Now Playing
Anthropic's $750B Compute Bet, Pentagon Oversight Gaps & Military AI Walkouts
Audio is available on Spreaker — see link below.
Anthropic's pre-IPO numbers reveal an $80B paper loss and a staggering $750B compute spending plan — while the Pentagon's AI oversight office sits 70% understaffed. Today's briefing connects the capital, the contracts, and the conscience crisis reshaping the AI industry.
Audio is available on Spreaker — see link below.
Anthropic just told the world it lost eighty billion dollars in the first half of twenty twenty-six. That number is doing a lot of work right now, and not in Anthropic's favor.
Here's what's underneath that: Anthropic has signaled plans to spend up to seven hundred and fifty billion dollars acquiring computing capacity over the coming years. That's not a typo.
Days before the IPO announcement, Anthropic also committed five percent of its shares to a public benefit fund focused on AI-related job displacement. If the company's valuation climbs above certain thresholds, that figure rises by two percent for every fifty percent gain.
The military dimension of this story is where things get more complicated. A government audit released in mid-October documented what many suspected but hadn't seen formally confirmed.
That tension surfaced in a very direct way this week when Sofia Liguori, an engineer at Google DeepMind, resigned publicly over the company's U.S. and Israeli military contracts, specifically a deal called GenAI.mil that enables battlefield AI targeting. She confirmed others have already left over similar concerns.
One quieter development worth tracking: the Department of Health and Human Services extended its Anthropic Claude contract through October thirty-first as it searches for a broader enterprise AI procurement agreement. Federal agencies are moving away from transitional deals toward permanent infrastructure commitments.
Chapter summary auto-generated from the verified script. Listen to the full episode for the complete content.