Publications |
11/04/26
By
Adam Savett
Few business stories have captured investors like artificial intelligence. Few have also produced lawsuits as quickly. When a company talks up its AI plans or the impact AI will have on their business and its stock later falls, shareholders increasingly sue, claiming they were misled.
The numbers: A Slow Rise, A Plateau, Then a Jump
The growth has come in two stages. Cornerstone Research, who produces an annual report on securities litigation with Stanford Law School, counted seven AI-related securities class actions in 2023, then 15 in 2024 and 16 in 2025. That is more than a doubling from 2023 to 2024, followed by a flat year. Other trackers differ slightly: NERA Economic Consulting counted 17 for 2025, while The D&O Diary, a respected blog on corporate liability written by Kevin LaCroix, counted 16.
Then came 2026. By midyear, Cornerstone had already counted 15 new AI-related cases and NERA had counted 18. NERA’s figure already topped its total for all of 2025. Cornerstone’s pace works out to roughly 30 for the year, about double 2025. The D&O Diary tallied 24 AI-related filings through September. AI-related cases accounted for nearly 14% of all new securities class actions filed this year.
AI cases punch above their weight. In 2025 they accounted for only about 8% of filings but accounted for 57% of the total investor stock losses Cornerstone measures, because the companies involved were often large and their share prices fell sharply.
The broader picture is also busy. NERA counted 118 federal securities class actions in the first half of 2026, a pace that would produce the most in a year since 2020. For comparison, Cornerstone counted 207 for all of 2025, down from 226 in 2024.
What investors are alleging
The AI-related securities class actions tend to fall into a few recurring themes.
“AI washing.” Like “greenwashing,” this means exaggerating how much AI a company really uses. Oddity Tech was sued after a short-seller report described its AI as little more than a questionnaire. A pension fund sued Apple, alleging it promoted a smarter Siri before it had a working version.
“AI Diversion.” Suits against companies accused of hiding the financial drag, capital expenditures, or margin compression caused by pivoting core infrastructure toward unproven generative AI tools, such as those against Microsoft, ZoomInfo, and Sprinklr.
Spending and infrastructure. Suits against companies that build the computing backbone for AI are a growing subcategory. CoreWeave was sued in January 2026 over its ability to meet customer demand, and Oracle was sued over statements about its AI-related capital spending. Commentators expect such “AI spend” claims to only grow as the costs of the AI arms race climb.
Stock promotion. Some suits claim companies used AI announcements to pump up their share prices, such as those against Richtech Robotics, ODDITY Tech, and Datavault AI. According to an analysis of NERA by The D&O Diary, pump-and-dump style cases never exceeded two a year from 2022 through 2025 but reached 11 in the first half of 2026 alone. These are not all AI cases, but AI is a common thread.
Notably, private securities litigation is increasingly filling a gap. Bloomberg Law reports that the SEC under Chairman Paul Atkins has stepped back from AI-washing enforcement, yet private suits keep coming.
Cases that have been dismissed
So far, courts have often been skeptical of AI-related securities class actions. To proceed, plaintiffs must point to specific facts showing a statement was false and that executives meant to mislead. Vague optimism is often dismissed as “puffery.” Recent outcomes include:
• UiPath: a robotic process automation company sued in 2024 over AI-related growth claims. In September 2026 a federal court in New York dismissed the second amended complaint with prejudice.
• TELUS International: in August 2026, a New York federal judge dismissed claims that the company exaggerated its AI capabilities, finding that the allegations did not change the core fact that the company did have AI capabilities, and that no specific false product claim was identified.
• Innodata: a New Jersey court dismissed the case in April 2026 but allowed plaintiffs to try again.
• iLearning: a federal court in Maryland dismissed this suit against an AI learning platform with prejudice in March 2026.
• C3.ai: in March 2026 a court threw out fraud claims in one older case, leaving only a narrow claim tied to its IPO paperwork. In a newer suit, the court dismissed the first complaint and the plaintiff refiled in August.
• GitLab: a software company sued in 2024 over AI-related growth claims. The case ended with a complete dismissal and final judgment in January 2026.
• Tempus AI: after it was sued in June 2025, the lead plaintiff acknowledged it lacked enough information to meet the legal pleading standard. The case was dismissed in November 2025.
Is AI-related securities litigation easier or harder to win than other kinds? Opinions differ. Some commentators have argued that AI cases settle more often and are dismissed less often than other securities suits. Others have noted that dismissal rates for AI cases rose as judges grew more familiar with the subject. With so few resolved cases, both views rest on sample sizes that are likely too small to be valid predictors.
Cases that have settled
Settlements are still rare. The clearest example is Evolv Technologies, which in April 2026 sought court approval of a $15 million settlement. Investors had alleged the company overstated the effectiveness of its AI-powered weapons detection system and improperly recorded revenue.
The larger picture suggests more are coming. Across all securities class actions, Cornerstone recorded 39 settlements in the first half of 2026, up from 32 a year earlier. The average settlement was $56.4 million, above the 2017-2025 average of $46.8 million.
The biggest AI cases, including Apple, CoreWeave Intuit, and Oracle, are still early. Motions to dismiss were pending in several as of this spring, and NERA reports the median case takes about 1.5 years to reach a dismissal. Settlements for the 2025 and 2026 wave of filings are therefore likely to appear in 2027 and beyond.
What it means
Securities litigation tends to come in waves: the dot-com bubble, subprime mortgage meltdown, SPACs, COVID-19, and crypto. AI appears to be the latest. It is important to note that these are otherwise conventional securities suits with AI as the subject matter, and the underlying legal rules have not changed.
For investors, the lesson is that a lawsuit is an allegation, not a finding, and many have been dismissed. For companies, the message is that AI claims in earnings calls and marketing now carry legal risk, and precision matters. Whether the wave continues will depend on how markets treat AI stocks and on what courts decide in the big pending cases.