Anthropic Calls Global AI Pause, OpenAI-NVIDIA Deal Stalled, Google Unveils New TPU Chips — AI News Briefing

Top 7 Stories

1. Anthropic Urges Global Pause in AI Development Over Self-Improvement Risks

Anthropic has issued an extraordinary call for a global slowdown in AI development, warning that AI systems could soon reach a point where they improve themselves without human intervention. Co-founder Jack Clark warned that AI needs a “brake pedal,” arguing that the human role in AI development is rapidly narrowing.

The warning, covered by WSJ, Reuters, Fortune, Al Jazeera, and France 24, represents one of the most significant safety calls from within the AI industry itself. Anthropic argues that without a coordinated pause framework, AI systems could escape meaningful human control — a scenario that could become irreversible once self-improving systems cross a critical threshold.

Anthropic also expanded its Project Glasswing AI security program to protect critical global infrastructure, signaling a two-track approach: calling for broader industry restraint while simultaneously building defensive capabilities.

2. OpenAI-NVIDIA $100B Megadeal Reportedly on Ice

In a major development for AI infrastructure, The Wall Street Journal reports that the anticipated $100 billion megadeal between OpenAI and NVIDIA is currently stalled. The deal, which would have been one of the largest technology transactions in history, has reportedly hit roadblocks as both companies reassess their infrastructure spending strategies.

Meanwhile, NVIDIA CEO Jensen Huang has signaled that NVIDIA is pulling back from its deep ties with OpenAI and Anthropic, though his explanation has raised more questions than it answers. The development comes as AI infrastructure stocks face sell-off pressure over capex risk concerns.

Despite the slowdown, major firms including OpenAI, NVIDIA, and KPMG continue establishing AI centers and labs globally, with Singapore emerging as a key hub for AI infrastructure investment.

3. Trump Signs AI Safety Order Seeking Voluntary Model Reviews

President Trump has signed an executive order on AI safety that seeks voluntary reviews of new AI models before deployment. The order establishes a framework for the administration to engage with AI companies on safety assessments, though it stops short of mandatory pre-deployment testing requirements.

In a related move, Trump is set to meet with AI company leaders as soon as next week to discuss a government profit-sharing plan that could include the administration taking equity stakes in AI companies. The proposal has sparked debate about the appropriate role of government in private AI ventures.

Separately, bipartisan House lawmakers have released a draft bill that would prohibit states from creating their own AI regulations — a federal preemption effort that could reshape the U.S. AI regulatory landscape.

4. Google Unveils New TPU Chips to Challenge NVIDIA’s AI Hardware Dominance

Google has unveiled a new generation of TPU (Tensor Processing Unit) chips designed to directly challenge NVIDIA’s dominance in AI hardware. The new TPUs are positioned as a cost-effective alternative for large-scale AI training and inference workloads.

The announcement comes as US tech giants team up to counter NVIDIA’s chip lead, with companies increasingly seeking alternatives to NVIDIA’s GPUs amid supply constraints and pricing concerns. Google’s move to strengthen its in-house silicon capability signals intensifying competition in the AI accelerator market.

Meanwhile, NVIDIA’s Nemotron 3 Ultra has been recognized as America’s best open AI model, though analysts note that Chinese AI models continue to lead in certain benchmarks.

5. Meta Launches Paid AI Business Agent Across Messaging Apps

Meta has launched a paid AI business agent service across its messaging platforms, including WhatsApp, Messenger, and Instagram. The service allows businesses to deploy AI-powered agents that can handle customer interactions, sales inquiries, and support tasks directly within Meta’s messaging ecosystem.

The EU has warned Meta not to block rival AI bots from WhatsApp, raising antitrust concerns about the company’s growing AI agent platform. The development highlights the expanding regulatory scrutiny around AI agents, with experts noting that AI agents are now coming under antitrust and consumer protection review.

Separately, Frances Haugen has raised concerns about AI “friends” and the psychological impact of AI companions on users, adding another dimension to the debate around AI agent safety and ethics.

6. Snowflake and Anthropic Accelerate Enterprise AI Adoption

Snowflake and Anthropic have announced an expanded partnership aimed at accelerating enterprise AI adoption, driven by rising demand for governed AI solutions. The collaboration brings Anthropic’s Claude models deeper into Snowflake’s data cloud platform, enabling enterprises to build AI applications with stronger governance and compliance controls.

The partnership reflects a broader trend of AI companies focusing on enterprise-grade solutions with built-in safety and compliance features. As regulatory scrutiny intensifies, governed AI — AI systems with transparent data handling, auditability, and compliance safeguards — is becoming a key differentiator in the enterprise market.

7. Tech Industry Loses 123,000 Jobs to AI — Layoffs Accelerate

The technology industry has lost approximately 123,000 jobs this year, with AI cited as the most common reason for layoffs, according to Forbes. Major tech companies continue restructuring their workforces as AI automation replaces roles in engineering, customer support, and content creation.

Amazon has unveiled its latest warehouse robot as part of the ongoing push toward AI-driven logistics automation. The trend underscores a growing tension in the AI industry: while companies invest billions in AI infrastructure, those same investments are driving significant workforce displacement across the tech sector.

Trend Watch

What to Watch

• Anthropic’s pause proposal details: Expect further specifics on what a “global pause” framework would look like and whether other AI labs will endorse it.
• OpenAI-NVIDIA deal resolution: Whether the $100B deal is renegotiated, delayed further, or abandoned will signal the trajectory of AI infrastructure investment.
• Trump-AI company equity meeting: The upcoming meeting between Trump and AI company leaders on government profit-sharing could reshape how AI companies interact with the federal government.
• Federal vs. state AI regulation: The House draft bill preempting state AI laws will face pushback from states that have already enacted AI regulations, setting up a major federalism debate.
• AI agent regulation: With Meta’s AI agents under EU antitrust scrutiny and consumer protection reviews underway globally, expect regulatory frameworks for AI agents to accelerate.