Anthropic filed its IPO prospectus, and nearly a third of it is risk factors. Per the Financial Times, the filing names specific behaviors Anthropic's own models have shown or could show, including attempts to resist shutdown, conceal or manipulate information, and conduct resembling blackmail. Reuters reports Anthropic lost more than $8 billion in 2025 while revenue jumped twelvefold to nearly $4.6 billion, and the company plans to spend $518 billion on compute in the years ahead. Backers reportedly want a valuation above $2 trillion, which would make this the biggest IPO ever, for a company that just told the SEC its product might be dangerous to humanity itself.

Meanwhile, AMD closed one of the year's biggest AI acquisitions today, and a viral AI assistant just tripled its valuation in a month.

• Anthropic's IPO filing names existential risk as a risk factor, alongside a $2 trillion valuation target.

• An $8 billion loss, $4.6 billion in revenue, and $518 billion in planned spending, all in one filing.

• AMD bought Fei-Fei Li's World Labs for $8.2 billion; Li becomes AMD's chief scientist.

• AMD says the deal will shape its chip roadmap as it chases Nvidia in AI-specific hardware.

• Instinct raised $1B at a $10B valuation, a month after its last round valued it at $2.5B.

• OpenAI scrapped a planned model release over safety concerns, the Wall Street Journal reports.

…and a whole lot more that you can read about below.

Anthropic's IPO prospectus devotes close to a third of its pages to risk factors, according to the Financial Times, which reviewed the filing. Among the disclosures: specific behaviors Anthropic says its models have already shown or could show, including attempts to resist shutdown, conceal or manipulate information, and conduct resembling blackmail. Reuters, first to report the financial details, says Anthropic posted an operating loss of more than $8 billion in 2025 as computing costs surged, while revenue jumped twelvefold to nearly $4.6 billion. Total operating expenses reached almost $13 billion. The prospectus also outlines plans to spend $518 billion on cloud and computing infrastructure in the coming years, on top of deals already signed with Google, SpaceX, and Nscale. The FT reports Anthropic's momentum has kept building into 2026, with second-quarter revenue alone reaching $11.5 billion and a second straight quarter of adjusted operating profit. Nearly a quarter of last year's revenue came from just two customers, whom the filing doesn't name. Backers reportedly believe the company could list above $2 trillion, more than double its $965 billion valuation from May, which would make it the biggest IPO on record. CEO Dario Amodei has spent the month publicly pushing labs to "pace the frontier" of AI development, telling the UN Security Council last week that AI is "the most important global security issue facing the world today." The filing lands one day after Amodei had his first one-on-one dinner with President Trump at the White House.

AMD is acquiring World Labs, the AI lab Fei-Fei Li founded in 2024 to build models with a grounded understanding of physical reality, for $8.2 billion. Li, a Stanford computer science professor and a pioneer of computer vision through the ImageNet database, will join AMD as executive vice president and chief scientist. In a post announcing the deal, Li wrote that scaling World Labs' breakthroughs "requires scaling our efforts, widening our reach, and getting closer to the hardware." World Labs' first product, Marble, generates entertainment experiences and simulated environments for training robots, work AMD says will shape its chip roadmap as it competes with Nvidia's existing suite of open-weight world models. AMD and World Labs formed an inference optimization and training partnership last year, and Li appeared at AMD's CES presentation earlier this year. World models are seen as central to training general-purpose robots, since there isn't enough real-world data to do it directly, the same problem companies like Tesla and Figure are racing to solve with synthetic data. The deal is expected to close by year-end, pending regulatory approval.

Solutions Engineer | Mach9
San Francisco, CA (Full-time, Full stack)
Work with customers on AI CAD software built for civil engineering.

Applications Engineer | Polymath Robotics
San Francisco, CA (Full-time, Robotics)
Build general autonomy software for industrial vehicles.

Software Engineer | Candid Health
San Francisco, CA / New York, NY / Denver, CO (Full-time, Full stack)
Build the autonomous revenue cycle management platform used by healthcare providers.

Full Stack Developer | Bountiful
San Francisco, CA (Full-time, Full stack)
Build the operating system for agriculture.

Senior Software Engineer | Beacons
San Francisco / New York, NY (Full-time, Full stack)
Build the AI-powered, all-in-one business platform for content creators.

Senior Mechanical Design Engineer | SiPhox Health
Burlington, MA (Full-time, Mechanical)
Design at-home health monitoring hardware built on silicon photonic chips.

Full Stack Product Engineer | Jiga
US / Remote (Full-time, Full stack)
Help companies source better parts by partnering directly with vetted manufacturers.

The International Quant Championship (IQC) is one of the world's largest quantitative research competitions, bringing together 156,000+ participants globally.

Participants have the opportunity to develop quantitative research skills, challenge themselves alongside peers from around the world and connect with a global community of quantitative thinkers.

Build skills on WorldQuant BRAIN®

Instinct Raises $1B Series C at a $10B Valuation

The viral AI assistant, invite-only since August, raised $1 billion from Sequoia Capital, Benchmark Capital, and Coatue, a month after a prior round valued it at $2.5 billion. Founder Noah Shinn said the funding helps bring Instinct "to more people" as it builds out the future of personal AI. The company now faces direct competition from Meta's Muse, which offers similar features plus deeper integration across Meta's own social apps.

Modulate Raises $25M for Voice Models and Analysis

The startup, which builds voice AI and analysis tools, closed a $25 million round to expand its product suite. Modulate's technology is used to detect toxic behavior and moderate voice chat in real time, a niche that's grown alongside the rise of voice-based social and gaming platforms. The round adds to a steady run of funding into companies building the audio layer of AI products.

Outmarket Raises $34.5M Just Months After Its Last Round

The insurtech startup closed a fresh $34.5 million round, following a quick return to the market after a prior raise earlier this year. The fast follow-on suggests investors are still willing to move quickly for insurtech companies showing real traction. Terms of the round, including its lead investor, weren't disclosed in the initial report.

Today's Growth Recommendation

Greenfield Robotics is giving farmers alternatives to herbicide-dependent weed control, with BOTONY as the first step in a broader robotic farming system. Their Reg A+ offering is now live for investors.

Invest in Robots Replacing Herbicides Below ↓

Greenfield Robotics is on a mission to give farmers alternatives to herbicide-dependent weed control. BOTONY is the beginning of a broader robotic farming system designed for a wider range of applications in the field.

The Reg A+ offering is now live for investors who want to be part of what comes next.

Invest in Robots Replacing Herbicides

^{This Reg A+ offering is made available through StartEngine Primary, LLC, member FINRA/SIPC. Please read the Offering Circular and related disclosures before investing. This investment is speculative, illiquid, and involves a high degree of risk, including the possible loss of your entire investment.}

Microsoft Copilot Lets You Pick a Different Brain

Buried inside Microsoft 365 Copilot is a setting most people never touch: which AI model actually answers you. Since earlier this year, Microsoft has quietly let Copilot users swap the default OpenAI model for Claude, Anthropic's model, on select features. The clearest example is Researcher, Copilot's deep-research agent that digs through your emails, files, chats, and the web to answer multi-step questions. Different models reason and write differently, so a flat Copilot answer might need a different model, not a better prompt.

Open Researcher inside Copilot Chat and start a query. Look for a model picker near the input box, sometimes labeled "Try Claude." Switch from the default model to Claude Opus and rerun the same question. If you don't see the option, ask IT to enable Anthropic models under Copilot settings in the Microsoft 365 admin center, since it's off by default in the EU and UK. Claude tends to write cleaner long-form analysis, while the default model is often faster for quick lookups, so it's worth testing both before deciding Copilot "isn't that smart."

The Model Comparison Prompt

Invite friends & get instant rewards

If you enjoy The Incredible Edge, share it to unlock free resources when they subscribe.