Anthropic IPO, Meta share offerings and GPT-6.1 Terra

Anthropic said Monday, Oct. 26, that it has filed a registration statement with the U.S. Securities and Exchange Commission (SEC) for an initial public offering of its shares. The same day, Meta launched two simultaneous stock offerings to bolster its cash as spending on AI infrastructure runs high, and OpenAI published the safety document for GPT-6.1 Terra, a new model in its GPT-6 series. All three are company announcements. OpenAI’s safety addendum gives no word on availability, regions or pricing, and none of the sources RaisLife (a Brazilian outlet) was able to open mentions Brazil.

Anthropic files for a U.S. IPO

According to Anthropic’s own announcement, the company submitted a Form S-1 to the SEC for a proposed initial public offering of Class A common stock. It plans to list the shares on the Nasdaq Global Select Market under the symbol ANT. The number of shares and the price range have not been set, the statement says, and it names Goldman Sachs and Morgan Stanley as the sources from which to request the preliminary prospectus once available.

The announcement cautions that the registration statement is not yet effective. Until it is, the shares cannot be sold and no offers to buy can be accepted, and the announcement is not an offer to sell. In practical terms, this is a formal step in the process, and no date for a stock market debut has been given. AI Updates, a daily AI news digest that RaisLife draws on, also cites an Associated Press report that RaisLife could not open, so this article relies on the company’s statement. Nothing here is investment advice.

Meta announces offerings tied to a plan to raise up to $95 billion

According to the company’s investor relations release, Meta launched two separate public offerings, neither contingent on the other. One is for 250 million depositary shares, each representing one-twentieth of a share of Series A mandatory convertible preferred stock, with an application to list them on Nasdaq under the symbol METAM. The other is for $22.5 billion of Class A common stock tied to forward sale agreements. In that case Meta does not receive the proceeds up front; the money comes in when the agreements are settled, no later than Nov. 19, 2027.

The release also mentions an at-the-market program, in which shares are sold gradually, of about $60 billion in Class C stock. No agreement has been signed for it, and it is expected to begin in the first quarter of 2027. The company says the proceeds will go to general corporate purposes, and Meta projects capital expenditures of $135 billion to $145 billion in 2026. Taken together, the two offerings (about $35 billion, according to the release) and the future program of about $60 billion add up to the $95 billion figure cited by AI Updates. AI Updates also cites a Financial Times report that RaisLife could not open. The terms of the preferred shares, such as the dividend rate, will be set only at pricing.

GPT-6.1 Terra: what the safety document says

OpenAI describes GPT-6.1 Terra, in an addendum to the GPT-6 Astra safety document published Oct. 26, as a fast, efficient model for high-volume professional work, with strong image-analysis and computer-use capabilities. The text says OpenAI is introducing the model but gives no dates, regions or access plans, and it lists no prices. The AI Updates summary says OpenAI cut its API prices; RaisLife could not confirm any figures or conditions.

The addendum contains evaluations that OpenAI ran on the model itself. The company rates it “critical” in cybersecurity capability and “high” in biology and chemistry, and says it uses the same safeguards as GPT-6.1 Sol and GPT-6 Astra. According to the document, the model beats GPT-5.6 Terra on HealthBench Professional (62.7 versus 57.7) but trails it by 1.9 points on the general HealthBench after length adjustment (a correction for response length), and also scores lower in dynamic mental-health simulations. The health figures are internal tests and do not indicate clinical use. The addendum also records a statistically significant regression in responses to images involving hate and extremism, which the company considers consistent with earlier releases. These are results measured by the maker, with no independent verification in the document.

Other highlights from the edition

The AI Updates edition also mentions three scientific research results presented by OpenAI, which according to the summary will go through academic discussion and verification, and a company dataset on professional use of ChatGPT in nearly 100 countries. RaisLife could not open the OpenAI pages cited, so it limits itself to noting that the announcements exist, as reported by AI Updates.

The summary of the edition is in AI Updates for Oct. 27 on RaisLife, and the full edition, with its list of sources, is in the original AI Updates edition. See also the technology page.

Sources

• Anthropic: announcement on the Form S-1 filing
• Associated Press: Anthropic IPO filing (not opened by RaisLife)
• Meta, Investor Relations: launch of the stock offerings
• Financial Times: Meta fundraising (not opened by RaisLife)
• OpenAI: GPT-6.1 Terra safety addendum
• OpenAI: GPT-6.1 Terra announcement (not opened by RaisLife)

Text by Rai, RaisLife’s virtual reporter (artificial intelligence), with automated editorial review. Data: technology on RaisLife.