Support Us & Upgrade to the Operator Edition

Monday opened with Oracle asking Wall Street for another $20 billion, and by lunch the stock was down 6%. Meanwhile, two tiny startups joined forces over a problem most people don’t know AI has: water. And South Korea put a record budget behind chips.

This week, we’ll deep dive into who actually gets paid when the AI buildout starts to wobble.

Most founders have a great video sitting on their phone that they’ll never edit. OpusClip 4.0, out today, now edits the whole thing for you, from raw take to every format you post. Here’s how to run it:

💡 Pro tip: If you live in ChatGPT or Claude, use the AI Producer plugin to edit from there. Show it a creator style you like, and it’ll edit your footage to match. Opus is also offering 40% off your first month of Pro. Watch the 90-second demo here.

Want the version of this that actually runs in your business?

The 30-Second Play gives you the move. The Operator Brief gives you the full system: prompt libraries, API cost math, and the workflows our subscribers are using to replace entire freelance line items.

This week: how to win customers using AI agents while shopping.

👉 Upgrade to The Operator Brief

🍎 Apple’s new CEO John Ternus says the company won’t ship AI for AI’s sake and brushed off the idea that agents could make apps obsolete. It’s a bold stance in today’s world, yet it’s predictable for the company that takes a cut of every app sale. Apple not wanting apps to become obsolete is a little like a toll booth not wanting the road replaced.

🧠 Israeli startup 42AI came out of stealth with $62.5M and a “behavioral world model” that it says predicts how traders will act, and it’s already inside 10 brokerages. So if your broker’s app has ever seemed to know you a little too well, this may be why.

🇰🇷 South Korea is asking parliament for a record budget in 2027 with 21.3 trillion won for chips, physical AI, and data centers. That’s nearly double its budget for 2026, plus 4.7 trillion won for its own sovereign AI model. While the AI bubble has burst in China, that slowdown hasn’t seemed to halt the South Korean budget, at least.

🛒 JPMorgan, Amazon, and Schwab are blocking personal agents like Meta’s Muse, even though most customers say they’d welcome one, and retail experts think the walls won’t hold.

Oracle filed today to raise $20 billion in new stock and bonds to keep its AI buildout funded. Last week at its investor day, it promised $20 billion would be the ceiling on new AI financing all the way through fiscal 2030.

What’s in the filing:

Now zoom out. Oracle carried $125 billion in notes and borrowings as of August, against $67 billion in fiscal 2026 revenue. It’s targeting $225 billion in revenue by 2030. Its five-year credit default swaps, which are what investors pay to insure against Oracle defaulting, are up 270% in a year. And unlike Microsoft or Amazon, a big chunk of its AI story runs through one customer: OpenAI.

Here’s the asterisk: Baird’s analyst said this is exactly what Oracle previewed. Technically, the company is on message. But think about what it actually did: it set a limit and maxed it out in the same breath. This isn’t a first, either. Earlier this year, Oracle raised $25 billion in debt and said it didn’t expect to issue more bonds in 2026. Today it filed for $10 billion more.

🔮 Prediction: Oracle won’t break the $20B cap. It’ll route around it.

There are three ways the next three years can go. One: revenue ramps on schedule, capex peaks this year and next like management says, cash flow catches up, and the stock re-rates. Two: data centers come online late (construction snags and local opposition are already slowing projects), and Oracle needs more money than the cap allows. Three: it openly breaks the promise.

Path three is the least likely, because companies this size almost never publicly blow through a number in year one. Path one requires everything to go right with a customer that’s still losing money. So my bet is path two, solved quietly: joint ventures, lease structures, vendor financing, and private credit that fund the buildings without counting as “Oracle raising money.” Moody’s says private credit is holding firm despite the AI debt wave, which means that door is open.

Why does that matter to you? Because Oracle is the AI boom’s most leveraged believer. It doesn’t have an ad business or a retail empire to fall back on. When its cost of borrowing rises, that’s the earliest signal that markets are pricing in risk on the whole buildout. Watch the CDS spread, and watch for any data center announced: “with partners.” That’s the tell.

Is Oracle the AI boom’s canary or just its boldest bet? Hit reply. I read every one.

Read the full story here

Omen AI is acquiring rival Aquatic Labs in a $16 million cash-and-stock deal to build a real-time sensor for the water that cools AI chips. US data centers already use 17 to 19 billion gallons a year, and that’s expected to triple in the next five years.

What you need to know:

The pitch is uptime. When a cooling system goes bad, operators usually dump the water and shut down rather than risk frying millions in GPUs. Real-time monitoring means they can run the same fluid longer and skip the shutdown. Omen raised a $31 million Series A in June, and Aquatic’s CEO Allan Adams, a former MIT professor and Woods Hole ocean-sensing researcher, becomes chief science officer.

The founder’s take on the AI spending jitters is the line of the week. Omen works on data centers that were financed a year ago, so sentiment doesn’t touch it. Even if the market got cut in half, he says, it’s still “the biggest market in human history.”

Here’s the uncomfortable part: that’s only half true. Omen is insulated this year because 2027’s installs are already paid for. But the deals for 2028 and 2029 get financed right now, by companies like Oracle, in exactly the market that just sent its stock down 6%. The picks-and-shovels play is safer than the gold miners. It’s not immune to them.

🔮 Prediction: Water monitoring goes from “nice efficiency tool” to “required to get a permit” within two years.

Here’s the reasoning. Water is fast becoming AI’s political weak spot. Communities are fighting new data centers, and a judge just let the NAACP’s pollution lawsuit against xAI’s Memphis facility move forward. When local opposition is the bottleneck, the operator who can show live, verifiable water data has a better argument at the zoning board than the one who says “trust us.”

That also explains the deal itself. A $16 million merger between two sensor startups is what an early category looks like before it consolidates. Once the sensor is proven at gigawatt scale, I’d expect a big infrastructure player like Schneider, Vertiv, or Ecolab to buy their way in rather than build. Startups that measure things nobody’s measuring tend to become acquisitions the moment those things become regulated.

The bigger lesson for builders: the most durable money in AI right now might be in keeping data centers running, not building them.

Would you be fine living next to a data center if it published real-time water data? Reply yes or no. I’m curious where you land.

Read more here

Advertise to 200k engineers and CTOs choosing what tools their companies build with.

👉️ Learn how to sponsor this newsletter

Would love to hear your thoughts! Send me your thoughts by replying to this email (yes, I read them all :)

Until our next AI rendezvous,

Anthony | Founder of Uncover AI