What public companies just told the SEC about AI, read straight from the filings. Every quote links the original document.
An insurer files to enter the AI cloud
Reliance Global Group (EZRA), a specialty insurance holding company, disclosed in an October 23 8-K that it has signed a letter of intent to acquire an unnamed AI cloud computing business. The target runs on-demand GPU capacity and a marketplace where third-party GPU owners earn revenue by listing their hardware. Reliance plans to buy additional GPUs after closing.
The proposed acquisition would mark Reliance's entry into the AI cloud market, with plans to scale rapidly through investment in a Company-owned GPU fleet and expansion of the acquisition target's proprietary marketplace. The business being acquired delivers on-demand computing capacity for artificial intelligence and other compute-intensive applications while enabling third-party GPU owners to earn revenue by offering their hardware through its platform.
No acquisition price was disclosed. The filing describes a letter of intent, not a signed deal, and Reliance has no prior track record running compute infrastructure. The sector pivot is unusual enough to flag; it is not complete.
West Texas stranded gas, routed to AI compute
Olenox Industries (OLOX) closed a $21.7 million acquisition of two companies near the Waha Hub in West Texas: Wildboy Holdings, which holds a power generation asset, and IPD, an infrastructure firm. Payment was almost entirely in Olenox Series F preferred stock convertible into common shares.
The acquisitions advance Olenox's strategy to build an integrated gas-to-power-to-compute platform, connecting natural gas resources with electricity generation and higher-value applications, including artificial intelligence, data centers and digital-asset operations such as the Company's CS Digital business.
The Waha Hub location is specific: pipeline congestion has made natural gas there among the cheapest in North America for extended periods, drawing compute ventures that need low-cost power. Olenox is a small company, but the geography and stated intent are concrete.
Agentic AI in bank operations, on the record
Valley National Bancorp (VLY) filed its Q3 2026 earnings presentation on October 22, with an AI roadmap that describes agentic AI already running in credit underwriting and loan servicing and an autonomous voice agent handling outbound auto loan collections. The bank does not frame these as pilots.
Agentic AI for Credit Underwriting & Loan Servicing Agentic AI automates elements of underwriting process and workflow orchestration to reduce manual steps for faster loan origination. improved client experience, and more efficient servicing...
SS&C Technologies (SSNC) reported $1,216.3 million in operating cash flow for the nine months ending September 30, 2026, up 10.4 percent year over year. CEO Bill Stone cited agentic AI as the next transformation driver for the company's financial services and healthcare clients at a September conference that drew more than 1,000 industry executives.
RingCentral (RNG) broke out agentic AI as its own named risk category in its October 23 10-Q, distinct from machine learning and large language models, noting that deploying it carries no assurance of business benefit. Naming agentic AI separately from AI broadly in a 10-Q risk section is newer practice for enterprise software filings.
Community banks warn of AI-assisted attacks; Mattel flags regulatory risk
Heritage Financial (HFWA), a Washington State community bank, went beyond standard cybersecurity boilerplate in its October 22 investor presentation, warning of
...sophisticated attacks using artificial intelligence and similar tools
against its IT systems and third-party processors. Primis Financial (FRST) took the competitive angle in its October 23 8-K, cataloguing rivals' growing use of AI as a source of intensifying pressure among financial institutions. Both filings show that AI risk language in bank disclosures has moved from generic acknowledgment to specific threat and competition categories.
The less-expected filer is Mattel (MAT), which itemized in its October 22 earnings release the regulatory risk of navigating AI across new investment and product development areas as a distinct risk factor, sitting alongside toy safety and supply-chain disclosures. Consumer toys are not where AI regulatory risk language tends to surface first.