Less dependence on Nvidia…
Building the world’s smartest AI used to be the finish line.
Today, it’s just the admission ticket.
Once you’ve built the model, you still need chips to run it, data centers to power it, cloud providers to host it, and—more recently—regulators willing to let you release it.
This week, OpenAI was reminded of all four.
Here is the story. ⇩
SPONSOR BREAK presented by MarketWise*
“Biggest Breakthrough in the History of Stock Trading”
A Maryland computer whiz recently created a new form of “Predictive AI” that can foresee the future prices of any of 2,334 stocks – to the penny – with 73% historical accuracy. It’s led to a huge anomaly that would’ve turned every $5,000 into over $15,000 in the 16 months following its creation in one study.
Claim your free demo here (no purchase required).
Jalapeño — built to beat Nvidia 🌶️
→ 9 months.
→ Better performance per watt, in early testing.
→ A direct swipe at the market leader.
OpenAI + Broadcom’s first custom AI chip
OpenAI and Broadcom announced Jalapeño on Wednesday — OpenAI’s first custom-designed chip, built specifically for inferencing (the process of running already-trained AI models).
That is a direct, unambiguous swipe at Nvidia, the company OpenAI buys the most chips from and competes hardest against for access to them.
Unlike Nvidia’s flagship GPUs, which are designed to do everything exceptionally well, Jalapeño is built for a narrower job.
If it performs as expected, OpenAI could lower inference costs, improve efficiency, and reduce its dependence on buying massive numbers of Nvidia chips.
It also joins a growing club.
→ Amazon has Trainium and Inferentia.
→ Google has TPUs.
→ Microsoft is developing Maia.
→ Meta builds its own AI chips.
Everyone is trying to own a little more of the AI stack.
Greg Brockman called Jalapeño part of OpenAI’s strategy to make compute “more abundant.”
Translation?
Less waiting in Nvidia’s checkout line.
SPONSOR BREAK presented by MarketWise*
The Billionaire Who Saved SpaceX Just Made a New Bet
He once rescued SpaceX from bankruptcy. He also helped launch Facebook, Airbnb, YouTube, and Spotify. And now he’s making a new bet – he’s just sold every single share of the Mag 7 companies in his portfolio. And he’s using that money to buy a shocking new kind of company instead. You should mirror his moves. And now you can, for as little as $50 a share.
Meanwhile… Nvidia
⚠️ Here’s the irony.
The same week OpenAI unveiled a chip designed to reduce its dependence on Nvidia…
…new market data suggested Nvidia’s grip on AI became even tighter.
Nvidia now controls roughly 74% of the AI inference market, up from about 66% a year ago.
That surprises many investors. Conventional wisdom suggested inference would become Nvidia’s weak spot.
Instead…
Nvidia kept winning there too.
→ It generated roughly $41 billion of AI inference revenue during the first quarter alone.
That’s more than Broadcom’s AI revenue and AMD’s data-center business combined over the same period.
SPONSOR BREAK presented by Brownstone*
Have you tried Elon Musk’s 70x AI agent?
I’m about to do a live demonstration.
Of Elon Musk’s latest genius invention.
It’s an AI agent…
Perhaps the most powerful ever created.
Elon himself believes it could 70x your money… in a short period of time.
The Numbers.
Growing fast and catching Nvidia are becoming two very different things.
Every time competitors identify an opening, Nvidia seems determined to close it.
→ Custom chips promised better efficiency?
Nvidia designed Vera Rubin specifically to improve inference economics.
→ Competitors focused on AI accelerators?
Nvidia expanded further into CPUs and complete AI systems.
Its strategy increasingly looks less like selling chips…
…and more like owning as much of the AI infrastructure stack as possible.
That doesn’t mean Jalapeño failed before it shipped.
Far from it.
Custom chips can still reduce costs, improve efficiency, and make OpenAI less dependent over time.
But building a better chip and replacing the industry’s dominant platform are two very different goals.
SPONSOR BREAK presented by Paradigm*
FREE Gold Ticker to Buy ASAP: (NYSE:___)
Jim Rickards – the world’s #1 gold expert – has just revealed one of his favorite gold plays… 100% FREE.
As Jim sees it, we’re witnessing the biggest gold boom of the last 100 years – and those who keep their money on the sidelines are missing their chance at a fortune thanks to gold’s epic run.
But smart investors who get in now, could make 10X their money in the coming months.
The Government
As if competing with Nvidia weren’t enough…
OpenAI also discovered this week that hardware isn’t its only dependency.
The company confirmed GPT-5.6 will first roll out to 20 government-approved partners before becoming broadly available.
The request came from the U.S. government.
Why?
Because the newest generation of AI models is becoming increasingly capable of autonomous coding, cybersecurity research, and advanced biological reasoning—areas with obvious national security implications.
OpenAI wasn’t blocked but slowed.
That’s an important distinction.
SPONSOR BREAK presented by MarketWise*
Elon Musk Warns We May Have Just Six Months Left
“Frontier AI” is a point of no return when AI surpasses human intelligence and gains free will. Elon Musk warns this moment could hit by the end of 2026.
According to 60-year Wall Street legend, Marc Chaikin, Frontier AI could soon become the only thing that determines which companies make money and which grind to a halt, That’s why he’s giving away a list of stocks to buy and sell absolutely FREE to help you position your money for a world driven by Frontier AI technology. Get Marc’s Frontier AI Hotlist right here…
It’s Not Just OpenAI
Anthropic’s situation is more severe.
Anthropic suspended its most capable models after being ordered to restrict access for certain foreign nationals, following reports that its Fable 5 model could be jailbroken to bypass cybersecurity safeguards.
OpenAI’s situation is less restrictive—a staggered rollout rather than a suspension—but the direction is the same.
OpenAI hopes an executive order signed earlier this month will bring more clarity. The order gives the administration and AI companies 60 days to develop a voluntary framework that could allow the government to review frontier models for up to 30 days before public release.
In other words, OpenAI isn’t pushing back against oversight—it wants predictable rules instead of last-minute negotiations.
The Bulit-In Delay
If a formal government review becomes standard across the industry, every major AI release could come with a built-in delay between building the technology and monetizing it. That could affect product launches, enterprise adoption, and ultimately, revenue.
A mandatory review window won’t stop innovation, but it could slow it down.
On the other hand, a standardized review process could also create something investors value: predictability. Clear rules help companies plan product launches, allocate capital, and commercialize new technology with greater confidence.
Don’t forget to cast your vote 👇
Lesson Of The Day:
Was this email forwarded to you? Don’t miss out on future stories — subscribe using the button below.
Also, help your friends blossom this spring! Share us with them.
💬 We Want To Hear Your Story:
Got a market or stock you want us to analyze next?
Just drop your request in the comments here.
P.S. – If you no longer want to receive occasional emails from us and you want to unsubscribe, click here 👉 “Unsubscribe” . Thank you!
