Sent to 20 readers
Race to AGI Daily Digest - Tuesday, July 7, 2026
Share:
TLDR
UN governance talks pivot from abstract AI risk to concrete safety and child‑protection goals.
Healthcare and pharma become testbeds for agentic operating systems, led by Assort Health and Katalyze AI.
Sherpa.ai raises $18M to make privacy‑preserving, data‑sovereign AI a core part of the enterprise security stack.
Bespoke Labs’ $40M round shows stress‑testing AI agents is becoming mandatory before enterprise rollout.
AI‑linked hardware stocks rebound, reinforcing the link between chip capacity and secure agent deployment.
The Full Story
Building on Monday’s UN kickoff in Geneva, the governance conversation just narrowed in on safety and kids. The updated dialogue is explicitly leaning into child protection and online harms, not just abstract “AI risk.” You can see that shift in the latest round of talks here: UN governance with safety push ->.
Now look at where the money is going today. On one side, you’ve got agentic operating systems for regulated worlds. Katalyze AI is raising to build an agentic OS for pharma, while Assort Health just pulled in $120M to run AI agents across real patient journeys. That’s not chatbots; that’s workflow plumbing inside hospitals. The Assort round is a good snapshot of where healthcare agents are headed: Assort Health funding details -> and their profile here ->.
On the other side of the stack, our active storyline kicks in: security. Sherpa.ai’s $18M round is all about privacy‑preserving, data‑sovereign AI, where models learn from your data without shipping it all to a central cloud. That’s basically a compliance and breach‑risk shield baked into the architecture: Sherpa.ai’s raise ->. Bespoke Labs is raising $40M to attack the next problem: how do you stress‑test enterprise AI agents before they go near customers or regulators? Think chaos engineering, but pointed at AI coworkers: Bespoke Labs funding ->.
The social chatter lines up: people are talking about AI as coworkers, not just coding helpers, and asking which capabilities are underrated. That’s exactly the world these agent platforms are aiming for.
Meanwhile, the market is re‑rating the hardware that feeds all this. Tesla, AMD, Qualcomm, TSMC, and Broadcom all ripped higher today, a sharp contrast to Monday’s sell‑off. It’s a reminder that the ongoing AI chip export reckoning -> is really about capacity for secure, high‑stakes agents. You can zoom in on AMD’s role in that stack here: AMD profile ->, and follow the funding flows across security and agents on our AI deals graph ->.
So the week’s story is getting clearer: regulators are tightening safety expectations at the top, while startups rush to harden the AI security stack and turn agents into trustworthy coworkers in some of the most sensitive industries we have.
Get This Delivered Daily
Join thousands of AI professionals who start their day with Race to AGI.