OpenAI wants tax credits for AI data centers—will Washington agree?

What began as a policy instrument to rebuild semiconductor capacity in the United States is now morphing into a broader reindustrialization debate. OpenAI, the company behind ChatGPT, has formally requested that the U.S. government expand the scope of the CHIPS and Science Act’s investment tax credit. The artificial intelligence firm wants the credit, originally meant to stimulate semiconductor fabrication, to also cover AI data centers, server production, and critical power grid components like transformers and switchgear.

In a letter dated October 27, 2025, addressed to the White House Office of Science and Technology Policy, OpenAI framed its request as a national competitiveness issue. The firm’s Chief Global Affairs Officer, Chris Lehane, made the case that hyperscale compute facilities, customized AI server infrastructure, and power grid hardware should be considered as strategic to American industrial policy as semiconductor fabs themselves. OpenAI believes the current 35 percent Advanced Manufacturing Investment Credit under the CHIPS Act should be extended to cover this expanded infrastructure stack.

Chief Executive Officer Sam Altman has reinforced this view in public commentary, stating that the infrastructure powering frontier artificial intelligence models is now a critical national asset. This includes not only advanced graphics processing unit clusters and custom inference engines, but also the physical environments that host them and the electrical systems that keep them online. The firm’s central thesis is simple but transformative: tax policy must evolve in parallel with the industrial scale and geopolitical relevance of artificial intelligence.

How big is OpenAI’s infrastructure vision and what role does public policy play?

OpenAI has publicly projected a $1.4 trillion capital commitment to build out its AI infrastructure footprint across the United States. This is not a marketing figure but a concrete projection of the multiyear investment required to scale training compute, inference capacity, chip design and manufacturing, data center facilities, and the associated grid upgrades required to deliver uninterrupted power to all of the above.

The firm argues that without favorable tax treatment and policy alignment, this investment will be slower, more fragmented, and less geographically inclusive. Just as the CHIPS Act was designed to revive U.S.-based chip fabrication to reduce foreign dependence, OpenAI is now asking policymakers to recognize the rest of the AI value chain as equally foundational. The company’s leadership has stressed that permitting bottlenecks, long grid equipment lead times, and unpredictable interconnection processes are already slowing progress. According to OpenAI, expanding tax credits to cover the full AI infrastructure stack would lower the cost of capital, accelerate timelines, and de-risk the kind of long-term investment now required to sustain American leadership in artificial intelligence.

Why is this proposal different from traditional tech-sector lobbying efforts?

Unlike past technology industry lobbying efforts that focused on R&D credits or telecommunications subsidies, OpenAI is targeting manufacturing and infrastructure. The request does not involve end-user technology, app ecosystems, or content platforms. Instead, it reflects the physical reality of the artificial intelligence build-out. That reality is grounded in heavy capital expenditure, multi-year asset timelines, and intensive power draw.

Importantly, OpenAI is not requesting direct subsidies for its own projects or government loan guarantees. The company is seeking structural tax incentives that would be available to the broader sector. This makes the ask more aligned with long-standing industrial policy tools than tech-specific handouts. Still, the politics are sensitive. Bringing commercial data centers under the umbrella of strategic manufacturing may face resistance from lawmakers who see such projects as private, profit-generating infrastructure.

The political optics are further complicated by the rapid monetization of AI systems. If tax credits flow to companies that are already extracting commercial value from AI tools, there is a risk of backlash. Critics have warned against privatizing gains while socializing infrastructure costs. Supporters, on the other hand, argue that the public benefits of more accessible, sovereign compute outweigh such concerns.

What are the proposed changes to the CHIPS Act and tax policy?

OpenAI wants to extend eligibility under the Advanced Manufacturing Investment Credit to include AI-optimized server manufacturing, hyperscale data center development, and electrical grid components such as transformers, electrical steel, copper cabling, and substation hardware. In practical terms, this means a wide range of suppliers and contractors involved in the build-out of AI facilities would qualify for a 35 percent investment tax credit, the same rate currently reserved for semiconductor fabrication projects.

The company also wants Congress and regulators to streamline grid permitting processes. Power interconnection queues have grown significantly in recent years, driven by both renewable energy and digital infrastructure demand. In some cases, delays can last for several years. OpenAI views these delays as incompatible with the pace of AI model development and deployment.

Taken together, these policy requests signal a shift toward treating AI infrastructure as public-interest infrastructure. OpenAI’s framing is not just about commercial efficiency, but about ensuring that the physical stack required for AI competitiveness is built and maintained inside U.S. borders.

Which companies and sectors could benefit if the CHIPS Act is expanded?

If OpenAI’s recommendations are adopted, the ripple effects would be felt across multiple industries. Companies manufacturing AI inference and training servers, such as NVIDIA Corporation and Super Micro Computer Inc., could benefit from a surge in demand as tax relief improves deployment economics. Server chipmakers including Advanced Micro Devices Inc. and Intel Corporation may also see expanded addressable markets for AI-dedicated compute nodes.

Grid infrastructure manufacturers are perhaps the biggest potential winners. Eaton Corporation, ABB Ltd, General Electric Company, and Siemens AG are among the companies currently navigating order backlogs and strong market demand for power distribution equipment. If tax credits stimulate more orders from AI infrastructure developers, these manufacturers could see significant tailwinds.

Specialty material producers in electrical steel, copper, and aluminum may also experience growth. Substation developers and electrical EPC firms would similarly be positioned for project awards if the permitting environment improves alongside capital incentives. Finally, real estate investment trusts focused on data centers, including Equinix Inc. and Digital Realty Trust Inc., could see enhanced margins and improved site selection flexibility.

The investor landscape is already reflecting this shift. Exchange-traded funds targeting U.S. infrastructure, smart grid development, and AI server ecosystems have seen increased institutional flows. Private equity firms are now aggressively pursuing power hardware manufacturers and hyperscale infrastructure developers, a trend that aligns closely with OpenAI’s policy strategy.

How are U.S. lawmakers and regulators responding to OpenAI’s proposal?

The White House has not issued a formal response, but early signals from administration officials suggest skepticism. A spokesperson reiterated that there are no current plans to provide federal guarantees or tax breaks for general-purpose data center projects. Congressional aides have echoed this view, suggesting that credits for server manufacturing or grid hardware may be viable, but tax relief for commercial server farms will likely face opposition.

This cautious sentiment is grounded in concerns about corporate welfare and the precedent of subsidizing infrastructure that enables profit-maximizing AI products. Lawmakers from both sides of the aisle have expressed interest in ensuring that any future tax policy changes include performance-based criteria, domestic content thresholds, or strategic use-case alignment such as national security, research, or education.

Nonetheless, OpenAI’s proposal is already shaping the legislative discussion. Several members of Congress are reviewing ways to expand the CHIPS Act tax credit to include high-assurance infrastructure that supports artificial intelligence growth, even if the scope is narrower than OpenAI originally proposed.

What are the broader implications for U.S. competitiveness and industrial policy?

OpenAI’s proposal may not pass in full, but the conversation it has initiated will have long-lasting effects. It marks a turning point in how artificial intelligence infrastructure is viewed, not just as a commercial necessity, but as a sovereign capability. As AI models grow in complexity and resource intensity, the physical stack that supports them is becoming as strategically important as the software they run.

This realization is prompting a policy rethink. Artificial intelligence innovation now depends as much on power, grid reliability, and data center throughput as it does on research breakthroughs. In this context, tax policy becomes a lever not only for economic efficiency, but for global competitiveness.

If the United States fails to scale its AI infrastructure in time, it risks losing its leadership position to other regions. China has already built state-supported AI parks. The European Union is funding sovereign compute platforms for research and defense. OpenAI is asking whether the United States is prepared to match those commitments with industrial-grade public policy.

Whether the CHIPS Act is expanded or not, the age of artificial intelligence is now anchored in physical infrastructure. That means wires, steel, and substations, and not just chips and algorithms. OpenAI’s intervention has clarified that reality for Washington and for Wall Street.