Two stories out of Silicon Valley this week point to the same shift: the balance of goodwill and money in AI is moving around fast. The Information reports that Anthropic may be losing standing with the AI research community, even as a group of AGI House alumni just closed a $25 million fund to back the next wave of founders.
Here’s what stands out.
Anthropic’s goodwill problem
Anthropic built its early reputation on safety, research credibility, and a culture that researchers wanted to be part of. That reputation is part of how it recruited against OpenAI and Google DeepMind, and part of how it raised tens of billions.
Now The Information signals that some of that goodwill is slipping among researchers. That matters more than it might sound. In AI, talent is the product. The handful of people who can train frontier models move between a few labs, and they talk. When sentiment turns, recruiting gets harder, retention gets shakier, and rivals pounce.
Why might the mood be souring? A few forces are worth watching:
- Commercial pressure. Anthropic has leaned hard into enterprise sales and revenue growth. Researchers who joined a mission-first lab sometimes bristle when the priorities tilt toward shipping and selling.
- Compensation wars. Meta, OpenAI, and others have thrown enormous pay packages at top researchers. Any lab that can’t or won’t match starts to look less attractive.
- Openness tension. Labs that publish less, or restrict what their teams can share, tend to frustrate researchers who value being part of the broader scientific conversation.
I’d treat this as a signal, not a verdict. Anthropic is still one of the most sought-after places to work in the field. But goodwill is an asset, and assets can be spent down. If the trend The Information describes deepens, expect it to show up in departures and in who wins the next round of recruiting fights.
$25 million for AGI House alumni
The second piece of news is more upbeat. According to The Information, alumni connected to AGI House have raised a $25 million fund.
If you’re not familiar, AGI House is a well-known hacker house and community in the Bay Area where AI builders gather, demo projects, and meet investors. It’s become a kind of on-ramp for early AI startups. A fund coming out of that network is a logical next step: the people who’ve been watching projects form at the earliest stage now have capital to back them.
$25 million is small next to the multi-billion-dollar megafunds, but that’s the point. Early-stage AI doesn’t need huge checks. It needs people who can spot a strong team before the rest of the market notices. A community-born fund has an edge there, because the partners already know the founders, the projects, and who actually ships.
Why both stories matter together
Put side by side, these two items sketch the current state of AI’s talent economy:
- The big labs no longer have a lock on the best people. Goodwill is fragile, and even a leader like Anthropic has to keep earning it.
- New money is flowing to the edges. Community-driven funds are forming to catch builders who want to start something rather than join a giant.
For practitioners, the takeaway is practical. If you’re a researcher or engineer, you have more leverage and more options than you did a year ago. If you’re a founder, capital is increasingly available from people who understand AI from the inside, not just generalist VCs writing AI checks.
What to watch next: whether Anthropic moves to shore up its standing with researchers, and where this new fund places its first bets. Both will tell us a lot about who gains ground in the second half of 2026.
The Information has the full details on both stories.