Hey folks, Pratham here.

Welcome to Paradox Weekly, a Masters' Union newsletter, where we break down the ideas, trends, and contradictions shaping business today.

Each week, you’ll find analysis on what’s actually happening in markets, tech, and strategy.

This week: why your investor becomes your competitor.

A few weeks ago, a friend sent me a Bloomberg headline: “OpenAI Signs $300 Billion Deal With Oracle.”

And then 2 days ago another announcement came in:

What happened:

New Flexibility:

Back to my friend’s message.

He asked: "Wait, isn't OpenAI basically Microsoft's company?

That question reveals everything wrong about how we think about tech partnerships.

Five years ago, OpenAI was broke.

Microsoft wrote a $1 billion check and demanded exclusivity.

Today, OpenAI is signing the largest cloud contract in history with Microsoft's biggest rival.

If you’re thinking this is a partnership gone wrong... Nope, that’s not what happened.

In this case, the power dynamics shifted and nobody planned for it.

Had the honor of gifting our "Learn by Doing" yearbook to Shri Piyush Goyal, Hon'ble Minister of Commerce and Industry, this week. Going through it again made me realize how insane our students actually are, 45L+ revenue from dropshipping in 2.5 months, some on Shark Tank while still studying, pitching to real VCs for actual funding.

Still can't believe we've built something this amazing in just 5 years tbh.

The bit about asking "what evidence would change your mind?" in heated arguments got me thinking.

Trying to be less attached to my own opinions these days: question more, defend less. Sounds simple, but actually pretty hard to do.

Reminded me of when he was teaching me at Upenn - one of the few classes where I had 100% attendance.

July 2019. OpenAI had a problem: they were running out of money. Building AI is expensive. Really expensive.

Microsoft showed up with $1 billion. But this wasn't charity.

The terms were simple: Microsoft gives money, OpenAI makes Azure their exclusive cloud provider. Forever.

At the time, it made sense. OpenAI got capital to survive. Microsoft got the hottest AI startup as their exclusive partner. Win-win.

After, Microsoft quietly invested more. Their total commitment reached nearly $13 billion across multiple funding rounds.

Nobody knew what would happen next.

ChatGPT launched on November 30, 2022.

One million users in five days. Instagram took 2.5 months. Spotify took five months. ChatGPT did it in five days.

By January 2023, just two months after launch, ChatGPT had 100 million monthly active users. UBS analysts wrote: "In 20 years following the internet space, we cannot recall a faster ramp in a consumer internet app."

Google declared a "code red" and rushed to build Bard.

Amazon, Meta, and Apple started throwing money at AI. 



Everyone wanted to partner with OpenAI.

But they couldn't. Microsoft had locked up exclusivity.

Microsoft's stock soared. AI services contributed 12 percentage points to a 33% increase in Azure cloud revenue. Satya Nadella told investors: "Every day that ChatGPT succeeds is a fantastic day for Microsoft."

The deal looked genius. Microsoft would get up to 49% of all profits from the fastest-growing consumer app in history.

But success creates new problems.

By 2024, OpenAI was trapped in what venture capitalists call the "golden cage problem."

They were wildly successful. Growing faster than anyone predicted. Valued at over $80 billion. But every single compute cycle ran through Microsoft Azure. Every ChatGPT conversation, every API call, every new model trained on Microsoft's infrastructure.

The European Union started investigating. They asked tough questions about monopolistic behavior and competition. For context, the EU made Google pay $5 billion in fines. They don't mess around.

OpenAI needed to diversify. But Microsoft had exclusivity.

Something had to give.

On his first full day in office, President Trump stood in the White House with Sam Altman, Larry Ellison, and Masayoshi Son.

When I saw that number, I had to double-check. $500 billion. That's more than Belgium's GDP. That's like buying Twitter 100 times.

Here's what made it brilliant: OpenAI told Microsoft, "You want exclusivity? Fine. But we're building our own infrastructure empire with your biggest cloud competitor. And the President just announced it."

Microsoft's response was telling: "We are thrilled to continue our strategic partnership with OpenAI and to partner on Stargate."

That language screams defeat. You don't say you're "thrilled" about your partner building with your rival unless you've already lost.

Having Trump announce your $500 billion project gives you political cover even Microsoft can't fight.

On his first full day in office, President Trump stood in the White House with Sam Altman, Larry Ellison, and Masayoshi Son.

That's $60 billion per year. For context, OpenAI's entire 2024 revenue was around $3.4 billion.

This wasn't just diversifying cloud providers. This was OpenAI declaring independence.

Reports emerged that the Microsoft-OpenAI partnership was restructuring. Revenue sharing would drop from 49% to just 8% by 2030. OpenAI gained rights to work with other cloud providers.

Microsoft went from owning nearly half of OpenAI's profits to getting 8%. An 84% reduction.

Microsoft and OpenAI are now simultaneously:

This isn't a normal business relationship. This is what happens when exclusive partnerships meet explosive growth.

Other tech giants watched and learned.

Google invested $2 billion in Anthropic without demanding exclusivity.

Amazon put $4 billion into Anthropic, also without exclusive terms.

They realized that exclusive partnerships in AI create more problems than they solve.

The OpenAI-Microsoft saga proves something important: exclusive partnerships are rocket fuel for early-stage companies but become constraints as they grow.

OpenAI needed Microsoft's $14 billion and Azure infrastructure to build ChatGPT. Without that partnership, ChatGPT probably doesn't exist.

But once OpenAI became successful, the exclusivity that helped them grow started choking them. They needed freedom to scale beyond what any single partner could provide.

Microsoft could have fought harder. They could have used legal pressure, withheld funding, played hardball. Instead, they adapted.

Why?

Because they realized something smart: it's better to own 8% of a thriving, independent OpenAI than 49% of a company that resents you.

Translation: we fought, we adapted, we're both better for it.

The Microsoft-OpenAI partnership that defined the AI boom is being rewritten in real-time.

Five years ago, OpenAI needed Microsoft to survive. Today, they need each other to thrive. That's not a failure of the original partnership. That's evolution.

The smart partnerships don't try to lock in forever. They build in flexibility to evolve as power dynamics shift.

Microsoft gets a more sustainable partnership. OpenAI gets freedom to scale. Both can focus on what they do best instead of fighting over control.

The companies that figure this out will dominate AI. The ones that don't will end up in court, burning billions on lawyers instead of innovation.

Here's my question for you: In your partnerships, are you building in the flexibility to evolve, or are you setting up future conflicts by trying to lock everything down today?

I read every email.

Until next week,
Pratham