Meta uses the US research tax credit for AI data centers and chips
Meta classifies AI data centers and Nvidia chips as research expenditures. According to the article, this saved the company 3.9 billion dollars in taxes in 2025, compared with 2 billion in 2024. In SEC filings, it acknowledges that these savings could be challenged.
Meta uses the US federal research tax credit by classifying AI data centers as “pilot models” and Nvidia chips as experimental materials. According to an article in The Decoder citing findings from New York Times, the tax savings reached 3.9 billion dollars in 2025, compared with 2 billion in 2024 and 700 million in 2023. According to the article, this made Meta the largest recipient of this credit among publicly traded companies.
Meta defends the approach by stating that, according to the company, it spent 200 billion dollars on research and development over the past five years. The article highlights a discrepancy between the experimental classification of the infrastructure and public statements that the data centers are intended to support the core products and business of Meta. In SEC filings, Meta acknowledges that the tax savings may be challenged. Reserves for uncertain tax positions rose by 45 % to 18.74 billion dollars.
According to the cited findings from New York Times, the audit firm EY approved the approach, helped develop it, and now offers it to other companies purchasing AI chips. The source does not state that tax authorities have already rejected this approach.
Why it matters
For large investments in AI infrastructure, tax classification can mean differences of billions in tax liability. For companies purchasing AI chips, a key point is that a similar approach is also being offered to other companies, but Meta itself acknowledges the risk of the savings being challenged. Investment planning therefore also depends on whether the tax claim can be defended.
Relevant practical impact
What this means
For a business
According to the article, a similar tax approach is being offered to companies purchasing AI chips. For financial planning, however, it poses a risk that the expected savings will later be challenged.
Risks and complianceCheck the original
Event sources
only one source so far · 1 publisher, 1 independent. We count feeds from the same owner only once.