The escalating standoff between the Trump administration and AI safety company Anthropic dominated tech and defense news this week, as federal contractors rush to sever ties with the firm while OpenAI moves to fill the vacuum with NATO and Pentagon partnerships. Meanwhile, Apple unveiled its next generation of silicon, and tensions in the Middle East entered a dangerous new phase.

AI INDUSTRY FACES NEW REALITY AS GOVERNMENT DRAWS HARD LINES

The Trump administration has banned federal use of Anthropic and designated the AI company a national security supply chain risk, a dramatic escalation following CEO Dario Amodei's refusal to remove safety guardrails around mass surveillance and autonomous weapons (Reuters, Politico). Defense contractors including Lockheed Martin are now purging Anthropic's tools from their operations to maintain access to the trillion-dollar defense budget, though legal experts question whether the government has authority to restrict private commercial activity (Defense News). Anthropic has announced plans to challenge the federal ban in court.

The administration's concerns were articulated by Under Secretary of Defense for Research and Engineering Emil Michael, who warned that commercial AI contracts with operational restrictions could jeopardize military missions—raising the specter of AI models shutting down mid-operation if terms of service are violated (Defense One).

OpenAI, by contrast, is positioning itself as the willing partner. The ChatGPT creator is reportedly in talks to deploy its technology on NATO's unclassified networks, following its recent Pentagon deal (Financial Times). CEO Sam Altman has sought to thread a needle, emphasizing the company's AI won't be used for domestic surveillance or by agencies like the NSA while apparently accepting broader military applications.

OPENAI TAKES AIM AT MICROSOFT'S GITHUB

In a move that could strain its relationship with primary investor Microsoft, OpenAI has begun developing its own code repository platform (The Information). The project, motivated by recent GitHub outages, remains months from completion but signals OpenAI's ambition to control more of the developer ecosystem. The company is considering making the service available to customers upon release.

Separately, TechCrunch reports that OpenAI has disbanded its mission alignment team—the internal group focused on ensuring AI systems remain aligned with human values—raising questions about the company's safety priorities as it pursues government contracts.

APPLE UNVEILS M5 SILICON AND PREMIUM DISPLAYS

Apple announced its M5 Pro and M5 Max chips, featuring an 18-core CPU built on a new "Fusion Architecture" that combines two 3nm dies into a single system-on-a-chip (Apple Newsroom). The processors power the newly revealed MacBook Pro models, with six "super" cores and 12 performance cores optimized for power-efficient multithreaded workloads. The company also launched the M5 MacBook Air.

The hardware refresh extends to displays: the $1,599 Studio Display and $3,299 Studio Display XDR both feature Thunderbolt 5 connectivity and upgraded cameras with Desk View support (The Verge). The XDR model offers a 120Hz mini-LED panel with 2,000 nits peak brightness and medical imaging features pending FDA clearance.

MIDDLE EAST CONFLICT ESCALATES INTO FIFTH DAY

Israel launched targeted strikes on security headquarters in Tehran while Iran retaliated with drone attacks on U.S. sites in Dubai and Qatar (Associated Press). The conflict has disrupted shipping in the Strait of Hormuz and triggered international political fallout, with President Trump criticizing UK Prime Minister Keir Starmer's response and threatening to terminate trade with Spain.

Four American soldiers have been killed in the conflict with Iran, according to Pentagon officials, as the administration warns of increasing casualties (NBC News). The escalation comes ahead of a planned late-March summit between Trump and Xi Jinping, with China calling for ceasefire while vowing to defend its sovereignty and "red lines" (Reuters). Negotiators are expected to meet in Paris next week to prepare for the presidential visit.

IN BRIEF

Tech and Immigration: A WIRED analysis reveals U.S. immigration agencies ICE and CBP have spent at least $515 million on technology from Microsoft, Amazon, Google, and Palantir in recent years, highlighting Big Tech's role in federal enforcement infrastructure.

Voice Rights in the AI Era: NPR's David Greene is suing Google, alleging the company's AI replicated his professional voice without authorization or compensation.

Domestic News: A Georgia father was convicted of second-degree murder for providing a firearm used by his son in a fatal school shooting (CNN). In Texas, Governor Greg Abbott won the GOP primary, while North Carolina's 4th congressional district race between Rep. Valerie Foushee and challenger Nida Allam remains too close to call and likely headed for a recount.

Startup Watch: AI inference startup Modal Labs is in talks at a $2.5 billion valuation, while Microsoft has warned of zero-day exploits targeting Windows and Office users (TechCrunch).