Macro investor and Real Vision co-founder Raoul Pal took to social media this week to deliver a sharp public verdict on Claude, calling Anthropic's flagship AI assistant "utterly unusable" and warning the company faces a serious competitive risk if it cannot scale its inference infrastructure quickly. The comments, which gained rapid traction online, reflect a growing frustration among power users who say Claude's performance has been degraded by capacity constraints.
What Pal Said and Why It Matters
Pal's criticism centered on availability and reliability rather than the underlying quality of the model. He argued that even a technically superior AI product loses its value proposition entirely if users cannot access it consistently. For Anthropic, the timing is uncomfortable. The company is navigating an expensive period of infrastructure investment while fending off aggressive competition from OpenAI, Google, and a growing field of open-source alternatives. Pal explicitly warned that slow action on inference capacity could cause Anthropic to "lose" ground it has worked hard to gain.
Key Facts
- Raoul Pal publicly called Claude "utterly unusable" on social media
- His criticism focused on inference capacity and availability, not model quality
- Pal warned Anthropic must scale infrastructure "fast" or risk competitive losses
- The comments amplified existing user complaints about Claude's reliability during peak demand
- Anthropic is currently spending heavily on infrastructure amid rapid user growth
The complaint is not entirely new. Users across developer forums and social platforms have periodically reported degraded response times and rate-limiting errors when demand spikes. What gives Pal's comments extra weight is his profile as a prominent voice in macro finance and technology investment circles. When an investor of his visibility calls a product unusable, it tends to surface quickly in the feeds of other investors and potential enterprise clients. This is the kind of reputational friction that compounds quietly until it becomes a business problem.
"Claude is utterly unusable. Anthropic needs more inference fast or they're going to lose this."Raoul Pal, via social media
A Recurring Pressure Point for Anthropic
Capacity crunches are a structural challenge across the AI industry right now, but Anthropic's situation draws particular scrutiny given its valuation and the expectations attached to it. The company has attracted billions in investment on the promise that it can compete at the frontier, and questions about whether it can keep pace with demand have surfaced before. Some analysts have pointed out that the cost of scaling AI infrastructure could slow Anthropic's growth trajectory, particularly if user experience suffers during periods of high demand and drives subscribers toward alternatives.
Pal's frustration also echoes a concern that has been voiced by other high-profile observers. Venture capitalist Chamath Palihapitiya, for instance, has raised pointed questions about whether Anthropic could end up on the wrong side of AI industry consolidation, comparing the risk to tech companies that led early but faded as the market matured. Both critiques share a common thread: the gap between building a technically impressive model and sustaining a durable, scalable product for real-world use.
For users who rely on Claude daily, the core issue is simple. A model that produces strong outputs when it works is still a liability if it goes down or slows to a crawl at critical moments. Anthropic has not publicly addressed Pal's specific comments, though the company has previously acknowledged that managing demand against infrastructure availability is an ongoing operational priority. How quickly it can close that gap will likely do more to determine its competitive position over the next year than any single model release. The pressure to deliver on both fronts, quality and consistent access, is only going to intensify as enterprise adoption grows and rivals continue to invest heavily in their own capacity.
