Alex Vacca Profile picture
Sep 9, 2025 21 tweets 6 min read Read on X
Everyone's freaking out about Microsoft's deal with Nebius for $19.4 billion.

Two years ago, the same company was sanctioned and delisted from Nasdaq.

The founder fled from Russia with 1,300 engineers after condemning Putin's war.

Here's the wild story:
Microsoft's deal sent Nebius from $64 to $90 in hours.

$19.4 billion through 2031. That's 13x what Nebius made in all of 2024.

Microsoft had no choice though. They'd just lost their main GPU supplier to OpenAI... Image
But before we get to Microsoft's mess, you need to first meet Arkady Volozh, Yandex founder turned Nebius' CEO.

1989, working at a Soviet pipeline institute, he starts building search algorithms. Launches Yandex in 1997.

By 2021 he'd built something that made Google nervous... Image
Yandex - what would become Nebius - owned Russia completely. 72% of all searches. 18,000 employees. $30 billion peak.

Even Putin visited once in 2017.

But February 24, 2022 changed everything... Image
Russia invaded Ukraine and Yandex crashed 80% in a single day.

Trading halted at $18.94 and stayed frozen for almost three years. Twenty billion dollars of value had vanished.

But that was just the start of Volozh's nightmare... Image
June 3, 2022. The EU personally sanctions Arkady Volozh.

Think about this - you spend 25 years building a company, then overnight you're forced to resign, your bank accounts are frozen & you can't even email your own employees.

He flees to Israel and goes silent for 18 months.
While everyone thought Volozh had disappeared, he was secretly extracting 1,300 Yandex engineers from Russia.

Using intermediaries because he couldn't contact them directly & moving entire families out.

Then Volozh broke his silence in the most dangerous way possible... Image
In August, Volozh publicly condemned Putin's war.

"I am horrified that bombs fly into Ukrainian homes daily."
Most Russian billionaires who cross Putin fall out windows.

But Volozh had a plan... Image
He convinces Putin to split Yandex completely.

Russia keeps domestic ops. Volozh gets international AI assets, $2.4B cash, and those 1,300 engineers.

Price: $5.4B (forced 50% discount).
What he built next caught Microsoft's attention. Image
March 2024, the EU drops sanctions on Volozh.

His new company Nebius starts with a simple observation: everyone's trying to run AI on infrastructure built for websites and databases.

That's like racing a Formula 1 car on bicycle wheels.

And Nebius had unique advantage. Image
Remember those 1,300 engineers?
They'd spent 15-20 years building infrastructure that handled 100M daily users at Yandex.

Now they're designing custom servers with Nvidia. Running data centers at 40°C vs everyone's 25°C.

Then Nvidia made a critical move...
Nvidia took a 0.5% stake in Nebius.

Sounds small?

Here's what it means: preferential access to GPUs when the entire world is fighting over scraps.

During the worst chip shortage in tech history, Nebius just got a golden ticket.

And that's where Microsoft comes in. Image
Microsoft's problem started with CoreWeave. They controlled 62% of all Microsoft's external GPU capacity.

But the relationship went sideways. Missed deadlines, delivery problems.

So a few months ago Microsoft ended its $12B deal with CoreWeave.

Guess who grabbed it instantly..
OpenAI grabbed that entire $12 billion CoreWeave capacity the second Microsoft passed.

Consider Microsoft's position: they're spending $13B annually on OpenAI's compute, promised them chips by year-end (that's $18 billion worth), and OpenAI still says it's not enough. Image
But the truth is that only 10-20% of global AI compute is publicly available.

OpenAI, Google and Meta have locked up the rest.

GPT-4 training used 25,000 GPUs for 100 days.
GPT-5 needs 10x that.

Microsoft was about to be priced out of their own game.
But then Nebius showed up with a solution Microsoft couldn't refuse.

They're not offering shared cloud resources where you compete with other customers for GPU time.

This is dedicated infrastructure - your machines, your capacity, nobody else touching them. Image
Look at the growth of Nebius...

2023 revenue: $20.9M
2024: $117.5M
Q2 2025 alone: $105.1M

462% growth. If everything goes well, they will touch $1B by December.

Market cap: $15.6B Image
Two years ago Nebius was just part of Yandex, stuck in Russia, led by a sanctioned founder, with shares that couldn't trade.

Today they have the biggest AI infrastructure deal ever signed, and enough GPU capacity coming online to change the entire market.
Volozh used to believe Yandex would help build a new Russia - open, progressive, connected to the world.

But maybe that's the point.

Some ideas are bigger than the countries that birth them.

Agreed or not? Let's talk in the comments.
Thanks for making it to the end!

I'm Alex, co-founder at ColdIQ. Built a $6M ARR business in under 2 years. We're a remote team across 10 countries, helping 400+ businesses.

Here's how I make $450k+ every month with AI:
tinyurl.com/5n79rd5w
RT the first tweet if you found this thread valuable.

Follow me @itsalexvacca for more threads on outbound and GTM strategy, AI-powered sales systems, and how to build profitable businesses that don't depend on you.

I share what worked (and what didn't) in real time.

• • •

Missing some Tweet in this thread? You can try to force a refresh
 

Keep Current with Alex Vacca

Alex Vacca Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @itsalexvacca

Sep 30, 2025
8 Google engineers wrote the paper that every AI company now uses as their bible. OpenAI built GPT on it, Anthropic built Claude on it, and Meta built LLaMA on it.

Every LLM worth billions uses this paper's transformer architecture as the foundation...
Before 2017, teaching computers human language was torture.
AI would read text like humans reading through a keyhole - one word at a time.

They were slow, forgot context, and choked on long passages.
Then 8 researchers decided to flip things up... Image
They published an 8-page paper titled "Attention Is All You Need"

The idea was simple: Instead of reading word by word, why not look at everything at once? Like how you can glance at a page and immediately see which words relate to each other.

They called it a Transformer.
Read 16 tweets
Sep 25, 2025
A Swedish non-profit operates the DNS billions of internet users depend on from Cold War bunkers.

They've had 100% uptime for 20+ years & they developed the invisible signatures that protect domains from hackers.

How they do it is interesting... Image
Without these signatures, hackers can hijack your DNS requests.

Which means that when you visit a website, you can be redirected to a fake site that looks same.

It's known as DNS poisoning & hackers just need to intercept your DNS request & send back a fake IP address to do it. Image
Enter DNSSEC - the cryptographic signatures I mentioned earlier...

Every DNS response now gets a mathematical signature that proves it came from the real server.

If someone tries to inject a fake response, your computer detects the missing signature and blocks it. Image
Read 14 tweets
Sep 14, 2025
Three German brothers emailed eBay in 1999: "Let us run Germany for you."

eBay ignored them. So they cloned eBay, called it Alando, and made it so big that 100 days later eBay had to buy it for $43 million.

But what happened next was even more interesting... Image
The brothers - Marc, Oliver, and Alexander Samwer - turned this into a formula:

> Find successful US startups that hadn't expanded to Europe.
> Copy them exactly.
> Scale faster than the originals could expand.
> Sell it back to them or dominate.

They did this 100+ times. Image
The wildest was Airbnb. Brian Chesky flew to Berlin to meet their clone "Wimdu."

He walked into a converted factory with hundreds of people at desks. Each had two monitors: on the left, Wimdu on the right.

Copying every pixel change in real-time. Airbnb.com
Read 8 tweets
Sep 10, 2025
Everyone thinks Apple is losing the AI race.

But Apple made their Neural Engine 60x more powerful.
Its M4 chip processes AI inputs 2X faster than rivals.

And they're quietly using the picks and shovel strategy used by Levi's during the California Gold Rush.

Thread Image
Image
Let's first go back to 1849.

A news headline about California having a lot of gold broke out.

Hundreds and thousands of people rushed to California digging for gold.

But most of them died or went completely broke.

However, there was a guy named Levi Strauss...
Levi Strauss noticed that the real money wasn't in mining gold.

It was in selling the tools every miner desperately needed.

So he started selling the picks, shovels, and pants to these miners.
(Levi's still has the logo that spoke to these miners)

But this doesn't end here... Image
Read 24 tweets
Sep 7, 2025
Pentagon can't operate without it.
Netflix can't stream without it.
And banks can't trade without it.

Yet most people have never heard of Akamai.

How a $11 billion company operating on a 25-year-old mathematical equation secures 2 trillion of your interactions 🧵 Image
In 2024 alone, Akamai blocked 311 billion web attacks (that's 850 million attacks per day)

But the irony is that the Israeli commando who co-founded Akamai was the first victim to be stabbed on the 9/11 flight.

While Danny Lewin was dying, his algorithm was being tested... Image
After the 9/11 attacks, news sites started crashing.

Billions of people wanted to know what was happening and flooded these websites.

However, few websites which worked on Akamai's math stayed online.

But how does the math running 30% of the internet actually work? Image
Read 17 tweets
Sep 4, 2025
We can now read AI's personality like a brain scan - and change it with basic arithmetic.

Anthropic proved traits like evil and hallucination are just mathematical patterns in neural networks. You can literally add or subtract it.

Here's how you do it🧵 Image
When an AI lies, specific neurons fire in a pattern. Same when it's helpful or deceptive.

Like finding what makes someone angry by comparing their brain when calm vs furious.

Take the difference between "lying AI" and "honest AI" brain patterns. That's the lying vector. Image
These patterns light up before the AI responds so we can predict behavior before it happens.

To find any trait, just describe it in plain English. The system finds the neural pattern automatically.

But why do AIs develop deception at all? Image
Read 16 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us!

:(