OpenAI's Q2 sales show tepid growth compared with Anthropic, according to WSJ

Private-company revenue prints of this kind arrive through press sourcing rather than filings, and episodes of this type have tended to trade on the listed proxies rather than the subject itself: Microsoft carries OpenAI exposure through its investment and its Azure inference and training workload, so the read-across runs through the AI-capex and cloud-growth narrative attached to that peer set. The distinction worth drawing is between growth rate and scale: a slower sequential rate at a smaller base versus faster growth at a larger one says more about competitive positioning in enterprise and API demand than about absolute trajectory, and markets have historically been quicker to reprice the proxy names on the direction of the gap than on either number in isolation. The deepened losses sit within an established pattern for the sector, where capital intensity and compute commitments have long outrun revenue, and have not on previous occasions derailed the listed sponsors unless funding terms or compute supply became the binding constraint. Attributed reporting without audited figures invites revision, so the follow-ons that matter are confirmation, any commentary from the proxy names on capacity and partnership economics, and whether enterprise spending surveys corroborate the share shift. As a single-sourced datapoint on private entities, the signal is directional and the market's established habit is to fade initial extrapolation absent corroboration.