Anthropic’s Cat Wu says that, in the future, AI will anticipate your needs before you know what they are

Anthropic is reportedly preparing a funding round that could value the company at about $950 billion, above OpenAI’s $854 billion March valuation, while business customers are increasingly preferring Claude over ChatGPT. A recent report said Anthropic has outpaced OpenAI among business users and quadrupled its market share since May 2025. The interview also highlights product momentum in Claude Code, coworker-style agents, and upcoming proactive automations, reinforcing a strong innovation trajectory.

Analysis

The near-term winner is less “AI in general” than the vendors that become the default operating layer for enterprise workflow automation. If the market is shifting from chat to proactive agent orchestration, the monetization pool moves from one-off inference to sticky seat expansion, workflow attach, and higher usage intensity — a better backdrop for hyperscalers and collaboration stacks than for pure model branding. The second-order effect is that enterprise buyers will increasingly compare AI on reliability, integration depth, and administrative control, not raw benchmark scores, which should advantage incumbents with distribution and security trust.

That said, this is also a classic spend-consolidation setup: once enterprises standardize on one or two agent frameworks, smaller point solutions and “wrapper” startups get squeezed. The biggest competitive risk is that model quality converges faster than product differentiation, compressing pricing power for standalone frontier labs while widening the gap for firms that own the workflow surface area. Over the next 3-6 months, watch for evidence that agentic features drive incremental net retention rather than merely shifting usage from one vendor to another; if it’s just substitution, the revenue lift will lag the hype.

The contrarian read is that the market may be underpricing how slowly “proactive AI” will scale in regulated environments. Autonomous setup of automations sounds compelling, but security, permissioning, and error tolerance will force most deployments into narrow, auditable use cases first. That suggests the first earnings upside accrues to infrastructure and security vendors, while pure application leaders could see slower-than-expected enterprise rollout despite strong sentiment.

Catalyst-wise, the next inflection is not a new model announcement but evidence of workflow lock-in: agent creation, task completion rates, and seat expansion in corporate accounts over the next 1-2 quarters. If those metrics inflect, the trade should favor the platforms embedded in daily enterprise work; if they do not, the current AI multiple expansion is vulnerable to a 10-15% reset as investors realize the adoption curve is still pilot-heavy.

Key Decisions for Investors

• Long MSFT vs. short equal-weight AI application basket over the next 3-6 months: Microsoft is best positioned to monetize proactive agents through distribution, identity, and workflow integration; target a 1.5-2.0x relative return if enterprise adoption broadens.
• Long AMZN on any 5%+ pullback, 2-4 month horizon: AWS should benefit from higher inference and orchestration load as agents become more autonomous; prefer entry after a weak tape because the upside comes from consumption acceleration rather than multiple expansion alone.
• Long CRWD as a portfolio hedge against agent proliferation, 6-12 month horizon: more autonomous enterprise automation increases security and permissioning spend; risk/reward is favorable because security budgets tend to expand even if AI application spend normalizes.
• Avoid chasing single-name frontier lab exposure near-term; if you want exposure, use call spreads rather than outright longs, since valuation sensitivity is high if enterprise monetization evidence does not materialize in the next 1-2 quarters.
• Pair trade: long infrastructure/security beneficiaries (MSFT, AMZN, CRWD) vs. short AI-app pure plays with weaker distribution, on the thesis that workflow ownership matters more than model quality as proactive agents roll out.