What public companies just told the SEC about AI, read straight from the filings. Every quote links the original document.
People Inc Names LLM Licensing as a Revenue Source
In a September 11 8-K, People Inc (PPLI) listed content licensing revenue by channel and named AI companies alongside Apple News+ as a distinct source:
Content licensing royalties are earned from our relationship with Apple News+ as well as other content use and distribution relationships, including utilization in large-language models and other artificial intelligence ("AI") related activities.
Publishers have been signing LLM licensing deals for roughly two years. What this filing adds is the accounting structure: People Inc treats AI utilization as a named revenue type within its content licensing segment, at the same tier as a disclosed Apple relationship. That is the level of specificity at which a revenue stream starts to matter to investors.
Agentic AI Reaches Substantive Risk Disclosures in Annual Filings
Guidewire Software (GWRE) filed its 10-K on September 11. The company sells core software to property and casualty insurers, and its annual report named agentic AI as a potential disintermediation threat:
as artificial intelligence technologies, including machine learning, generative and agentic AI, and automated decision-making technologies (collectively, "AI technologies") continue to evolve, our existing and potential customers may leverage these capabilities to develop or enhance their own solutions that could reduce or eliminate the need for our solutions.
A large insurer building agentic AI pipelines could, in theory, route around a software vendor. Guidewire naming that scenario in its 10-K is a more specific disclosure than the generic AI competition language most companies use.
Rent the Runway (RENT) included agentic AI in its September 11 10-Q risk factors, listing it alongside machine learning and automated decision-making as technologies the company must keep pace with. A fashion rental company itemizing agentic AI in a quarterly filing suggests the phrase has completed its move into general corporate boilerplate.
Zeta Global (ZETA) named agentic AI regulatory risk in a 424B5 prospectus supplement filed the same day, citing potential failure to comply with evolving frameworks governing agentic AI systems as a factor that could affect results. That is a capital-raise document, where risk language is reviewed closely before printing.
An ETF Puts "Agentic AI" on a Registered Product Label
ETF Opportunities Trust filed a 485BPOS on September 11 updating prospectuses for a shelf of Tuttle Capital products. Among them is the Tuttle Capital Agentic AI Income Blast ETF, registered alongside AI Healthcare and AI Emerging Leaders variants. Tuttle's covered-call income products are small, but the filing marks the moment agentic AI became a named category on a registered investment product label, not just a theme described in marketing copy.
MindWalk Holdings (HYFT) filed a 6-K on September 14 describing itself as building infrastructure specifically for life sciences AI and agentic AI workloads, with a proprietary biological pattern representation refined over 20 years of curation at its core. Small-cap and early-stage, but the self-description is precise: agentic AI infrastructure for drug discovery is a narrower and more testable claim than general AI positioning.
SeeQC's S-1 Frames Quantum Hardware as AI Infrastructure
SeeQC (SEQC) filed an S-1 on September 11 to raise capital for its cryogenic quantum chip business. The company describes tightly coupled integration between quantum processing units and GPUs as a path to supporting AI workloads, and frames its technology as relevant to both AI-driven quantum computing and quantum-accelerated AI. The framing is deliberate: it ties a nascent quantum hardware story to the AI infrastructure spending cycle where capital has been concentrated. SeeQC is candid in the filing that the commercial market for this hardware does not yet exist at scale.