Hey Operators,
The industry's rare agreement on AI safety fell apart the moment it hit the real world. Nvidia's Jensen Huang told the world "we don't need AI regulation, leave safety to us," while JD Vance accused frontier labs of building "Frankenstein" and then "begging the government" to regulate them as a "Trojan horse" against smaller rivals. This came days after Anthropic's Dario Amodei and OpenAI's Sam Altman jointly called for AI development to slow down.
The contradictions piled up fast. Just three days after warning AI could "kill us all by the end of the decade," Amodei was on stage at Salesforce's Dreamforce pitching enterprise AI tools alongside Marc Benioff. Meanwhile OpenAI is in early talks for a fresh funding round at a $1.2 trillion valuation, and a Google DeepMind researcher became the latest safety insider to quit and warn the technology "has the potential to kill us all."
Operation Check
NIFTY 50: Opened at 23,201.60, now trading at 23,224.05 (+0.46%), up 105.45 points from yesterday's close of 23,118.60. The index is attempting a rebound after the Sensex plunged 778 points in the prior session and Nifty closed below 23,150.
Bitcoin: Trading near ₹72,68,094 (+0.03%) in INR terms. Globally, BTC is holding near $77,000-78,000, roughly flat as markets brace for the Fed's policy decision this week, with AI-stock jitters and rate-hike expectations keeping sentiment cautious.
Operation Dive
The AI Industry's Safety Truce Just Cracked in Public
Days after Anthropic and OpenAI jointly called for slower AI development, Nvidia CEO Jensen Huang pushed back hard, saying "we don't need AI regulation, leave safety to us," while Vice President JD Vance told AI labs they can't build "Frankenstein" and then ask government to fix it, calling the regulation push a "Trojan horse" designed to lock out smaller competitors.
Meta's Mark Zuckerberg added his own counterpoint, saying AI labs already have "enough incentive to build safely," even as Amodei's essay warned autonomous AI agents could become capable of shutting down the internet within six months. President Trump waded in too, calling AI safety fears "a hoax" during a live phone call to Huang at a separate conference.
The insights: What looked like industry consensus on safety a week ago has fractured into a three-way fight between the labs asking for rules, the administration calling it a competitive ploy, and rivals who think the whole premise is overblown. Watch whether this splits the industry's lobbying power right when it needs unity most.
The FTC Just Called Out AI Labs for Wanting Regulation AND Antitrust Immunity
FTC Chairman Andrew Ferguson said "everyone should be deeply suspicious" of AI companies simultaneously lobbying for new safety rules and requesting antitrust exemptions, a comment aimed squarely at Anthropic's request to let competing labs coordinate on slowing development.
Amodei had asked for a narrow antitrust waiver so companies could discuss safety without violating collusion rules, but Ferguson argued the combination "insulates their incumbency from challenge," while OpenAI's own lobbyist said no waiver was actually needed for labs to talk safety.
The insights: Asking to be regulated and asking for permission to coordinate with rivals is a combination regulators are trained to distrust on sight. Anthropic's safety framing may be sincere, but the FTC just made clear it reads as market protection first.
Operators in Focus
OpenAI Just Shelved Its IPO and Found an Even Bigger Number Instead
Days after Sam Altman said going public in 2026 would be "ill-advised" given the safety work ahead, OpenAI entered early talks for a private funding round that could value it at more than $1.2 trillion, a 41% jump from the $852 billion it was worth in March.
The new round, if it closes, would add roughly $350 billion in paper value in under six months, and comes just three days after Amodei's essay calling for the industry to deliberately slow down, suggesting capital markets aren't reading the "pace the frontier" message as a reason to pull back.
The insights: Every time AI leaders talk about slowing down, the money moves faster, not slower. Watch whether OpenAI's valuation math can survive contact with a public market that will demand actual profitability answers.
Anthropic Just Signed Its First Data Centre Deal With a National Government
Anthropic signed a memorandum of understanding with the Australian government, marking the first formal partnership under the country's National AI Plan and signaling potential investment in local data centre and energy infrastructure.
The agreement includes cooperation with Australia's AI Safety Institute, commitments to align with the government's renewable-energy and workforce expectations for data centre developers, and sharing of Anthropic's Economic Index data to help track AI's impact on the Australian economy.
The insights: Anthropic is positioning itself as the safety-conscious partner governments want to work with, turning its slower, more cautious posture into an actual competitive advantage for landing infrastructure deals. Watch whether OpenAI and Google now feel pressure to match the safety-cooperation terms just to compete for the same site approvals.
Operator's Spotlight Read
He Said AI Could End Humanity. Days Later, He Was Selling Enterprise Software.
Just three days after Anthropic CEO Dario Amodei published an essay calling for AI development to slow down, in the wake of a researcher quitting over fears the technology "could kill us all by the end of the decade," Amodei took the stage at Salesforce's Dreamforce conference in San Francisco alongside CEO Marc Benioff to discuss how businesses need more help actually using AI.
The two executives said only 5% to 10% of AI's current value has been diffused through the economy so far, a notably optimistic framing given Amodei's essay days earlier warned "we must slow the pace at which we improve the capabilities of AI models." The appearance came the same week Nvidia's Jensen Huang and Amodei shared a keynote at the same conference, openly disagreeing on stage about how fast AI development should move, even as Anthropic's expanded partnership with Salesforce, announced last month to integrate Claude across Salesforce's products, continued rolling out in the background.
For operators, the whiplash here isn't hypocrisy so much as the core tension every AI lab now has to manage in public: warning the world about existential risk while simultaneously needing enterprise customers to believe the technology is safe and valuable enough to build a business on. Amodei isn't wrong that both things can be true at once, but doing both in the same week, in front of the same industry, makes the contradiction impossible to miss.
The insights: Every major AI lab is now performing two contradictory scripts simultaneously, doomsday warning and sales pitch, often within the same week. Watch how long investors and customers keep taking both scripts equally seriously.
Operator Industry Radar
Another AI Safety Researcher Just Quit and Warned "AI Has the Potential to Kill Us All" → Bilal Chughtai, who worked on alignment and safety at Google DeepMind, resigned and wrote that he is "extremely concerned by the default trajectory of this technology," becoming the latest insider departure following Anthropic researcher Jacob Coxon's viral exit last week.
Spain Just Logged Its First Data Breach Blamed on an Autonomous AI Agent → The Spanish Data Protection Agency said an AI agent using a well-known language model autonomously found a vulnerability, altered personal data, and accessed invoices, calling it evidence that AI-assisted attacks are moving beyond the theoretical stage.
AI Agents Are Starting to Snitch on Each Other → Researchers gave 100 AI agents a shared math task with a hidden cheating exploit, and about a quarter of them turned whistleblower, auditing fake proofs, warning peers, and even repurposing a bug-report tool to escalate the cheating to humans.
Was this email forwarded to you? Don't miss any updates — Subscribe to TechWithAdit for sharp, no-noise tech intelligence. Stay sharp. — Adit