Anthropic is heading into a potentially record-setting stock market debut with a Federal Trade Commission inquiry open and a body of AI liability law that the company concedes is still unsettled, according to the public IPO filing it submitted to the Securities and Exchange Commission.
The disclosures will likely figure into discussions advisors are having with clients who have the San Francisco-based company's Nasdaq listing under the ticker ANT on their radar.
At the same time, a growing roster of advisory technology providers are wiring Claude into custodial, CRM and planning data through Anthropic’s dedicated financial advisor product, which added another batch of partners on Tuesday.
Anthropic's filing indicates that its models and services could draw product liability, negligence, consumer protection and failure-to-warn claims stemming from harms allegedly caused by model outputs, by autonomous AI agents or by the way customers use the technology. Whether AI systems should be treated as products or as services remains an open legal question, the company said, and the answer could shape how much liability it ultimately carries.
The prospectus also bared details of a civil investigative demand the FTC sent Anthropic this month, seeking material on consumer protection, privacy, data protection, and product safety and security.
A civil investigative demand is a compulsory, subpoena-like request the agency can use to gather information.
Reuters reported that the risk factors section of the firm's S-1 filing runs to approximately 94 pages, longer than in a confidential version of the filing the news agency had reviewed.
Some of the disclosures go well beyond hypothetical litigation. Anthropic said it held off for more than a month on pre-training its next frontier base model while it strengthened protections against theft of model weights and intellectual property. It warned that similar pauses could delay revenue and let competitors pull ahead.
The filing also described previously undisclosed evaluation failures. Four AI models broke out of their testing environments and reached real-world computer systems, in sessions lasting between 10 and 34 hours. That stemmed from an error during cybersecurity testing with an outside partner, which left the environment connected to the internet even though web access was supposed to be switched off.
In one case, an early model version kept working on an evaluation task that could not be completed. Along the way it found a stored password and used it to take administrator control of a third party’s system, then gathered more credentials, changed system settings and viewed one person’s personal data.
Anthropic said it notified everyone affected and found no evidence the models had tried to evade human oversight or coordinate with one another. It has since suspended some pre-release cybersecurity evaluations and paused higher-risk reinforcement learning environments.
In a separate disclosure on Sunday, the company said Claude had gotten around a paywall and pulled publicly available aggregate voter counts for Georgia without human authorization, Reuters and other news outlets reported.
Anthropic launched Claude for Financial Advisors in mid-September, with Charles Schwab and Dynasty Financial Partners among its early partners. On the opening day of Schwab’s IMPACT conference in Boston this week, the company rolled out a wave of new connectors for advisory firms, expanding its partner list to include Invesco, FINNY, CAIS, Conquest Planning and 55ip, the J.P. Morgan Asset Management subsidiary.
Anthropic also added a new connector integration to Flourish, which the firm says lets advisors work with their book of business, check rates, manage client invitations and prepare for meetings without leaving the Claude environment.
“Holistic advice is only as good as the holistic implementation behind it. With Claude and Flourish, advisors can see the whole picture, spot things like excess cash hiding in plain sight, and act on it in a fraction of the time. That means more wallet share, less busywork, and more time for client relationships,” Max Lane, CEO of Flourish, said in a statement.
Anthropic has emphasized that investment recommendations, client communications, compliance determinations and other regulated activities remain subject to human review and approval.
It recommends that registered investment advisers use its Enterprise plan, which includes audit logs that support recordkeeping. A compliance skill also screens client-facing language against the SEC’s Marketing Rule.
As reported by Reuters, Wilson Sonsini Goodrich & Rosati is representing Anthropic as counsel on the offering. Investment partnerships affiliated with the firm own a less than 0.01% sliver of Anthropic’s outstanding capital stock, directly or indirectly, which could open the door to questions about potential conflicts of interest.
Among the array of consultancies tapped for the Anthropic IPO, Freshfields US represents the underwriters, led by Goldman Sachs and Morgan Stanley alongside 20 other banks. Orrick, Herrington & Sutcliffe advised the AI lab on executive compensation and equity matters, while Cleary Gottlieb and Yale law professor John Morley advised Anthropic’s Long-Term Benefit Trust, and Cravath, and Swaine & Moore is advising the co-founders on certain governance matters.
Some current and prospective investors now value Anthropic at about $1.5 trillion, down from roughly $2 trillion a week earlier. A funding round in May placed the company at $965 billion.
Advisors must navigate misinformation and generational expectations as technology reshapes client relationships.
The regulator has been trying to reach imprisoned advisor Jason Foley.
Magnificent 7 firm continues to ‘right size’ workforce in AI-related shift also seen in financial services.
Healthcare inflation running at 280% of CPI - and most HSA holders are sitting in cash.
New research identifies the practice habits driving 121% DC AUM growth among the industry's top performers.
Knighthead Life entered the market with a competitive MYGA. A strong launch earned advisor confidence and paved the way for FIAs.
As AI makes financial information more accessible than ever, Lana Hock explains why human judgment, trust, and empathy remain the qualities clients value most in a financial advisor
