Meta faces US pressure for AI model review as capex surges
Meta Platforms Inc. is under pressure from the Trump administration to voluntarily submit its AI models for federal security review, a move competitors like OpenAI and Microsoft have already accepted. The company remains the sole major holdout, though it expects to finalize an agreement soon. Separately, Meta's capital expenditure guidance has risen to $135 billion, more than double the previous year's spend, contributing to a recent 3% decline in its stock price.

*this image is generated using AI for illustrative purposes only.
Meta Platforms Inc. shares are down approximately 3% over the past week as the company faces pressure from the Donald Trump administration to submit its advanced artificial intelligence models for federal voluntary review. The stock decline is compounded by investor concerns over a significantly increased capital expenditure budget, which has risen to $135 billion for the year. The administration's push aims to allow federal agencies to evaluate the capabilities and vulnerabilities of cutting-edge AI models before broader deployment, focusing on national security risks such as cyberattacks and exploitation by foreign adversaries.
Meta remains the only major U.S. AI developer that has not yet agreed to provide the government with voluntary access to its models for testing. Administration officials have been pressing the company through email communications to participate in the voluntary government review program, the New York Times reported on Tuesday, citing four people familiar with the confidential request. In a statement, Meta spokesperson Francis Brennan said that the company supports the administration’s efforts and expects to finalize an agreement soon. Commerce Department spokesperson Ben Kass stated that the department’s Center for AI Standards and Innovation routinely works with companies on voluntary agreements as part of its mandate.
Capital Expenditure Surge
Beyond regulatory pressure, the primary factor weighing on the stock is Meta's capital expenditure trajectory. The company plans to spend $135 billion this year, up from prior guidance of $115 billion and more than double last year's $72 billion in capex. This scale of spending continues to pressure near-term multiple expansion, even as the underlying business remains strong, and remains the primary reason the stock has struggled to reclaim its highs despite strong revenue growth.
Industry Participation and Executive Order
Several leading AI companies have already entered similar arrangements with the federal government. OpenAI and Anthropic have been working with U.S. officials to evaluate unreleased AI systems. Additionally, Alphabet Inc.’s Google DeepMind, Microsoft Corp and Elon Musk’s xAI agreed in May to provide early access to future models for national security assessments. President Trump signed an executive order on June 2 giving the government responsibility for AI reviews, allowing voluntary government reviews of certain frontier AI models up to 30 days before release.
| Company | Participation Status | Recent Action |
|---|---|---|
| Meta Platforms | Not yet agreed | Expects to finalize agreement soon |
| OpenAI | Participating | Working with U.S. officials on unreleased systems |
| Anthropic | Participating | Suspended access for foreign nationals |
| Google DeepMind | Participating | Agreed to early access in May |
| Microsoft Corp | Participating | Agreed to early access in May |
| xAI | Participating | Agreed to early access in May |
At the time of publication, Meta shares are trading 0.15% lower at $556.83.
How will the voluntary review process impact the release timeline for Meta's upcoming Llama 4 and other future AI models?
Will the $135 billion capital expenditure budget be sufficient to maintain Meta's competitive edge against rivals who have already secured government partnerships?
Could the delay in agreeing to federal reviews result in stricter regulatory mandates compared to the voluntary terms accepted by competitors?

































