Content About Stocktwits
No Image Available

'As the AI World Turns.' Oracle, OpenAI & Leopold. ARD #132

Author
Michael Parekh AI: Reset to Zero
14 hours ago
19 MIN READ

Today’s theme is borrowed from evergreen daytime television. ‘As the World Turns’. Adapted today to ‘As the World of AI Turns’.

The original ‘soap opera’ ran on CBS for fifty four years. And the reason I am borrowing it today, is that it weaves through three separate AI stories on ‘Borrowing’, aka ‘Leverage’, in the world of Tech and AI.

Oracle and its 81 year old founder, Larry Ellison, through every major tech wave from the PC to AI.

OpenAI and Anthropic, the sibling rivalry that will not stop. Pepsi besting Coke for now.

And a twenty five year old AI wonder kid investor named Leopold, who echoes a lesson on leverage, also being painfully re-learned by millions of South Korean stock market investors. At the same time.

Big tales, very dramatic, plenty salacious, with a lot of soap opera elements. But there is a singular lesson running through all three, and that is the part I want to underline. Lessons I learned painfully in my twenties. Although not as dramatically.

It is leverage. And it is a lesson as old as human time.

This is the AI Tech Wave at the point where individual conviction, and individual borrowing, decide what happens next.

(1) Larry Ellison, 81, Betting Across Every Tech Wave

MP TAKE: There is a very detailed, terrific article in the New York Times about Larry Ellison and how he has made big bets across almost every technology wave. He is unique amongst the founder CEOs of technology. His peers were Bill Gates, Steve Jobs, and later on came Jeff Bezos and everybody else after that.

But Larry is unique because as one of the world’s richest men, he has hit the peak richest man in the world a couple of times. Once at the peak of the dot com wave, and once recently in this AI boom. That is a signal in itself.

He has essentially bet all he has, or a lot of what he has, around the tech wave of the moment. Basically convinced that he could make a big difference using his resources. Financial and otherwise, political as well, to make things happen.

His story is amazing. It will be turned into books and movies if it has not been already. There is a lot of interesting, salacious drama in there, worthy of dozens of soap operas. If you are into that sort of thing, please do read it. It is very well written.

But the core lessons here are around the current AI Tech wave. In terms of how he has financed it using his Oracle company’s resources. Using leverage to the hilt. Debt and equity, on and off balance sheet, SPVs and the like. Maybe three times, maybe five times his underlying equity, to fund the building of hundreds of billions of dollars worth of data centers for OpenAI and others.

And of course that can swing either way on leverage. That leverage is one of the core themes running through all three of the stories today.

Right now the jury is not in. We do not know whether this will work or not. Oracle’s debt is rated just short of junk, if not at it. And he has, again, many times leverage on his equity for building these data centers. We will see how it all turns out.

It is the same question plaguing all the bigger financed peers. Meta, Google, Amazon, Microsoft. All of which we talked about the last week and a half around their earnings season that just finished up for that stage.

But this is Oracle. A company that has been around for decades, the leader in databases around the world, the creator of one of the biggest fortunes in the personal computer wave. And now on its way to either making it or breaking it in the world of AI.

The other related part of the Ellison story is his son David. Who has taken the lessons from his father, and is also using his father’s balance sheet, leveraging a fair bit of it. An eighty billion dollar plus acquisition of the Warner Discovery studios, trying to create his own empire. Which has gone through all kinds of ups and downs in the politics of DC, and has also been delayed through lawsuits by several states, including California.

So, like father, like son. Leverage again. This thing runs in the family. And this is a drama bigger than Succession. That show that we all kind of liked.

Sources:

For longtime readers:

(2) OpenAI and Anthropic, and the Leverage That Is Not Money

MP TAKE: I have written a lot about how Anthropic came from behind. Essentially the Pepsi becoming the Coke.

It happened over the last year through its tremendous focus on just AI coding and enterprise coding. Taking its core Claude Code and Claude Cowork products into serious revenue growth.

OpenAI is not that far behind. But OpenAI was focused on a number of other things, including the consumer market, gadgets, smartphones and more. So less focus, more leverage across a lot of other priorities. And as a result they fell behind, and now they are trying to catch up.

There is a detailed article in the Wall Street Journal this time, on how that happened to Sam Altman, founder CEO, and his various senior management people. One of whom had to step back because of health reasons, unfortunately.

All of this at a time when both companies are racing toward their own trillion dollar plus IPOs. Anthropic this year. OpenAI potentially next year.

There is a lot of drama here. A lot of soap opera. A lot of Blips and Blip 2.0s and a lot of other ups and downs. Very salacious again.

The core theme here is that the leverage is less of money, and more on priorities and focus. Basically thinking one can do so many things at once. Leveraging one’s abilities, as it were.

And there is some ‘circular’ leverage of the other kind too. OpenAI has committed to hundreds of billions, potentially close to a trillion dollars worth of obligations for AI data centers, when their revenues are ramping in just the tens and hundreds of billions of dollars. That is equally spectacular, but far less than the debt or the obligations the company is taking on.

Sources:

For longtime readers:

(3) Leopold, and the Mirror Image in South Korea

MP TAKE: The third one is a young fellow, Leopold Aschenbrenner, who is now giving the world an extraordinary story of Schadenfreude in terms of his up and down.

He was fired from OpenAI, wrote an extraordinary treatise on AI and how it would change the world, 165 pages. It captivated some of the most accomplished young billionaires of our time, the Collison brothers of Stripe fame and so on. They basically funded him for hundreds of millions of dollars on a maiden hedge fund, which he grew to forty five billion dollars through very focused private and public bets in AI.

His fiancee at the time ended up being the chief of staff to founder CEO Dario Amodei. So he had lots of relationships deep in the network, as it were.

And his whole fund came crashing down before the destination weekend of his wedding in Carmel, CA.

A lot of drama. This would make an extraordinary set of scenes in a soap opera, filmed Succession style. He had to sell what looks like most of his public, and maybe part of his private portfolios, for cents on the dollar to Ken Griffin and his Citadel. One of the big billionaire financiers of our day.

Lots of ups and downs. Lots of salacious detail. The Journal and a bunch of other places have written all about it.

But the core mirror image of this story is the same story of leverage, and fear and greed, that I have talked about going on in South Korea.

There a nation of over fifty million people, over the last few weeks, egged on by their government approving single stock leveraged ETFs back in April and May, went the same way. One in five, one in six Koreans, basically young people mostly, bet their personal savings on two companies. SK hynix and Samsung, which make up half the Korean market. And took advantage of these one times, two times, three times levered ETFs.

That resulted last week, with the volatility in these stocks after they reported earnings, in both being down eight, nine, ten percent. And even though they had been up hundreds of percent prior, the leverage is what killed a lot of people in margin calls. The estimates are anywhere between one and a half to two million plus Korean citizens, out of fifty million, may have had margin calls and were essentially wiped out.

Which is kind of what happened to Leopold. Until he dusts himself off and goes back to making billions more of course.

One individual, a billionaire, versus millions of individuals who are not billionaires. All hit by the curse, the downside of leverage.

It works wonderfully on the way up. It works terribly on the way down.

We have all learned these lessons. I learned it in my twenties at Goldman Sachs, in the crash of eighty seven. Very painful lessons. Far less public drama.

Absolutely a hand on the stove when it is hot kind of lesson if taken to heart.

Sources:

For longtime readers:

MP OVERALL TAKE

Those three stories are the highlight of ‘As the AI World’Turns’, with a lot more stories and dramas to come.

And my takeaway on all of it is that as entertaining as they are, as dramatic as they are, as soap opera as these stories are, it is a fundamental question of leverage.

The importance of focus. The importance of discipline. The importance of all the things that our rational brain will always tell us to do when we are doing anything. Whether it is investments, or other things professional or personal.

And most of the time our animal brains can let us get carried away. Because of the pendulums of fear versus greed, and the other emotions in between.

So that is a key takeaway. It is a lesson worth learning and relearning.

It is there through millennia. It has nothing to do with AI. Nothing to do with technology. It is just being a human, and emotions.

Which is worth holding onto, because this AI Tech Wave is going to keep producing stories like these three. The technology is new. The leverage is not.

That is how the AI world turns.

GADGET AI

India Starts Paying for Its Apps

MP TAKE: Here I wanted to talk about the country of India. One point four plus billion people, the most populous country after China now, a lot of them young.

They have one of the biggest online populations. Over seven, eight hundred million people using smartphones to do everything, as we all do as well.

But the difference in India for the last decade or so is that their population had not been paying as much for the apps that they downloaded.

And that is turning. The research firms are capturing it now. Indians are actually paying for a lot more applications than they did.

This is an important thing, because US companies like Meta, Google and many others, Amazon included, have been investing aggressively in India. They have one of the most appealing markets of young people, in the hundreds of millions.

There is a behaviour change signal in the mainstream, that they are starting to pay for applications. There are a lot of statistics in the articles in the show notes below. TechCrunch had a lot of detail from a couple of other sources.

It is important because it means big things for opportunities in India, which has been one of the potentially most appealing technology markets, especially for AI services from a consumer point of view, because of their large young market.

It is dramatic too, because the other takeaway is that sometimes these things take a lot longer than we think. People thought India would turn to be more proactive in terms of paying for things a lot earlier. It took the better part of a decade.

And that is the other kind of dramatic lesson to be learned from this.

Sources:

For longtime readers:

Q1. What did I take away from the use of smartphones in India?

Answer: I have been to India a number of times over the last couple of years. And the core thing for me was how ubiquitous it is at every level of society in India. Urban, rural, everywhere. Everyone is on smartphones.

And the app of choice is actually WhatsApp, Meta’s core application. They do everything on it.

WhatsApp has essentially become the WeChat of India. WeChat being the Chinese super app owned by Tencent in China, where over seven, eight hundred million people use it for everything. Commerce, payments and the rest.

WhatsApp is taking on that role in India. And it is one of the crown jewels of Mark Zuckerberg’s global empire. He actually just recently appointed an Indian CEO to be the CEO of WhatsApp globally. I wrote about that separately.

They are going to be employing AI agents. So he has the ability to create essentially a super app in India with AI, and that is what all of this is about.

I was impressed by that WhatsApp take rate in India, as it were.

Q2. What is one thing that could be improved on?

Answer: The core thing that came across in India is that young people there are relatively more comfortable with English than almost any other big country around the world. And so most of the online usage is really centred around English. AI data is trained around English, and so on.

India of course has hundreds of languages, dozens of major cultures and religions and more.

So one of the big opportunities in India is AI, as data centers get built, to be trained on local languages. To go far deeper into the substrate of the cultural milieu of India, and leverage it for AI data that does not exist outside India.

That would be a big piece. It would be interesting to see a lot get invested into that area. Whether it is by the US companies, or the local indigenous native AI companies out of India.

WRAP

Today’s AI-RTZ #1167 is on the two ‘Forever Problems’. AI hallucinations and prompt injections. There was a report last week that Anthropic may have fixed prompt injections, which is a very geeky thing, but very important to understand in terms of how AI has a big vulnerability that is very difficult to fix.

I go into a lot of deep technical detail, as well as a high level overview, on why they may have fixed it for their specific application. But it is not at all fixed for the AI world at large. I still call these two things the Forever Problems of AI, and they are very different from anything we have had over the last fifty, sixty years of traditional computing. Just by the nature of what AI is. The probabilistic nature of AI, versus the deterministic nature of traditional computing.

AI Ramblings Daily on AI-RTZ is here to think through AI and reset. Together.

Tomorrow, ARD 133 and AI-RTZ #1168.

Thanks for joining us today, AI Curious Folk. Stay tuned.

— MP

Full Source Reading

For the broader context, see the canonical sources for ARD 132, in today’s narrative order:

Event 1. Larry Ellison and Oracle

Event 2. OpenAI and Anthropic

Event 3. Leopold and the Korea Mirror

Gadget AI. India Starts Paying for Its Apps

Clips from today

Clip 1. India Finally Starts Paying for Its Apps

But the difference in India for the last decade or so is that their population had not been paying as much for the apps that they downloaded. And that is turning.

MP Take: This is an important thing, because US companies like Meta, Google and Amazon have been investing aggressively in India. They have one of the most appealing markets of young people, in the hundreds of millions. A behaviour change signal in the mainstream is a big deal. And there is a second lesson in it about time. People thought India would turn to be more proactive in paying for things a lot earlier. It took the better part of a decade.

Clip 2. WhatsApp Is Becoming India’s WeChat

The app of choice is actually WhatsApp, Meta’s core application. They do everything on it.

MP Take: WhatsApp has essentially become the WeChat of India, the Chinese super app owned by Tencent, where over seven, eight hundred million people use it for everything including commerce and payments. It is one of the crown jewels of Mark Zuckerberg’s global empire, and he just recently appointed an Indian CEO to be the CEO of WhatsApp globally. They are going to be employing AI agents. So he has the ability to create essentially a super app in India with AI, and that is what all of this is about.

Clip 3. Leverage Works Wonderfully Up. Terribly Down.

One individual, a billionaire, versus millions of individuals who are not billionaires. All hit by the curse, the downside of leverage.

MP Take: It works wonderfully on the way up. It works terribly on the way down. We have all learned these lessons. I learned it in my twenties at Goldman Sachs, in the crash of eighty seven. Very painful lessons. Absolutely a hand on the stove when it is hot kind of thing. The arithmetic does not care whether you are a forty five billion dollar fund in Manhattan, or one of a million and a half retail accounts in Seoul.

Clip 4. How Anthropic’s Focus Beat OpenAI’s Breadth

Anthropic came from behind. Essentially the Pepsi becoming the Coke.

MP Take: It happened through its tremendous focus on just AI coding and enterprise coding, taking Claude Code and Claude Cowork into serious revenue growth. OpenAI was focused on a number of other things, including the consumer market, gadgets and smartphones. Less focus, more leverage across a lot of other priorities. Here the leverage is less of money, and more on priorities and focus. Basically thinking one can do so many things at once, leveraging one’s abilities as it were.

About AI Ramblings Daily (ARD), and AI-RTZ

Both are daily. Both are free. Both are about AI. But they’re different mediums carrying different messages.

AI-RTZ is the morning text, a deeper written take on one idea, published by at least 5 AM EST. Today: post #1167.

AI Ramblings Daily is the afternoon video + podcast, my ad hoc takes and perspective on the day’s AI issues and news flow, around 20 minutes, with short 1-2 minute clips for quick topic views. Today: episode #132.

Subscribe to either or both on michaelparekh.substack.com. They run as separate Sections you can opt into or out of.

(NOTE: The discussions here are for information purposes only, and not meant as investment advice at any time. Thanks for joining us here.)





Want the latest?

Sign up for Michael Parekh's Newsletter below:


Subscribe Here

More like this

No Image Available
Monday links: AI assumptions
Tadas Viskanta Abnormal Returns
Markets Weed stock came in with a lot of hype, but have been a disaster. (trendlabs.com) New all-time highs are not, in and of themselves, any kind of...
4 months ago
2 MIN READ
BROWSE ALL ARTICLES