AI News Roundup: US AI Giants Unite Against Chinese Model Theft, Anthropic Overtakes OpenAI in Revenue, NVIDIA Rides 10-Day Streak

OpenAI, Anthropic, and Google Unite Against Chinese AI Model Theft

In a rare show of solidarity, OpenAI, Anthropic, and Google are now sharing threat intelligence through the Frontier Model Forum to combat what they describe as systematic model distillation by Chinese AI labs. The collaboration mirrors how cybersecurity firms exchange threat data — when one company detects an adversarial pattern, it flags it for the others.

Anthropic published the most detailed accounting yet: three Chinese labs — DeepSeek, Moonshot AI, and MiniMax — created approximately 24,000 fraudulent accounts and ran more than 16 million exchanges with Claude to distill its capabilities. MiniMax alone accounted for 13 million of those exchanges. The goal was straightforward — use a powerful model’s outputs as training data for a cheaper, smaller model that replicates its reasoning, coding, and vision abilities without the R&D cost.

U.S. officials estimate that unauthorized model copying costs American AI companies billions annually. The collaboration adds urgency to the debate over whether API access to frontier models needs stricter controls, especially after DeepSeek’s R1 release in January 2025 wiped nearly $1 trillion off U.S. and European tech stocks in a single day by demonstrating near-parity at a fraction of the cost.

Anthropic Surpasses OpenAI in Revenue for the First Time

Anthropic announced that its annualized revenue run rate has crossed $30 billion, while OpenAI sits at $25 billion — marking the first time Anthropic has outearned its larger rival. The milestone is driven by explosive enterprise adoption: over 1,000 enterprise customers now spend more than $1 million per year on Claude, double the 500 reported during Anthropic’s Series G in February.

To fuel the next phase of growth, Anthropic also secured a new agreement with Google and Broadcom to deliver approximately 3.5 gigawatts of next-generation TPU capacity starting in 2027, on top of roughly 1 gigawatt of Google compute already committed for 2026. The infrastructure deal underscores just how capital-intensive the frontier AI race has become — and how Anthropic is betting that enterprise demand for Claude will continue to outpace supply.

NVIDIA Posts Longest Winning Streak Since 2023 on Rubin Demand

NVIDIA’s stock logged 10 consecutive winning sessions through April 14, rising 18% to approximately $192 and breaking out of a multi-month consolidation. The rally was fueled by surging demand for the company’s next-generation Vera Rubin AI platform, which CEO Jensen Huang says is part of more than $1 trillion in GPU orders through 2027.

The Rubin platform — built on 3nm process nodes with HBM4 memory — promises to reduce AI inference costs by 90% compared to the current Blackwell generation and cut training GPU requirements by 75%. Meta has already locked in early access through a $21 billion deal with CoreWeave for Rubin-based clusters to power its next generation of agentic AI models. Microsoft’s Azure is also building “Fairwater AI superfactories” around Rubin NVL72 systems.

India Launches AI Governance Body with Labor-Market Mandate

India’s government constituted the AI Governance and Economic Group (AIGEG), a high-level inter-ministerial body chaired by Union IT Minister Ashwini Vaishnaw. What sets AIGEG apart from the growing list of national AI committees is its explicit labor-market mandate: the body is tasked with assessing which job profiles AI adoption will hit first and developing mitigation and transition plans.

The move comes as India balances its position as both a major AI talent hub and a country with hundreds of millions of workers in roles vulnerable to automation. AIGEG will also coordinate AI governance policy across ministries, aiming to avoid the fragmented regulatory approaches that have slowed AI adoption in other large economies.

Gartner: Worldwide AI Spending to Hit $2.5 Trillion in 2026

Gartner’s latest forecast projects global AI spending will reach $2.52 trillion in 2026, a 44% increase year-over-year. AI infrastructure alone will add $401 billion, with AI-optimized server spending surging 49%. Data center spending is expected to cross $650 billion, up from nearly $500 billion last year.

But the headline number comes with a caveat: Gartner places AI firmly in the “Trough of Disillusionment” for 2026, meaning enterprises are more likely to buy AI through their existing software vendors than launch new moonshot projects. Organizations with successful AI initiatives invest up to four times more in data quality, governance, and change management than those reporting poor outcomes — a signal that the infrastructure gold rush alone won’t guarantee returns.

NVIDIA Launches Agent Toolkit with 17 Enterprise Partners

At GTC 2026, NVIDIA debuted its Agent Toolkit alongside 17 launch partners including Adobe, Salesforce, SAP, ServiceNow, Palantir, and Red Hat. The toolkit bundles NVIDIA’s OpenShell runtime and Nemotron models as open plumbing for building autonomous enterprise agents — AI systems that can execute complex, multi-step workflows across local and cloud environments without human intervention.