Wall Street Just Put a $2 Trillion Price Tag on One AI Company

Anthropic's revenue jumped seven times over in ten weeks, and it's headed toward a $2 trillion IPO. Nvidia just backed $105 billion for one Ohio building that won't turn on until 2028. If the AI money moves feel bigger and stranger every day, that's because they are. Today's issue is the numbers, the fine print, and what any of it means for a business that isn't printing money like these guys.

When you need more leads, you want to do more, right?

But what do you do if your leads are on LinkedIn? LinkedIn outreach still depends on the limited number of profiles already on your team.

Aimfox Avatars lets you rent dedicated, customizable LinkedIn profiles that start conversations and hand interested replies to your existing sellers.

Add outreach capacity without hiring another SDR or turning every employee account into a prospecting channel..

Book more calls with Aimfox Avatars

What happened: Anthropic's annualized revenue run rate topped $65 billion at the end of July, up from about $9 billion at the end of last year, according to TechCrunch. The company, which makes Claude, has filed confidential paperwork for an IPO expected as soon as this fall, and investors reportedly want it valued at $2 trillion or more, which would make it one of the largest public debuts ever. Anthropic was last valued at $965 billion in May.

Why it matters to you: This is what "AI bubble or not" actually turns on. When a company's revenue multiplies seven times in under a year, either the demand is real and about to reshape how every business buys AI tools, or the market is pricing in growth that can't keep pace. Either way, prices and features on tools built with Claude are going to keep shifting fast.

What to do about it: Don't lock into a long AI vendor contract right now. Month to month gives you room to move if pricing changes once the IPO dust settles.

What happened: Nvidia guaranteed up to $105 billion in lease and power obligations for a massive OpenAI data center being built in Ohio, plus a direct $1.5 billion investment in the project's developer, SB Energy, according to Axios. The campus will eventually deliver 8 gigawatts of computing power on land that used to be a uranium enrichment site, and it won't reach initial capacity until 2028.

Why it matters to you: Nvidia isn't just selling chips anymore. It's guaranteeing the debt behind the buildings that house them, which means if OpenAI's growth ever slows down, Nvidia is on the hook too. That kind of circular financing, one company financing its own customer's growth, is exactly the pattern that worries people who study bubbles.

What happened: OpenAI rolled out ChatGPT for Teens on August 18 for users ages 13 to 17, according to Fortune and NBC News. It blocks conversations about suicide, self-harm, and sexual content, skips pretending to have feelings, and leans on a "study mode" that walks kids through homework instead of just answering it. OpenAI says it estimates a user's age from how they write and talk, not from an ID, and automatically switches suspected minors into the teen version.

Why it matters to you: If anyone under 18 in your house or business, a teenage employee, your own kids, uses ChatGPT, they may already be in this mode without you doing anything. Parental controls exist, quiet hours and safety alerts, but only if both the parent and teen opt in.

What to do about it: If you have a teenager in your life who uses ChatGPT, ask them to help you set up the parental link. Age detection by how someone types is new and imperfect, so don't assume it's already protecting them.

What happened: OpenAI said this week it's rewriting its safety rulebook, called the Preparedness Framework, and pausing training on Astra, an unreleased model, after the system approached what the company defines as a "critical" level of cyber capability, according to Axios. The rulebook hadn't been updated since 2023. OpenAI is now adding more monitoring earlier in development and putting more compute toward understanding how its own models reason and act.

Why it matters to you: A company slowing itself down, on its own, before regulators forced the issue, is unusual enough to notice. It also confirms something worth sitting with: the same AI skills that write your code or answer your customers can, in the wrong setup, find and use security holes.

What to do about it: Keep your business software patched and don't skip updates. The gap between "AI can help defend your systems" and "AI can help attack them" is thinner than most vendors let on.

What happened: Google won a bankruptcy auction for a chunk of Spirit Airlines' internal data, paying $10 million for about 100 million emails, 500 million Microsoft Teams chats, financial records, HR files, and presentations, according to Axios and 9to5Google. Spirit shut down earlier this year owing $8.1 billion. The data has been stripped of passenger and frequent flyer details, and a federal judge still has to approve the sale.

Why it matters to you: If your business ever closes, gets acquired, or goes through bankruptcy, your internal emails and chat logs aren't just deleted. They can be sold as an asset, and AI companies are now bidding on that kind of data specifically. Mercor, an AI hiring startup, was the runner-up bidder at $7.5 million for the same trove.

What to do about it: Check what your business's data retention and deletion policy actually says, especially for anything stored with a third party like Microsoft or Google Workspace. Don't assume closing the company means the data goes with it.

What happened: Cursor, the AI coding tool SpaceX bought for $60 billion earlier this year, launched Origin, a built-in code hosting platform that syncs both ways with GitHub, on August 18, according to SiliconANGLE. By coincidence, GitHub suffered its seventh outage in two weeks that same day, a six-hour, forty-two minute incident with error rates near 20% worldwide, according to Cryptonomist.

Why it matters to you: This one's for the business owners who have a developer or dev team. If your projects live entirely on GitHub, that outage was a reminder you don't have a backup plan. Cursor is betting that reliability gap turns into customers.

What happened: xAI launched Grok Bot, an AI agent that gets its own cloud computer to log into websites, apps, and inboxes and complete multi-step tasks on its own, without needing an API connection, according to Unite.AI. It can research leads, update a CRM, or process invoices while you're away.

Why it matters to you: The catch is in the fine print. xAI's own documentation reportedly warns not to treat separate bots as a security boundary, because every bot on your account shares the same browser cookies, files, and login sessions, according to Tech Times. If one bot logs into your bank or your email, every other bot on the account can reach that same session, and researchers cited in the report found prompt injection attacks pulled off credential theft in most test runs against agents like this.

What to do about it: If you try Grok Bot or anything like it, start with a read-only task that touches nothing sensitive. Don't hand it your financial logins on day one.

Zoom out and today is one story wearing seven different outfits. Money is moving at a scale that stopped making intuitive sense a while ago: revenue that multiplies seven times in ten weeks, a single building worth more than most countries collect in taxes, a bankrupt airline's inbox worth more dead than the planes were. None of that is a reason to panic or swear off AI tools. It's a reason to stay a beat behind the hype, ask where your data actually goes, and remember that "critical mass" for these companies and "useful for your business" are two very different bars to clear.

If today's issue helped you make sense of the AI money chaos, forward it to one person who needs it. New here? Subscribe free and get tomorrow's issue too.

Talk tomorrow,
Mark Shilensky

Follow me personally on Social Media:
Facebook: https://www.facebook.com/MarkShilenskyPage
Instagram: https://www.instagram.com/markashilensky/
Threads: https://www.threads.com/@markashilensky
LinkedIn: https://www.linkedin.com/in/shilensky/
X: https://x.com/markshilensky
TikTok: https://www.tiktok.com/@mark.shilensky
BlueSky: https://bsky.app/profile/markshilensky.bsky.social