Thomas Przybylowski: Securities Plaintiffs' Bar Retools as AI, Tariff and Private Credit Lawsuits Surge in 2026

Federal securities class actions totaled 118 in H1 2026, implying roughly 236 for the full year and exceeding the prior 2023 peak, despite a 2025 decline in filings and a 25% drop in plaintiffs' attorneys' fees. AI-related suits reached 18 in H1 alone, surpassing the 17 filed in all of 2025, while six tariff cases, 10 largely non-U.S. pump-and-dump cases, and growing private-credit valuation claims broaden litigation exposure. The SEC's September 2025 decision to stop considering mandatory arbitration provisions when accelerating registrations could enable companies to divert investor disputes from class actions, setting up potentially material legal challenges.

Analysis

The investable implication is not broad AI multiple compression; it is a higher disclosure-risk premium for companies monetizing AI narratives ahead of auditable revenue, customer retention, or deployment evidence. Boards will likely shift toward more qualified language and slower KPI disclosure, which can reduce near-term promotional upside for AI-adjacent small caps while favoring scaled vendors with transparent backlog, recurring revenue, and conservative guidance. NVDA is comparatively insulated because its valuation debate rests more on hyperscaler capex and gross-margin durability than unsubstantiated capability claims, although customers' AI project delays would still affect the demand chain.

QUBT faces asymmetric downside because thin-liquidity, high-short-interest issuers are vulnerable to litigation-triggered financing constraints: a stock decline can impair warrant exercise proceeds, raise ATM dilution, and make customer or supplier diligence harder. The more important read-through is to speculative quantum/AI issuers rather than semiconductor leaders; expect greater dispersion over the next 1-3 months as underwriters, auditors, and exchanges demand tighter substantiation of technical and commercial claims.

Private-credit exposure is a 6-18 month watch item, not an immediate trade solely from filing trends. Public BDCs with externally marked, lower-transparency portfolios could face a valuation discount if credit losses or non-accruals reveal optimistic marks; that would pressure both NAV credibility and capital-raising capacity. The counter-consensus view is that litigation risk itself is usually immaterial to large-cap earnings: the actionable signal is whether legal scrutiny forces guidance cuts, restatements, delayed filings, or a higher cost of equity.

The principal falsifier is evidence that courts consistently dismiss complaints relying on expert or short-seller material before discovery, limiting settlement leverage. Conversely, a sustained rise in AI-related restatements, SEC comment letters, auditor changes, or adverse rulings on mandatory-arbitration provisions would validate a broader governance-risk repricing.

Key Decisions for Investors

• Maintain NVDA as a relative long versus a basket of speculative AI/quantum names; use a 3-6 month pair horizon. The thesis is quality-of-disclosure and financing dispersion, not a new NVDA earnings catalyst. Exit the relative trade if hyperscaler capex guidance weakens materially or NVDA data-center backlog/conversion commentary deteriorates.
• Avoid or maintain a tactical short bias in QUBT on rallies until it provides independently verifiable commercial revenue, customer concentration, and cash-runway disclosures. Size modestly because borrow cost, short-squeeze risk, and retail-flow volatility can dominate fundamentals; cover on a credible third-party customer validation or non-dilutive financing.
• Screen publicly traded BDCs, including ARCC, OBDC and FSK, for rising PIK income, Level 3 asset concentration, NAV marks above observable secondary loan prices, and non-accrual migration. Do not initiate a sector short from this article alone; trigger a relative short only when those indicators coincide with a guidance or NAV-mark revision.
• For AI holdings over the next reporting cycle, require an explicit checklist: AI revenue definition, contracted versus pilot customers, implementation timing, gross-margin impact, and model-risk disclosures. Treat withdrawal of prior AI KPIs, delayed 10-Q/10-K filings, auditor turnover, or an SEC disclosure inquiry as immediate position-review catalysts.