The Misguided Rush to Invest in AI
As private investors eagerly dive into AI investments, there’s a misconception among politicians that governments should jump on the bandwagon, too.
NVIDIA has seen its market valuation soar, surpassing $5 trillion, indicating a robust demand for AI technology. Major players like Microsoft, Google, Amazon, and Meta have poured $230 billion into AI in 2024 alone and plan to invest $325 billion this year in AI infrastructure. Venture capital is also flooding into numerous AI startups.
The Trump administration’s America’s AI Action Plan, released in July 2025, highlights this misguided view. It calls for federal agencies to use funding mechanisms to provide various grants for AI infrastructure, asserting that “winning the AI race will usher in a new golden age of human prosperity.”
Additionally, President Trump has signed a presidential order directing the Commerce Department to offer “financial support, including loans, grants, and tax incentives,” for AI data centers that require heavy power consumption.
This trend isn’t limited to the U.S. Canada is planning to allocate $2 billion aimed at developing AI computing infrastructure, while the EU is launching InvestAI, with an ambitious goal of raising $263 billion for AI initiatives.
The enthusiasm behind these initiatives signals that private markets are functioning well on their own. Subsidies are more justified when there’s a systematic underfunding of societal needs, not when private investment is already thriving.
Strict policies should focus on identifying specific gaps that private investors might overlook and target those areas. However, the AI Action Plan and Executive Order employ a broad approach, offering subsidies for everything from data centers to workforce training without clarifying where private investment might be insufficient.
A bigger question looms: can governments effectively pinpoint socially beneficial projects that lack funding? Private investors are risking their own capital, relying on their expertise, while politicians often allocate taxpayer money to appease regional interests. Using taxpayer funds to pick winners in a thriving private sector usually leads to inefficiency rather than innovation.
This scenario is reminiscent of the Obama administration’s push for solar manufacturing, where a $535 million loan guarantee to Solyndra ultimately resulted in bankruptcy. Some justify such failures as the cost of exploring valuable industries, but that reasoning doesn’t hold for AI.
Unlike the solar energy sector at that time, AI is currently flush with private capital. The grants in question do not address any alleged underfunded social aims but merely add taxpayer money to existing private investments.
It’s true that governments play a crucial role in AI development. Concerns about Chinese control over AI are valid, but their competitive focus in AI tends to leans toward military capabilities and advanced research, rather than commercial data centers, which is where private capital is rapidly being invested.
If subsidies are warranted, they should be channeled into areas where private markets fall short—like defense research or national security—rather than to profitable businesses that are already thriving. Simply put, showering American companies with subsidies while they succeed on their own won’t effectively counter China’s military advancements in AI.
Yes, AI might transform the economic landscape, but this is precisely why policymakers should exercise caution instead of giving in to the urge to subsidize. Taxpayer funds should be reserved for areas private investors overlook, such as defense and fundamental research, instead of bolstering cash-rich data centers. The AI surge doesn’t require government assistance; what’s needed is a bit of restraint.
