Trump Cancels AI Executive Order, Microsoft Drops Claude Over Cost, White House Funds $9B Spy AI — AI News Briefing

🗞️ AI News Briefing — May 23, 2026 (18:00 CST)

Top 7 Stories

1. Trump Abruptly Cancels AI Executive Order Hours Before Planned Signing

In a dramatic reversal, President Trump cancelled the signing of an executive order that would have established federal oversight of AI models — halting the ceremony just hours before it was scheduled to begin. According to AP News and The New York Times, Trump called off the order over concerns it could weaken America’s technological edge against China and impose excessive regulatory burdens on the AI industry. Politico reported that Trump told advisors he “didn’t like certain aspects” of the proposed order, which would have granted federal agencies authority to vet AI models before public release.

The cancelled order had been in development for months and represented the most significant AI regulatory framework proposed by the current administration. Its abrupt cancellation marks a sharp pivot from the administration’s earlier signaling that some form of AI oversight was necessary. USA Today reported that Trump cited concerns about overregulation, while The Washington Post noted that industry lobbyists had mounted an aggressive last-minute campaign against the order. The decision effectively leaves the United States without any federal framework for pre-deployment AI safety testing — a gap that European regulators have already filled through the EU AI Act.

The cancellation has significant implications for the global AI governance landscape. With the US stepping back from model-level oversight, China and the EU remain the only major powers actively regulating frontier AI development. For companies like OpenAI, Anthropic, and Google, the decision removes a potential compliance hurdle but also eliminates the possibility of a unified US regulatory standard that could have provided clarity for investment and deployment decisions. The Wall Street AI rally continued on the news, with the S&P 500 approaching its eighth consecutive winning week, powered by optimism that the AI sector will face lighter regulatory headwinds.

2. White House Approves $9 Billion for Intelligence Community AI Capabilities

The New York Times reported that the White House has approved $9 billion in funding for US intelligence agencies to accelerate their artificial intelligence capabilities — one of the largest single investments in government AI to date. The funding package is designed to help the intelligence community catch up with the rapid pace of AI advancement in the private sector, where companies like OpenAI and Anthropic are deploying models with capabilities that far exceed current government systems.

The investment will fund AI-powered signals intelligence, automated threat analysis, language translation, and cyber defense capabilities across agencies including the CIA, NSA, and the National Reconnaissance Office. Intelligence officials have long warned that the US is falling behind adversarial nations — particularly China — in applying AI to espionage and defense applications. The funding package is expected to support partnerships with private AI laboratories, potentially creating new procurement channels for Anthropic’s Claude and OpenAI’s models within classified environments.

The scale of the investment — $9 billion in a single package — signals that national security AI is transitioning from experimental programs to core infrastructure. For context, the entire federal AI R&D budget across all agencies was approximately $2.5 billion in fiscal year 2024. This new package more than triples that figure in a single appropriation. The decision also creates a new revenue channel for frontier AI labs: classified government contracts that, while not publicly disclosed, could represent billions in annual recurring revenue for companies that win the intelligence community’s trust.

3. Microsoft Cancels Claude Licenses as AI Costs Exceed Human Labor — Fortune Exposes the Real Numbers

Fortune published a significant report revealing that Microsoft’s internal data shows using AI technology — specifically Anthropic’s Claude — is now more expensive than paying human employees for many tasks. India Today reported separately that Microsoft is actively cancelling internal Claude licenses across several departments due to escalating costs. The revelation is the first major public acknowledgment from a frontier AI customer that the unit economics of enterprise AI deployment may not yet pencil out at scale.

Microsoft’s experience challenges the narrative that AI is an inevitable cost-saver for enterprises. While Claude’s capabilities have driven extraordinary revenue growth for Anthropic — contributing to the company’s $10.9 billion quarterly revenue and first-ever operating profit — the cost per seat for enterprise Claude deployments has reportedly risen sharply. For knowledge workers performing document analysis, code review, and content generation, Microsoft found that the combination of license fees, compute costs, and required human oversight made AI-augmented workflows more expensive than traditional human-only processes for many use cases.

The implications extend far beyond Microsoft. If one of the world’s largest enterprise software companies — and Anthropic’s strategic partner — is rolling back Claude adoption due to cost, it raises fundamental questions about the sustainability of AI lab revenue growth models. Anthropic’s 80x year-over-year growth has been driven by enterprise deployments, but Microsoft’s experience suggests that retention may be harder than acquisition. The broader market implication is that AI companies may face a reckoning on pricing: either reduce costs through model efficiency gains and infrastructure optimization, or risk losing enterprise customers who discover that the total cost of AI adoption exceeds the savings. For now, Microsoft’s decision represents a rare data point in an industry that rarely shares honest assessments of AI return on investment.

4. Anthropic Acquires Developer Tools Startup Used by OpenAI, Google, and Cloudflare

TechCrunch reported that Anthropic has acquired a developer tools startup whose products were widely used across the AI industry — including by competitors OpenAI and Google, as well as enterprise customers like Cloudflare. The acquisition represents Anthropic’s first major move into the developer tooling ecosystem and signals an ambition to own more of the AI development stack beyond just model APIs.

The acquired startup’s products — which reportedly include API management, model evaluation frameworks, and developer workflow integrations — were considered essential infrastructure for teams building on top of frontier models. By acquiring this capability in-house, Anthropic gains direct control over tools that influence how developers interact with AI models. This is strategically significant: developer tooling creates sticky, recurring revenue streams and establishes platform lock-in effects similar to what AWS built with its developer services ecosystem.

The acquisition also raises questions about how Anthropic will manage the startup’s existing relationships with competitors. If OpenAI and Google were customers of the acquired tools, Anthropic will need to decide whether to continue supporting them, potentially providing competitive intelligence, or restrict access to its own ecosystem. The move mirrors a broader pattern in the AI industry where companies are acquiring capabilities across the stack to build integrated platforms — Google with its Gemini Omni and Universal Cart, OpenAI with its autonomous reasoning capabilities, and now Anthropic with developer infrastructure.

5. Meta Shares Slide as Massive AI Spending Spooks Investors; Employees Report Growing Misery

BBC reported that Meta’s shares declined sharply after the company announced plans to spend billions more on AI infrastructure, with investors growing increasingly concerned about the capital intensity of Meta’s AI ambitions. The spending plans come at a time when Wall Street is beginning to scrutinize AI investments more closely — the same week that OpenAI filed its IPO prospectus and Anthropic disclosed its profitability, creating a comparative framework for evaluating AI spending efficiency.

The New York Times separately published a report detailing growing employee dissatisfaction at Meta related to the company’s aggressive AI pivot. According to the Times, employees have expressed frustration with the pace of AI-driven organizational changes, including layoffs, restructuring, and the increasing expectation that all products integrate AI features. TechTarget’s weekly tech roundup also noted that Meta has been conducting layoffs in parallel with its AI investments — a pattern that has become common across the industry but continues to create internal tension.

The dual pressure from investors and employees puts Meta in a difficult position. On one hand, CEO Mark Zuckerberg has staked the company’s future on AI leadership, arguing that massive investment in compute and model development is essential to compete with OpenAI, Anthropic, and Google. On the other hand, the market is beginning to demand evidence that AI spending translates to revenue growth — and Meta’s ad-dominated business model may not benefit from AI in the same direct ways that Anthropic’s Claude Code or OpenAI’s API services do. The stock decline suggests investors are not yet convinced.

6. California Governor Newsom Signs First-of-its-kind AI Worker Protection Executive Order

Governor Gavin Newsom signed a first-in-the-nation executive order on May 21, 2026, directing California to prepare workers, small businesses, and communities for the economic disruption that artificial intelligence will bring to the workforce. The order mobilizes state agencies, labor experts, economists, universities, and industry leaders to develop policies, gather data, and identify early warning signs of workforce disruption — while ensuring workers share in the gains created by AI-driven productivity.

The executive order directs the state to explore an ambitious range of policy interventions including severance standards, employment insurance and transition support for displaced workers, worker ownership models, universal basic capital concepts, expanded workforce training, and stronger tracking of hiring and payroll trends to help California respond faster to potential layoffs and economic disruption. “California has never sat back and watched as the future happened to us — and we won’t start now,” Newsom said. “We have taken the lead on advancing innovation, safety, and transparency. But we must think bigger.”

The order also addresses AI-driven layoffs within California state government itself. CalMatters reported that the order directs state agencies to find ways to ease the pain for workers displaced by AI automation, creating a framework that could become a model for other states. The timing is notable: the order came just days after reports of AI-driven layoffs across multiple sectors, including Meta, and as CBS News reported that AI job cuts are rising nationwide. California’s approach — combining worker protection, data-driven early warning systems, and exploration of structural reforms like universal basic capital — represents the most comprehensive state-level response to AI workforce disruption in the United States.

7. Japan Government to Access Anthropic’s Latest AI; EY and Microsoft Launch Global Enterprise AI Initiative

Japan’s finance minister confirmed that the Japanese government will gain access to Anthropic’s latest AI models, marking a significant expansion of Claude’s international government deployments. Japan Today reported that the agreement will give Japanese ministries access to Claude’s analytical and language capabilities, potentially across financial regulation, tax policy, and economic planning workflows. The deal underscores the growing trend of sovereign AI adoption, where national governments partner with frontier AI labs to modernize public services.

In parallel, Microsoft and EY announced a global initiative to help enterprise clients scale AI beyond the experimentation phase. The partnership combines EY’s consulting expertise with Microsoft’s Azure AI platform to help organizations deploy AI across audit, tax, advisory, and operational workflows. This follows the pattern established this week by PwC’s deployment of Claude to hundreds of thousands of employees and JPMorgan’s launch of its AI agent platform — signaling that the Big Four accounting firms and major financial institutions are moving AI from pilot programs to core infrastructure.

Together, these developments paint a picture of an AI industry where the center of gravity is shifting from consumer-facing products to enterprise and government deployments. Japan’s government adoption of Anthropic’s technology and the EY-Microsoft enterprise initiative represent the kind of large-scale, long-term contracts that generate predictable revenue — the foundation on which Anthropic built its path to profitability and the model that OpenAI will need to validate as it prepares for public markets.

📊 Trend Watch

🔭 What to Watch

• Federal AI Policy Vacuum — With Trump’s cancelled executive order, the US lacks a federal framework for AI model oversight. Watch for Congressional action or state-level legislation (following California’s lead) that could fill the regulatory gap — and whether European AI Act compliance requirements become the de facto global standard for US companies.
• AI Cost Curve Inflection Point — Microsoft’s Claude cost reversal is the first public signal that AI economics may not be improving fast enough. Watch for other enterprises to publish honest ROI data, and for AI labs to respond with pricing changes, efficiency improvements, or new product tiers.
• Meta’s AI Spending Reckoning — The stock slide over AI spending suggests investor patience is finite. Watch for Meta’s next earnings report to see whether AI investments are translating to measurable revenue growth or whether the company faces pressure to moderate its AI capex plans.
• Intelligence Community AI Procurement — The $9B funding package will create new classified procurement channels. Watch for which AI labs win intelligence community contracts — these deals could represent billions in revenue but will also raise questions about AI capabilities being used for surveillance and cyber operations.