StackAtlas · Weekly Briefing

AI News This Week

The most important AI stories from Mar 1, 2026 – Mar 7, 2026 — ranked by significance and public attention. Updated every Sunday at midnight EST.

Top Story

FundingFeb 27, 2026

OpenAI Raises Record $110B in Largest Private Funding Round Ever at $730B Valuation

OpenAI closed a historic $110 billion funding round on February 27, 2026, with $50 billion from Amazon, $30 billion from SoftBank, and $30 billion from Nvidia, valuing the company at $730 billion pre-money ($840 billion post-money). The deal includes a $100 billion expansion of OpenAI's AWS compute partnership and commits OpenAI to consuming at least 2GW of AWS Trainium compute. This marks the largest private funding round in history, surpassing OpenAI's own $40 billion March 2025 round, and signals massive corporate confidence in AI infrastructure despite growing questions about profitability and sustainability.

Read more:TechCrunch ↗Crunchbase News ↗

Policy & SafetyFeb 27, 2026

Anthropic Designated 'Supply Chain Risk' After Refusing Pentagon Restrictions on AI Use

Defense Secretary Pete Hegseth designated Anthropic a "Supply Chain Risk to National Security" on February 27 after the company refused to allow unrestricted military use of its Claude AI models without prohibitions on mass surveillance and autonomous weapons. The unprecedented move—the first time an American AI company has received this designation—followed a collapse in contract negotiations and triggered a six-month off-boarding period. Despite the blacklist, reports confirm Claude is still being used in Operation Epic Fury, the U.S. war against Iran, highlighting the Pentagon's reliance on the technology. The controversy sparked public support for Anthropic, pushing Claude to the top of app store charts as users switched from ChatGPT.

FundingFeb 27, 2026

Anthropic Raises $30B in Third-Largest Venture Round Ever Amid Pentagon Controversy

Anthropic announced a $30 billion funding round in late February 2026, valuing the company at $380 billion post-money, making it the third-largest venture round on record. The timing coincided with the company's high-profile dispute with the Pentagon over AI ethics and military use restrictions. The massive capital injection reflects continued investor confidence in Anthropic's approach to AI safety and responsible development, even as the company faces backlash from defense hawks. Anthropic's Claude models have emerged as serious competitors to OpenAI, with many developers and businesses citing superior performance on complex reasoning tasks.

Read more:Crunchbase News ↗TechCrunch ↗

Model ReleaseMar 3, 2026

OpenAI Releases GPT-5.3 Instant and GPT-5.4 Thinking with Major Finance Focus

OpenAI released GPT-5.3 Instant on March 3, 2026, focused on improved conversational flow and web search accuracy, followed by GPT-5.4 Thinking optimized for finance workflows. The company announced ChatGPT for Excel in beta, powered by GPT-5.4, which integrates directly into spreadsheets to build financial models and analyze data. Performance on OpenAI's internal investment banking benchmark improved from 43.7% with GPT-5 to 87.3% with GPT-5.4 Thinking. The releases include new integrations with FactSet, Dow Jones Factiva, LSEG, Daloopa, and S&P Global, targeting professional finance users and marking a strategic push into enterprise vertical markets.

Product LaunchMar 3, 2026

Apple Unveils MacBook Air with M5 Chip and Major AI Performance Upgrades

Apple announced the new MacBook Air with M5 chip on March 3, 2026, delivering up to 4x faster AI performance than the M4 model and featuring double the base storage at 512GB. The M5 includes a 10-core CPU, enhanced GPU with Neural Accelerators in each core, and Apple's N1 wireless chip for Wi-Fi 7 and Bluetooth 6 connectivity. Available in 13- and 15-inch models starting at $1,099, the new MacBook Air began pre-orders March 4 with availability March 11. Apple positioned the release as part of a broader AI push, emphasizing on-device AI capabilities through Apple Intelligence integration.

Read more:Apple Newsroom ↗MacRumors ↗

Product LaunchMar 3, 2026

Meta Tests AI Shopping Assistant to Compete with ChatGPT and Google Gemini

Meta began testing an AI-powered shopping research feature within its Meta AI chatbot on March 3, 2026, rolling out to select U.S. desktop users. The tool displays product carousels with images, pricing, and merchant links in response to shopping queries, competing directly with similar features from OpenAI's ChatGPT and Google's Gemini. Unlike competitors, Meta's tool doesn't include in-app checkout, requiring users to click through to merchant websites. CEO Mark Zuckerberg previously told investors the company plans to launch "agentic shopping tools" to help users discover products from Meta's business catalog, signaling a major push into AI-driven e-commerce.

Netflix Acquires Ben Affleck's InterPositive AI Filmmaking Startup

Netflix announced the acquisition of InterPositive on March 5, 2026, an AI filmmaking tools company founded in 2022 by actor-director Ben Affleck. The startup, which operated in stealth mode with 16 employees, develops AI-powered tools that help filmmakers with post-production tasks like relighting shots, background replacement, and visual effects without text-to-video generation. InterPositive's proprietary models are trained on controlled soundstage footage rather than scraped internet data, focusing on technical filmmaking assistance rather than replacing human creativity. Affleck joins Netflix as senior advisor on technology, and the deal reinforces Netflix's strategy of using AI to enhance quality rather than cut costs, though it has raised concerns among Hollywood unions.

Read more:Variety ↗NPR ↗Deadline ↗

Policy & SafetyFeb 3, 2026

International AI Safety Report 2026 Warns Capabilities Outpace Safety Measures

The second International AI Safety Report, published February 3, 2026, concluded that AI capabilities are advancing faster than safety measures can keep pace, with over 100 experts from 30+ countries contributing to the assessment. Led by Turing Award winner Yoshua Bengio, the report found that AI systems can assist in cyberattacks, biological threats, and disinformation campaigns, though they're not yet executing attacks autonomously. The report noted that criminal groups and state actors are actively using AI in operations, and that "open-weight models cannot be recalled once released," creating persistent security risks. The findings are informing policy discussions globally, including at the India AI Impact Summit.

RegulationDec 11, 2025

Trump Executive Order Targets State AI Laws, Creating Federal-State Clash

President Trump signed an executive order on December 11, 2025, establishing a national AI policy framework designed to preempt state regulations deemed burdensome to industry competitiveness. The order directs the Attorney General to establish an AI Litigation Task Force to challenge state laws on constitutional grounds and requires federal agencies to identify "onerous" state AI regulations by March 11, 2026. The move targets laws like California's Transparency in Frontier AI Act and Colorado's AI Anti-Discrimination Law, creating a potential showdown between federal deregulation efforts and state-level AI governance. Legal experts note the order lacks congressional authority to overturn state laws, setting up court battles throughout 2026.

#10

AI AgentsMar 1, 2026

AI Agents Move from Demos to Production as Industry Shifts to Pragmatism

March 2026 marks a turning point as AI agent systems transition from prototypes to real-world deployments, with Anthropic's Model Context Protocol (MCP) emerging as an industry standard. OpenAI and Microsoft have publicly embraced MCP, and Anthropic donated it to the Linux Foundation's new Agentic AI Foundation to standardize open-source agent tools. Major releases include Basis reaching $1.15 billion valuation with its agentic accounting platform, and multiple "AI agent for X" startups securing funding. However, industry analysts note that reliability and trust remain critical unsolved problems, with agents still complementing rather than replacing human workers despite improved capabilities in software engineering and workflow automation.