📈 Anthropic S-1, 🤖 Meta hacked and 💰 Alphabet raises $80B — The most expensive Monday in history

Anthropic files its S-1 to go public valued at $965 billion. Meta's chatbot gives away Instagram accounts to hackers — including Obama's White House account. Alphabet borrows $80 billion to build more AI. Three stories from the same Monday that together tell something big: AI isn't an industry — it's a minefield where those who don't invest are left out and those who invest badly get burned.

📈 Anthropic files its S-1: the world's most valuable AI company prepares to go public

On June 1, Anthropic confidentially filed its registration statement (Form S-1) with the SEC. The company behind Claude, which just closed $65B and is valued at $965 billion, begins the process to go public amid the triple IPO of AI giants. The number of shares and price range have not yet been set, and the company warns the offering will depend on market conditions.

What makes this announcement historic isn't just Anthropic. It's the context: Elon Musk has said SpaceX will start its IPO roadshow this week with a target valuation of $1.75 trillion. And OpenAI is preparing its own exit, according to Reuters, with an estimated valuation of up to $1 trillion. Three of the most important AI companies — each competing to be first to public markets — in the same months.

Wedbush analysts sum it up well: "This represents the opening of the floodgates for the IPO market, which has been relatively dormant for years." That's not an exaggeration — if all three go public, we'd be looking at the largest concentration of initial public offerings in tech history.

Anthropic arrives with numbers that justify the valuation: an estimated annualized run-rate of ~$44B, revenue that grew 130% from Q1 to Q2 2026, and its first quarterly operating profit of ~$559 million. The question is no longer whether Anthropic is worth nearly a trillion — it's how much it will be worth in the public market when it prices the "ethical AI company" premium.

From money to security: while Anthropic prepares its IPO, Meta's chatbot showed that poorly designed software can be costly.

🤖 Hackers trick Meta's chatbot into stealing Instagram accounts — even Obama's White House account

This weekend, several users reported massive Instagram hacks. And the attack vector wasn't social engineering against humans — it was directly against the Meta AI Support Assistant, the company's AI-powered support chatbot.

The technique is as simple as it is terrifying. The hacker used a VPN to simulate the victim's location and bypass Instagram's automatic protections. Then they opened a chat with Meta's AI assistant and asked it to add a new email to the target account. The bot sent the verification code to that email, the hacker shared it with the chatbot, and the assistant — trusting the information it had just generated itself — showed a "Reset Password" button. One click later, the account belonged to the attacker.

Among the compromised accounts — confirmed by Meta on Monday — are the Instagram profile of Obama's White House account (inactive since 2017) and that of US Space Force Senior Enlisted Advisor John Bentivegna. Security researcher Jane Wong also lost access to her account.

Meta has confirmed they "have resolved the issue" and are working to recover affected accounts. But the damage is done — not as a technical breach, but as a proof of concept that support chatbots with too much power are a real security risk.

And to close the circle of money: Monday also showed that not even the giants can fund the AI infrastructure arms race alone.

💰 Alphabet raises $80 billion in shares to fund its AI spending

And to close the money circle: on Monday, Google's parent company announced a plan to raise $80 billion through a combination of public stock offerings, a long-term sale program, and a $10 billion private investment from Berkshire Hathaway, Warren Buffett's holding company.

The goal is to fund AI infrastructure expansion, which the company itself describes as something that "exceeds available supply." Alphabet had already raised its annual capital expenditure forecast to between $180B and $190B in April. This move — one of the largest stock sales in history — confirms that not even Google, with its multi-billion-dollar cash reserves, can fund the AI infrastructure race alone.

That Berkshire Hathaway — the most conservative fund on the planet — is putting $10B into the deal is a signal that shouldn't be overlooked. Warren Buffett doesn't invest in bubbles. If he's putting money into Google's AI expansion, it's because he believes the return will be real and massive.

And the context is even more revealing: in the last 48 hours, Anthropic announced its S-1, SpaceX its IPO roadshow, Alphabet its mega-capital increase, and Meta its AI security crisis. All on the same Monday. Capital isn't going to AI — it's going at full speed.

Anthropic files its S-1, Meta's chatbot gives away Instagram accounts, and Alphabet asks for $80B for infrastructure. The same Monday, three stories that paint AI as an industry of extremes: nearly-trillion-dollar valuations, embarrassing security failures, and capital expenditures only giants can afford. AI isn't an industry — it's a minefield where those who don't invest are left out and those who invest badly get burned.

— Max