The AI Bazooka Reloads: How OpenAI’s Mega-Deals Change Everything

Editor’s note: “Say Hello to the $400 Billion AI Bazooka Aimed at the Market” was previously published in October 2025 with the title, “Say Hello to the $400 Billion AI Bazooka Aimed at the Market.” It has since been updated to include the most relevant information available.

The skeptics are getting louder again. They’ve got their charts, their bubble analogies, and their “AI stocks can’t keep this up” narratives. And yet – the market keeps proving them wrong.

Just look at this past month: OpenAI signed more than $400 billion worth of next-generation chip and compute deals with Nvidia (NVDA), AMD (AMD), Broadcom (AVGO), and Oracle (ORCL) – locking in over 26 gigawatts of GPU power for its coming models. Meanwhile, Big Tech’s AI capex continues to surge toward record highs.

It all underscores one simple fact: most of the money in the modern economy is still going to AI. That’s why AI stocks have been soaring… and why they’re likely to keep doing so.

When you follow the money, the trail doesn’t just point to AI; it leads straight to the beating heart of the global economy itself…

Big Tech Is the Global Economy

Let’s stop pretending otherwise: Big Tech doesn’t just influence the global economic order – they are the global economic order. 

Nvidia, Microsoft (MSFT), Apple (AAPL), Alphabet (GOOGL), Amazon (AMZN), Meta (META), Broadcom, Taiwan Semiconductor (TSM), Tesla (TSLA) – the valuations alone tell the story.

All are trillion-dollar companies. If Nvidia and Microsoft were countries, they would rank as the fifth and sixth largest economies in the world, bigger than India, the U.K., France, Italy, Brazil, Canada, Russia, and Mexico. 

And these are the behemoths that are going all-in on AI spending.

Not dabbling or cautiously allocating. They’re loading up a financial superweapon – what I call the AI Bazooka – and firing it almost exclusively at expanding AI infrastructure.

The AI Capex Explosion

The numbers are staggering. Let’s start with Meta.

In 2022, the company spent about $32 billion on servers and data centers. This year, analysts expect spending to approach $40- to $45 billion, with AI infrastructure being the biggest driver.

And it isn’t alone.

Microsoft’s annual capex has nearly doubled since the pandemic, topping $50 billion, and guidance suggests it could push higher as AI cloud demand accelerates.

Alphabet has boosted annual capex from roughly $25 billion in 2021 to more than $50 billion in 2025, with a majority directed at AI compute and data centers.

Amazon remains the heavyweight: consensus estimates peg AWS-related spending above $70 billion this year.

Together, the “Hyperscale Five” are projected to exceed $250- to $300 billion in annual AI-related capex – a dramatic leap from under $150 billion in 2022.

That means that in less than three years, we’ve seen a 2.5x increase – and it’s still accelerating. Just look at that chart above. Like AI stock prices, it’s going vertical.

And just when it looked like the hyperscalers were maxing out their budgets, OpenAI stepped in and turned the spending dial to 11…

The OpenAI Effect: Supercharging the AI Arms Race

According to recent reporting from Yahoo Finance, OpenAI has inked a series of massive hardware deals worth up to 26 gigawatts of GPU capacity… enough computing power to run more than 20 million homes.

In September, the ChatGPT creator signed a landmark agreement with Nvidia, which will invest up to $100 billion in OpenAI in exchange for the company buying roughly 10 gigawatts of GPUs over the next several years. The first deployment will begin in the second half of 2026, when Nvidia rolls out its next-generation Vera Rubin superchips.

Not stopping there, OpenAI also struck a deal with AMD, purchasing a stake equal to about 10% of the chipmaker in return for another 6 gigawatts of GPU supply. Those will also arrive in 2026, when AMD launches its MI450 series.

And just this week, OpenAI announced yet another blockbuster partnership – a 10-gigawatt co-development deal with Broadcom to design and deploy custom AI accelerators purpose-built for OpenAI’s next-generation models.

Add in OpenAI’s $300 billion “Stargate Project” with Oracle, which includes another 4.5 gigawatts of compute, and the scale becomes clear: this is the largest private buildout of computing capacity in human history.

All told, these commitments represent an unprecedented wave of spending that could reshape global chip, data center, and energy markets. And they underscore a simple truth: AI infrastructure is the new oil.