September 9, 2025
The landscape of American innovation shifted dramatically in the summer of 2025. Congress, through the One Big Beautiful Bill Act (OBBBA), rewrote the tax code to favor research and intellectual property at home, while California, with its Transparency in Frontier Artificial Intelligence Act (SB 53), demanded that developers of powerful AI systems expose the very processes they once guarded as trade secrets. Taken together, these two laws frame the new reality: Washington is enticing businesses to invent boldly on U.S. soil, even as states are raising the cost of secrecy and insisting on public accountability. For companies whose survival depends on research pipelines and intellectual property portfolios, innovation is no longer shaped only by market forces. It is being sculpted—sometimes smoothed, sometimes chiseled—by the hand of the law.
When Congress enacted OBBBA, it struck at the core of the 2017 Tax Cuts and Jobs Act. That law had required research expenditures to be spread out over years, a mandate many companies viewed as a drag on innovation. Section 174A now reverses course. As the statute itself provides:
In the case of domestic research or experimental expenditures which are paid or incurred by the taxpayer during the taxable year, the taxpayer may elect to treat such expenditures as expenses which are not chargeable to capital account. Such expenditures shall be allowed as a deduction for the taxable year in which they are paid or incurred.
The text matters: “shall be allowed as a deduction” restores what innovators had lost. Congress went further, offering choices for the years when amortization had already taken effect. A company may, in its first taxable year after 2024, “deduct the remaining unamortized amount of domestic research or experimental expenditures paid or incurred in taxable years beginning after December 31, 2021, and before January 1, 2025.” Small businesses, defined in the Act as those “with average annual gross receipts not exceeding $25,000,000,” are permitted to amend prior returns outright.
Foreign research tells a different story. The law requires those costs to be “deducted ratably over the 15-year period beginning with the midpoint of the taxable year in which such expenditures are paid or incurred.” In other words, Washington is not simply easing the burden of research. It is placing a thumb on the scale: innovate at home and deduct now, innovate abroad and wait fifteen years.
Though Section 174A speaks in the dry idiom of the tax code, its ripple effects extend into licensing and corporate structuring. By favoring domestic research and tightening the benefits of offshore IP, the law heightens scrutiny of agreements that cede influence over U.S.-originating intangibles. A reversionary clause or perpetual license once viewed as routine may now be read in light of Congress’s clear policy: that “domestic research or experimental expenditures” deserve favored treatment and protection. The practical result is a licensing environment where economic terms are inseparable from questions of regulatory exposure.
If OBBBA restores financial oxygen, SB 53 forces a measure of transparency. The Act, codified in California’s Business and Professions Code, requires that:
A covered developer shall clearly and conspicuously publish on its internet website the developer’s safety and security protocol, including a description of the testing procedures used to assess catastrophic risks posed by the developer’s covered model, and the risk mitigation measures adopted.
The law defines catastrophic risk with statutory precision. A “catastrophic risk” means a risk that, if realized, could “result in the death of, or serious injury to, more than 100 persons, cause severe economic damage of $1,000,000,000 or more, or facilitate the use of chemical, biological, radiological, or nuclear weapons, or a cyberattack resulting in comparable harm.”
Developers must also establish internal whistleblower channels. As the statute directs:
A covered developer shall establish and maintain an anonymous reporting mechanism to allow covered employees to disclose, in good faith, information regarding potential catastrophic risks or false or misleading statements regarding risk mitigation. An employer shall not take retaliatory action against a covered employee for making such a disclosure.
The law requires incident reporting within fifteen days of discovery and empowers the Attorney General to bring enforcement actions. In short, California has written transparency into law, turning once-internal safety documents into public, and legally significant, artifacts.
California’s experiment is not isolated. New York’s proposed RAISE Act, Michigan’s AI Transparency Act, and Colorado’s AI Act all echo similar concerns. And even in states without statutes, Attorneys General have invoked existing authority. The patchwork is spreading, but California’s statute provides the most vivid template of how accountability and intellectual property collide.
What emerges from these statutory texts is a compact. Section 174A invites companies to bring their innovation home with the promise that their research costs “shall be allowed as a deduction” in the year they are incurred. SB 53 insists that when that innovation takes the form of frontier AI, the public has a right to know “the testing procedures used to assess catastrophic risks” and to rely on protected insiders if disclosures fall short.
For companies, this means intellectual property is not just an asset—it is a legal touchpoint, defined and constrained by statutes that speak in clear imperatives. Counsel cannot treat tax elections, licensing clauses, or risk assessments as isolated matters. They are interconnected expressions of a legal order that is reshaping how innovation occurs. In 2025, the law’s own words are as much a part of the story of invention as the patents and protocols that innovators produce.