By: Joyce Jia

In early December 2025, the Financial Times reported that Anthropic retained the Palo Alto-based law firm Wilson Sonsini to help facilitate a public offering as early as 2026. The move may signal a strategic effort to outpace rival OpenAI, which had also been eyeing a 2026 listing. Following the report, OpenAI executives privately expressed concern about being beaten to market. Regardless of who lists first, 2026 is shaping up to be a landmark year for AI in public markets, and a long-awaited opportunity for investors and observers to lift the veil on the economics of the generative AI giants driving the Fourth Industrial Revolution.

More than an IPO: A Governance Blueprint for the AI Era

As AI systems increasingly power healthcare systems, financial markets, communications, and public services, they are functionally becoming part of the nation’s core infrastructure. Unlike pharmaceuticals subject to FDA approval, financial intermediaries registered with the SEC, or telecommunications providers licensed by the FCC, frontier AI developers operate without a comparable, unified regulatory gatekeeper. Existing oversight has been fragmented across state privacy regimes and FTC consumer protection authority, leaving the public with significant gaps in understanding how these companies operate, what their models optimize for, and how deeply they depend on cloud infrastructure, semiconductor supply chains, and energy capacity. Moreover, propelled by unprecedented growth trajectories, they have financed their expansion almost entirely through venture capital and private placements, without being subject to public disclosure obligations and public accountability that securities registration demands. 

When Anthropic or OpenAI files for its initial public offering, it will become the first frontier AI company to submit its business model to the full discipline of SEC disclosure. The filing will not merely determine one company’s market valuation. It will likely shape the disclosure template and regulatory precedent governing how artificial intelligence integrates into public capital markets, compelling for the first time a legally enforceable accounting of how frontier AI companies actually operate, what system risks they pose, and whether traditional corporate governance frameworks are adequate to contain them. 

Tip of the Iceberg: What SEC Disclosure Will Finally Force Frontier AI to Reveal

The S-1 registration statement must include all information specifically required by Form S-1, as well as any material information needed to prevent the included statements from being misleading. Current SEC disclosure requirements do not specifically address modern AI risks, though the SEC’s Division of Examinations has identified AI as a priority focus area. Under established materiality doctrine, disclosure is required for any matter to which a reasonable investor would attach importance in deciding whether to purchase a security. Furthermore, Item 105 of Regulation S-K requires a “Risk Factors” section discussing material factors that make an investment speculative or risky, with a concise explanation of how each risk affects the registrant.

Under this framework, investors will likely see for the first time the size of an LLM company’s accumulated deficit and its projected path to profitability. They will also gain visibility into the terms of related-party contracts and material agreements with cloud and semiconductor suppliers who serve as both vendors and strategic investors. Disclosure will additionally include training data liabilities such as the Bartz v. Anthropic settlement. Finally, “Business” and “Management’s Discussion and Analysis” sections may provide a first-of–its-kind monetization narrative, explaining what these companies’ algorithms are optimized for, what commercial incentives are encoded in model training, and how ecosystem allocation decisions generate indirect value. Together, these disclosures will expose the structural economics of frontier AI development in a way that no private financing round has ever required.

Disclosure priorities will also differ across companies. Anthropic’s Long-Term Benefit Trust grants outside directors veto power over decisions conflicting with the company’s safety mission, even where those decisions would maximize shareholder value. This structure directly challenges the shareholder primacy doctrine established in Dodge v. Ford, the 1919 Michigan Supreme Court decision holding that corporate directors must prioritize shareholder returns over broader social objectives. The arrangement raises a critical question: Can mission-driven governance survive activist investors and proxy battles once the company goes public? 

OpenAI, having completed its conversion to a Delaware public benefit corporation in October 2025, faces a structurally distinct set of disclosures: the terms of its nonprofit’s retained equity stake, the scope and exclusivity of its Microsoft dependency, and the resolution of Elon Musk’s active litigation challenging the conversion will each require disclosure under existing SEC frameworks in configurations not previously tested at this scale.

Materiality Standards for the AI Future

Much of what frontier AI companies should disclose is already compelled by existing SEC rules. What is needed is rigorous application of the comment process to establish materiality standards specific to the AI sector. Anthropic published a new model Constitution last month articulating how its AI systems should reason and prioritize competing values. OpenAI, for its part, has committed to a governance structure that retains a nonprofit equity stake precisely to preserve the primacy of safety and security obligations even after its corporate restructuring. Both are well positioned to lead on disclosure voluntarily, setting best practices that regulators and competitors will be compelled to follow. That makes these securities filings as consequential as any AI safety research. Markets, no less than models, need alignment.

By: Anusha Nasrulai

The Washington state legislature is currently considering the Driver Privacy Act, a bill regulating automated license plate readers (ALPRs). ALPRs are cameras that capture license plate information of a vehicle along with location and time information. Currently, many agencies’ retention and sharing of ALPR data is subject only to internal policy, if any exists. The surveillance capabilities of ALPR systems have profound consequences, such as chilling the exercise of civil liberties and invading the privacy of vulnerable individuals, including immigrants or people who come to Washington to access reproductive or gender-affirming health care. The Driver Privacy Act presents the opportunity to raise the floor of privacy protections afforded by the Federal and Washington State Constitutions.

The Driver Privacy Act limits authorized uses of ALPRs to law enforcement, parking and toll enforcement, and transportation agencies. The current bill also puts forth warrant thresholds, a 21-day data retention period, and auditing requirements for agencies. Additionally, the bill places restrictions on accessing and sharing ALPR data and creates civil and criminal liability for violations under the Act. The bill puts specific prohibition on ALPR use around protected activities, including exercising First Amendment rights and accessing healthcare.

Advocates are calling on lawmakers to strengthen the bill by further limiting the data retention period and prohibiting third party vendors and agencies from sharing ALPR data without a warrant.

Eyes on Washington

ALPRs have faced increased scrutiny in Washington state in the past year. In October 2025, the UW Center for Human Rights released a report exposing how immigration enforcement and other out-of-state law enforcement access data from ALPR systems operated by Washington agencies. This is despite the Keep Washington Working Act and Shield Law that are intended to limit local law enforcement assisting federal immigration enforcement and “protect[] people in Washington from civil and criminal actions in other states that restrict or criminalize reproductive and gender-affirming care.” Many municipalities have halted their use or procurement of ALPRs, at least till the state passes guidance. Washington now is at a turning point, to either implement guardrails that protect individuals’ privacy from undue government surveillance, or pass legislation that sanctions expanded use of ALPRs across Washington state.

Flock Safety, a widely-adopted ALPR vendor in Washington, has faced “growing controversy” for enabling ALPR data operated by public agencies to be accessed by federal agencies and other jurisdictions. As of this June, Washington has 80 cities, six counties, and three tribes using Flock cameras. While Flock Safety has received nationwide attention recently, there are other prominent ALPR vendors used in Washington, including Vigilant Solutions (Motorola) and Axon (formerly Taser). 

How ALPR data is used, stored, and shared matters because these cameras capture more than license plate information. Law enforcement use ALPRs during real-time investigations, by checking a vehicle’s license plate information against a “hot list” of vehicles associated with an investigation or reported crime. The information collected by ALPRs can be cross-referenced with other law enforcement or public agency databases to identify individuals which vehicles are registered to. Law enforcement can also search historic ALPR data to track the direction, speed, and travel patterns of a vehicle. In aggregate, ALPR data can reveal sensitive information about where an individual frequents and their travel patterns. Furthermore, ALPR photos can capture the likeness of drivers, passengers, and nearby surroundings.

Constitutional Concerns

The specific protections provided by the Driver Privacy Act are guided by Constitutional principles. The surveillance capabilities of ALPRs can have a “chilling effect” on the exercise of Constitutionally protected activities. Awareness of constant surveillance may alter or deter people from exercising their protected rights of expression, association, and religion. These concerns are valid given historic law enforcement surveillance of political rallies, protests, and places of worship.

ALPRs also implicate the Constitutional right to privacy. The Supreme Court has not required a warrant for law enforcement to collect and search license plate information because there is a diminished expectation of privacy due to the systematic regulation of vehicles and drivers’ movements taking place on public roads. Though, the Court has also addressed in United States v. Jones , Carpenter v. United States, and Kyllo v. United States, whether law enforcement’s use of emerging surveillance technologies infringes on Fourth Amendment protections.

The Supreme Court has found that a warrant is required before installing a GPS tracker or obtaining cell site location information (CSLI) to track an individual’s long-term movements. Justice Sotomayor wrote a concurring opinion in Jones highlighting how emerging technologies have enhanced law enforcement surveillance capabilities without physical intrusion: “GPS monitoring generates a precise, comprehensive record of a person’s public movements that reflects a wealth of detail about her familial, political, professional, religious, and sexual associations.” The aggregation of ALPR data to reveal historical travel patterns raise concerns similar to those articulated in the Jones concurrence. Similarly, the Court in Carpenter, was concerned by how CSLI time-stamped data provides an intimate view into a person’s life and is cheaper and more easily accessible compared to other surveillance strategies. This is comparable to retained, historical ALPR data.

The Supreme Court recognized in Kyllo that people’s Fourth Amendment protections should not be left to “the mercy of advancing technology.” Law enforcement use of sense-enhancing technology in the form of infrared scanners to collect information “that could not otherwise have been obtained without physical ‘intrusion into a constitutionally protected area,’” was found to be a search subject to Fourth Amendment protections. While cars have lesser Constitutional privacy protections than homes, modern ALPR systems with embedded AI also provide law enforcement with extra-sensory capabilities that may implicate the Fourth Amendment.

Federal courts have yet to conclude that police use of ALPRs violate Fourth Amendment search and seizure requirements. The Washington State Constitution provides an affirmative right to privacy, enshrined in Section 1, Article 7. Washington courts have interpreted this article to create a higher standard for lawful search and seizures. As of December 2024, 18 states already have ALPR laws, with more states considering passing ALPR legislation. Washington is now contemplating joining these states in passing ALPR regulations.

Where the Bill Stands Now

The Driver Privacy Act has already passed in the Senate with amendments. The House Civil Rights & Judiciary Committee will hold a public hearing in consideration of the amended bill on February 18^th.

The Driver Privacy Act not only regulates the use of ALPRs in Washington state, but it can also create meaningful privacy protections for all Washingtonians.