AI trade, and that the market's expectations leave little room for error.

  “The Nasdaq valuation right now assumes that AI is gonna produce a winner take all outcome. I think that the most likely outcome is commoditization.”at 4:13

• Steve Eisman · The Real Eisman Playbook

  The whole tech space will sell off if business-to-business AI sales start slowing.

  Asked what happens to semiconductors and the rest of tech if business-to-business AI sales slow, Eisman says the whole tech space sells off. Semiconductors already sold off in July. He dismisses the idea that Treasury yields or rate support would change that outcome. His reasoning is brief and conditional on the B2B slowdown.

  “I think the whole tech space sells off.”at 15:29

AI pricing & unit economics

• Steve Eisman · The Real Eisman Playbook

  The AI token spending binge has ended; buyers are now more price-sensitive.

  Eisman describes 'token maxing' around June and July, when customers spent on AI tokens with no sensitivity to cost. That ended, and people are now far more self-conscious about spending. From this he expects the third and fourth quarters to be more interesting for Anthropic and OpenAI than the first half of the year, and he hopes Anthropic's S-1 will give clarity on how the spending is trending.

  “Token maxing, which is where people were just spending whatever they... Whatever they had no sensitivity to spending. They just spend money like crazy on tokens. Token maxing ended, Now people are a lot more self conscious.”at 14:45

Enterprise AI adoption & AI software

• John Clark Levin · Kurzweil Technologies

  Demo-level AI capability outpaces reliability; markets won't tolerate that gap for three years.

  Levin argues AI capabilities will not plateau, but the gap between peak demo capabilities and reliability (tasks AI can do roughly as reliably as humans) is large. Market hype is driven by demos, but real-economy adoption depends on reliability. Reliability is improving fast but still trying to catch up to peak capabilities, so he doubts markets will tolerate that gap persisting for the next three years.

  “Although market speculation and hype are driven by the peak capabilities in the demos, adoption in the real economy is driven by reliability. And although reliability is increasing very fast, it's huffing and puffing trying to catch up to peak capabilities...”at 21:17