Corporate tax revenues in the United States have declined significantly in recent years despite a surge in corporate profits and investment, according to newly released government data and expert analyses. For the fiscal year ending October 1, tax receipts from the corporate income tax dropped by about 15 percent, totaling $381 billion. This represents roughly a $70 billion decrease compared to the previous year and a more than 25 percent decline since fiscal year 2024, the last full year under the Biden administration.

The reduction in revenue coincides with tax legislation passed by Republicans and signed into law by former President Donald Trump, which allows companies to accelerate how quickly they can write off the costs of new investments. This accelerated depreciation provision permits businesses to fully deduct the cost of qualified assets, such as computing equipment, in the year the investment is made rather than spreading deductions over several years.

This change initially aimed to stimulate business investment, and academic research supports the view that such tax breaks can encourage companies to boost spending. However, the recent surge in investment spending has largely been driven by major technology firms racing to expand artificial intelligence (AI) infrastructure. Companies like Microsoft, Google, and Meta are funneling hundreds of billions of dollars into data centers, chips, and related systems, leveraging the tax incentives extensively.

Experts note that this rush into AI investment has intensified the disparity between rising corporate profits and shrinking tax revenue. While profits have reached record highs, the accelerated write-offs suppress the taxable income reported by these companies. This divergence is considered unusual but understandable given the current tax structure and investment patterns.

Critics argue that the tax break primarily benefits large corporations that would have invested heavily in AI infrastructure regardless of the incentive, thus serving as a subsidy for expenses companies were already planning to incur. Some Democrats, including Senator Ron Wyden of Oregon, have proposed limiting or excluding specific AI-related components, such as data center equipment, from the investment deduction to ensure the tax code captures more revenue.

Supporters of the tax provisions caution that the revenue loss is likely temporary, as deductions are tied to actual investments and will eventually phase out. Over time, as companies realize profits from their investments, the tax system is expected to collect taxes on these earnings. Advocates also stress that encouraging investment can have broader economic benefits, including increased productivity.

The broader fiscal picture includes growing government deficits, driven in part by higher military spending and increased interest costs on the federal debt, which has reached levels comparable to the country’s annual economic output. Despite these pressures, the economy remains relatively strong, with corporate profits and stock markets near historic highs.

Looking ahead, some economists estimate that AI-related infrastructure spending could total more than $10 trillion between 2025 and 2032, suggesting that accelerated depreciation will continue to play a significant role in the taxation of corporate earnings for years. The ongoing debate centers on balancing the objectives of incentivizing innovation and investment with ensuring sustainable tax revenues for the government.