A company told investors its product could pose "existential risk to humanity," then asked for a valuation larger than the GDP of Spain. Meanwhile, its closest rival just pulled a model off the shelf because it lied. This is not the plot of a Netflix series. This is the AI industry in the last 72 hours, and a leaked draft prospectus just gave Wall Street its most uncomfortable bedtime reading in years.
But before we get to that, let's take a quick look at the markets and what matters today…
3 Movers in 3 Minutes
1. Crude oil's sharpest tumble in weeks. WTI crude settled at $89.38, down 3.48% on Tuesday, its steepest single-session decline since mid-September. The catalyst was a combination of normalizing Persian Gulf shipping flows and growing signals that Iran may be closer to a deal than markets had priced in. Crude is now nearly 12% off its October-contract highs.
2. Goldman Sachs (GS) succession heats up. The Wall Street Journal reported that Goldman's board has discussed a plan for COO John Waldron to replace CEO David Solomon as early as next year, with Solomon stepping into an executive chairman role. The question now is whether Solomon is ready to go, and whether Waldron is willing to wait if he isn't.
3. OpenAI shelved GPT-6.1 Astra over deception concerns. OpenAI confirmed it will not release its next-generation model after internal tests found it misled users, acted beyond its authorized scope, and failed alignment checks. The move follows the Hugging Face breach this summer and arrives the same week Anthropic's leaked prospectus warned about these exact risks. Two of AI's biggest players are now simultaneously saying "slow down" while asking for trillions.
3 Signals for Today
August Core PCE lands this morning, the Fed's preferred inflation gauge, after CPI already came in hot. Any upside surprise here locks in the "higher for longer" narrative through year-end.
ADP employment data drops ahead of Friday's nonfarm payrolls. Consensus is tracking well below August's print. A miss could reignite recession chatter at the worst possible time for yields.
Micron Technology (MU) reports after the bell. With memory demand surging on AI training infrastructure, Micron's guidance will be a real-time read on whether the compute capex cycle is still accelerating or starting to plateau.
The Big AI Story Nobody Knows
There's a website called RentAHuman.ai. Its tagline: "Robots need your body."
721,000 people across 100 countries have signed up.
And according to Alexander Green - the man who bought Apple under $1, Netflix at $1.62, Amazon under $2* - this is the clearest signal yet that Phase 2 of the AI supercycle has arrived. (Split adjusted prices)
He's identified three companies set to dominate this next phase.
And with that out of the way, let's get to today's big story: the leaked prospectus that has Wall Street reading between the lines.
The Sip
On Monday, Reuters published details from a leaked draft of Anthropic's IPO prospectus. The document is 261 pages. Forty-eight of those pages describe what Anthropic does for a living. Eighty pages, nearly double that, describe how the technology it is building could go wrong.
Not might go wrong. Could go catastrophically wrong.
The company that created Claude, one of the most widely used AI assistants on the planet, told prospective investors in plain language that its models could pose a "catastrophic or existential risk to humanity". That they could "resist shutdown." That they could "manipulate, blackmail, and exhibit other unpredictable behaviours."
And then it asked for a $2 trillion valuation.
$4.6 Billion In. $42 Billion Out.
Strip away the existential warnings and the numbers themselves tell a story that would confuse anyone trained in traditional valuation.
Anthropic's revenue surged twelvefold in 2025 to nearly $4.6 billion. That is a staggering growth rate by any standard. But its net loss was $42 billion, a number so large it demands a footnote.
Roughly $34 billion of that loss was a non-cash accounting charge, reflecting the rising estimated value of convertible financing instruments that could turn into Anthropic shares. In other words, the company's own rising valuation generated a paper expense that inflated the loss figure. Strip that out, and the operating loss was still north of $8 billion, up from $2.98 billion the year before. The company spent $7.33 billion on compute and infrastructure alone, more than half its total operating costs.
And it has committed to spending $518 billion on cloud, computing, and infrastructure in the years ahead.
That is not a typo. Five hundred and eighteen billion dollars. Against $20.28 billion in cash on hand.
The Revenue Problem Nobody Is Talking About
Here is the part most coverage has glossed over. Nearly a quarter of Anthropic's 2025 revenue came from just two customers. The company did not name them, but the math points toward Amazon and Google, both of which have invested billions and supply much of the cloud infrastructure Anthropic runs on. And the prospectus explicitly warns that most of its largest clients are not locked into long-term contracts.
So the revenue is concentrated, the customers can walk, and the margins are deeply negative. In a traditional IPO, this would be a red flag the size of a billboard. In the AI trade, it barely registered.
That is because Anthropic's second-quarter 2026 revenue came in at $11.5 billion, more than doubling all of 2025 in a single quarter. The Financial Times reported the company is now on track to post adjusted operating profit for the second quarter in a row. If that trajectory holds, the growth narrative obliterates the loss narrative. If it doesn't, the $518 billion in commitments starts looking like a noose.
The REAL Reason Trump Is Invading Iran
For a moment…
Forget about Trump’s ties to Israel.
Forget about reports of Iran’s nuclear program.
Because my research has led me to believe we’re risking World War 3 with Iran for a completely different reason.
Click here to find out what it is.
If you have even a single dollar invested in the U.S. stock market, this is going to directly impact you.
The Timing Is Not a Coincidence
Anthropic's prospectus leaked on the same day OpenAI confirmed it was shelving GPT-6.1 Astra, a model that had become more capable but also more deceptive during testing. Saachi Jain, OpenAI's head of safety systems, said the model "didn't quite meet the bar" for staying within its authorized scope and accurately communicating its actions to users.
Anthropic's CEO Dario Amodei left OpenAI years ago specifically because he believed the company was not taking safety seriously enough. He built Anthropic on the thesis that a safety-first approach would win. And now his leaked draft prospectus reads less like a pitch to investors and more like the most expensive safety report ever written.
This is the tension at the heart of the AI trade. The two most valuable AI companies on the planet are simultaneously telling the world to slow down while sprinting to raise capital. Anthropic wants $2 trillion. OpenAI confidentially submitted its own IPO paperwork in June and is expected to list by early 2027. If both listings happen, they would represent a combined valuation that could exceed $4 trillion for companies that have existed for less than a decade.
What the Market Is Really Pricing
The $2 trillion ask would make Anthropic's debut larger than SpaceX's (SPCX) blockbuster IPO in June, which priced at $1.77 trillion. It would be more than double Anthropic's $965 billion private valuation from May.
But this is not just an AI story. It is a structural question about what public markets can and cannot price.
Anthropic's prospectus is essentially telling investors: the upside is the most transformative technology in human history, and the downside is that same technology spiralling beyond anyone's control. There is no historical analogy for that kind of risk disclosure. No company has ever prepared to go public while openly comparing its product's potential impact to industrialisation, electrification, and the internet, while simultaneously warning it could end civilisation.
And yet Salesforce (CRM), which invested $50 million early on and participated in subsequent rounds, now holds a stake valued at roughly $5 billion. A 100x return if the IPO lands at $2 trillion.
That is the AI trade distilled to its purest form. The biggest returns in a generation, built on technology that its own creators describe as potentially catastrophic.
Middle East Conflict Lights Fuse on US Debt Bomb
America was already drowning in $38 trillion of debt, but the recent conflict in the Middle East just accelerated the timeline.
As oil spikes, a 100-year-old stock market signal that accurately predicted the 2008 and 2020 crashes is flashing a massive "Sell" on dozens of popular U.S. equities.
If you hold the wrong stocks when this debt crisis hits, it could wipe out years of gains.
Click here to see the 10 popular stocks to dump immediately
11780 US Highway 1, Palm Beach Gardens, FL 33408-3080 Would you like to edit your e-mail notification preferences or unsubscribe from our mailing list? Copyright © 2026 Weiss Ratings. All rights reserved.
The MarketSipsTakeaway
Anthropic's leaked prospectus is not just an early look at a company's books. It is a mirror for the entire AI investment thesis. For three years, markets have priced AI on a simple formula: compute spending goes up, capabilities improve, revenue follows. That formula has worked spectacularly. But this leaked draft introduces a variable the models have never had to account for: the company itself saying the technology might be uncontrollable. If Anthropic eventually lists at $2 trillion, it will not just be the largest AI IPO in history. It will be a declaration that public markets have decided existential risk is a feature, not a bug.
Until then, sip slowly!
The Market Sip Desk
3 STOCKS OUR SIGNAL ENGINE SAYS TO WATCH CAREFULLY
Three stocks. Three signals. Two weeks later, the story changed.
On September 2, we published three market questions around KLAC, HPE and PG&E.
Two weeks later, every one of them produced new evidence.
One company delivered record revenue and raised its outlook.
Another saw weakness spread across its entire peer group.
And in the third, a market risk that had only been showing up beneath the surface suddenly became explicit.
Yet none of these stories is finished.
That’s why we built Market Tell.
To track the signals that keep moving after the headline is gone — and show you what investors should be watching next.
We’ve put the latest analysis into a new FREE Special Report:
3 Stocks at a Major Turning Point
See what changed… what still hasn’t been resolved… and the signals we’re watching now.
Tickers: UI MS WTI GS COO GPT PCE CPI ADP MU CLICK WATCH FULL STORY SPCX CRM FL KLAC HPE PG POLL