AI Pulse · Frontier Labs Watch

Two of the biggest names in AI just filed confidentially for public offerings, and the numbers floating around tell very different stories about who’s actually making money.

Anthropic is reportedly tracking toward a late-2026 debut off a $965 billion private valuation. According to aggregated reports, the company posted roughly $10.9 billion in Q2 revenue with about $559 million in operating income. That would mark its first profitable quarter. OpenAI, meanwhile, is aiming for a valuation around $1 trillion but is still bleeding cash. Projected 2026 losses sit near $14 billion. Both filings are confidential, so these figures come from aggregators, not the S-1 documents themselves. Treat them as directional until the actual paperwork goes public.

What this means for you

If Anthropic really turned a profit, it’s the first major signal that foundation model companies can operate in the black without infinite venture backing. That matters because it suggests the underlying economics of large language models might actually work at scale. For business owners, it’s a hint that the tools you’re paying for today, whether Claude or GPT, are likely to stick around and keep improving rather than vanish in a funding crunch.

OpenAI’s continued losses, even at massive revenue, show how expensive it is to stay ahead in this space. That cost pressure will shape pricing, feature rollout, and which companies survive the next few years. If you’re building AI workflows into your operations, this is the kind of market intelligence that helps you pick stable partners and plan around realistic timelines. It’s also the sort of thing we wire into the Omni Command Centre so you can see which models are gaining traction and where the real momentum sits, not just the hype.