Alec J. Berin, Matthew P. Suzor, and Quintin C. Cerione of Miller Shah LLP
Since the debut of OpenAI’s ChatGPT in late 2022, artificial intelligence (AI) has exploded from an experimental tool to a global industry. The exponential rise of generative AI, although providing companies and consumers with greater levels of efficiency and productivity, is putting pressure on American antitrust law to play catch up in regulating the growing AI market.
As AI becomes commonplace today, one of the greatest challenges it poses is that its building blocks—chips, cloud infrastructure, and large-language models—are largely controlled by only a handful of companies.[1] A major concern, therefore, is whether American antitrust law, which was largely designed during an industrial period dominated by railroads and manufacturing, can address the competitive risks of the AI era. Regulators and courts have started to express their perspectives about these issues, yet more questions than answers have emerged.
The Intersection of American Antitrust Doctrine and AI
The core of the American antitrust framework is comprised of the Sherman Antitrust Act (1890), Clayton Act (1914), and Federal Trade Commission Act (1914).[2] The Sherman Act was initially enacted in an effort to target monopolization by barring exclusionary practices, while the Clayton Act filled its holes by prohibiting mergers and acquisitions whose effect “may be substantially to lessen competition, or to tend to create a monopoly.”[3] Historically, courts have applied these laws to industries defined by physical assets, such as steel, oil, and operating systems.[4] Today, however, the market power increasingly consists of control over intangible items: data and algorithms.
Regulators are attempting to offer guidance on how these statutes apply in a digital and data-driven era. For example, in 2023 the FTC and DOJ issued revised Merger Guidelines, which warned that a merger could undermine competition if it “creates a firm that can limit access to products or services that its rivals use to compete.”[5] Although this is not directed exclusively at tech companies, this language nonetheless suggests antitrust law’s expanded focus on vertical integration—especially relevant for companies’ partnerships aimed at combining the control of AI infrastructure and data services.
The particular challenge for regulating market power in the AI sector is defining the relevant market. Because AI depends on key inputs—vast amounts of data and computational resources – rather than traditional products and services that have historically defined markets, delineating the relevant market is uniquely complex. This is clearly indicated in a 2025 report from the Congressional Research Service, which warns that “limited access to data” may threaten competition, regardless of whether AI services remain free to consumers.[6] In the coming years, determining whether AI regulation will be concentrated on the models, chips, or cloud services used for these products—or if they will be considered a single integrated stack—will be critical in influencing enforcement outcomes.
Early AI-Antitrust Legal Battles
In recent months, lawsuits against major tech companies have begun to address how far traditional antitrust principles extend into the AI space.[7] This October, a class-action lawsuit filed against Microsoft[8] alleged that its financial relationship with Open AI—particularly a deal granting Microsoft exclusive cloud computing that restricts the supply of computational resources needed to run ChatGPT—both limited market competition and artificially drove up ChatGPT subscription prices while diminishing product quality for millions of Open AI users.[9] Similar concerns are being raised by antitrust experts regarding Nvidia’s $100 billion partnership with OpenAI,[10] as experts fear that building such a relationship will give both companies an unfair advantage over their competitors.
Perhaps most notably, a September ruling by a federal judge in a landmark antitrust case against Google illustrated how AI may continue to be an obstacle in regulating monopolies.[11] Although the judge affirmed that “Google cannot use the same anticompetitive playbook for its GenAI products that it used for Search,” he insisted that the emergence of generative AI has granted companies a greater ability “to compete with Google than any traditional search company developer has been in decades” and ultimately spared Google from the harsh penalties.[12] This exemplifies the inherent tension of AI; a technology capable of fostering and hindering competition will prove only more difficult for regulators to address in years to come.
Critical Legal Questions to Consider
Going forward, courts will need to answer a series of questions to best address the competitive concerns of AI. First, as AI blurs product boundaries—with single companies being involved in many layers of the supply chain—determining whether these layers represent distinct or integrated markets has big implications for assessing anticompetitive behavior.
Second, because several of the most popular AI products offer services for free or at low costs, harm to consumers may lie outside the scope of price fixing but instead resulting from diminished product quality and restricted access to inputs.[13] It will be up to courts and regulators to determine when harm is being committed in the AI market.
Third, defining the line between integration and exclusion will become increasingly urgent. Though partnerships and acquisitions may accelerate innovation, unlawful exclusion may arise through integrated companies’ restriction of rivals’ access to essential inputs or result in self-preferencing through exclusive supply arrangements. Though this risk is outlined in the 2023 Merger Guidelines, it remains to be seen how courts will approach this issue in the coming years.
Notes
[1] See e.g., Jay Stanley, Will Giant Companies Always Have a Monopoly on Top AI Models?, ACLU (Aug. 20, 2025), https://www.aclu.org/news/racial-justice/will-giant-companies-always-have-a-monopoly-on-top-ai-models; Steven Levy, There Is Only One AI Company. Welcome to the Blob, Wired (Nov. 21, 2025 at 11:00), https://www.wired.com/story/ai-industry-monopoly-nvidia-microsoft-google/.
[2] See Sherman Antitrust Act of 1890, 15 U.S.C. §§ 1–38; Clayton Act of 1914, 15 U.S.C. §§ 12–27; Federal Trade Commission Act of 1914, 15 U.S.C. §§ 41-58.
[3] The Clayton Act of 1914, 15 U.S.C. § 18.
[4] See e.g., United States v. Columbia Steel Co., 334 U.S. 495 (1948) (applying the Sherman Act to the steel industry); FTC v. Sinclair Ref. Co., 261 U.S. 463 (1923) (applying the Federal Trade Commission Act and Clayton Act to the oil industry); United States v. Microsoft Corp., 346 U.S. App. D.C. 330 (2001) (applying the Sherman Act to operating systems).
[5] Federal Trade Commission & U.S. Department of Justice, Merger Guidelines (issued Dec. 18, 2023),
https://www.justice.gov/atr/2023-merger-guidelines.
[6] Congressional Research Service, Artificial Intelligence and Competition Policy (2025), CRS Insight No. IN12458, https://crsreports.congress.gov/product/pdf/IN/IN12458.
[7] Mike Scarcella, AI Users Sue Microsoft in Antirust Class Action Over OpenAI Deal, Reuters (Oct. 13, 2025 at 17:47 CDT), https://www.reuters.com/legal/government/ai-users-sue-microsoft-antitrust-class-action-over-openai-deal-2025-10-13/.
[8] Class Action Complaint, Samuel Bryant et al. v. Microsoft Corp., No. 3:25‑cv‑08733 (N.D. Cal. filed Oct. 13, 2025) (alleging anticompetitive restraints arising from Microsoft’s partnership with OpenAI).
[9] Scarcella, supra note 7.
[10] Jody Godoy, Nvidia’s $100 Billion OpenAI Play Raises Big Antitrust Issues, Reuters (Sept. 23, 2025),
https://www.reuters.com/technology/nvidias-100-billion-openai-play-raises-big-antitrust-concerns-2025-09-23/.
[11] See generally, United States v. Google LLC, 803 F. Supp. 3d 18 (D.D.C. 2025) (remedies decision addressing generative AI’s competitive effects).
[12] Id. at 99, 128.
[13] Scarcella, supra note 7.
