The $2 Billion Seed Round That Proves AI Has Lost Its Mind

How Mira Murati just raised more money for a 6-month-old startup with ZERO product than most companies see in their lifetime - and why this signals the peak of AI madness

This week, the AI industry officially jumped the shark.

Mira Murati raised $2 billion for a company that doesn't exist. Meanwhile, Trump is preparing to boost energy supply to power AI expansion with executive orders because power demand from AI data centers could grow more than thirtyfold by 2035.

We're watching the biggest bubble in tech history inflate in real time, funded by taxpayers and powered by delusion.

Let me show you the numbers that will make your head spin.

The $2 Billion Fever Dream (It's Real, and It's Spectacular)

Thinking Machines Lab, the secretive AI startup founded by OpenAI's former chief technology officer Mira Murati, has closed a $2 billion seed round at a $10 billion valuation. The 6-month-old company's work remains unclear.

Read that again: $2 billion for unclear work.

To put this insanity in perspective:

• The next-largest U.S. seed financings have all been in the $200 million to $450 million range

• This deal is 4.5x larger than the previous record

• Investors had to commit at least $50 million just to participate

• It's by far the largest seed round in the Crunchbase dataset. It's not even close

What Murati is selling: "AI systems that are more widely understood, customizable and generally capable"

What investors are buying: Hope, dreams, and FOMO.

📊 BRUTAL MATH: $2B raise ÷ 6 months = $333 million per month burn rate

That's $11 million per day. For a company with no product.

The Energy Crisis Nobody's Talking About (But Should Terrify You)

While VCs throw billions at vapor, the real crisis is hitting: America's power grid is about to collapse under AI demand.

The numbers are absolutely terrifying:

• Between 2024 and 2029, U.S. electricity demand is projected to grow at five times the rate predicted in 2022

• AI data centers could require 68 gigawatts of power globally by 2027 — almost equivalent to California's 2022 total capacity

• Goldman Sachs Research forecasts global power demand from data centers will increase 50% by 2027 and by as much as 165% by the end of the decade

• 72% of power and data center executives say power grid capacity is very or extremely challenging to AI infrastructure expansion

The government response? Trump's executive actions to make it easier for power-generating projects to connect to the grid and provide federal land for data centers. The government is literally reshaping energy policy around AI hype.

🔥 REALITY CHECK: Large hyperscale data centres have power demands of 100 MW or more, with an annual electricity consumption equivalent to the electricity demand from around 350,000 to 400,000 electric cars.

The $500 Billion Government Bailout Disguised as "Innovation"

Remember when private companies were supposed to fund their own infrastructure? Those days are over.

President Donald Trump announced a joint venture investing up to $500 billion for infrastructure tied to artificial intelligence by a new partnership formed by OpenAI, Oracle and SoftBank. They're calling it "Stargate."

Here's what's really happening:

• The companies expect to commit $100 billion to Stargate initially and pour up to $500 billion into the venture over the next four years

• Stargate will build 20 datacenters, each measuring 500,000 square feet, according to Larry Ellison, chief executive of Oracle

• The first data center built under the initiative will be in Texas

• Oracle said in a statement that the Stargate project "will secure American leadership in AI, create hundreds of thousands of American jobs"

Translation: Taxpayers are funding the infrastructure for private AI profits through federal land deals and energy subsidies.

The Training Cost Reality Check That Exposes the Bubble

Want to know what it actually costs to build these "revolutionary" AI models?

Google spent $192 million on Gemini 1.0 Ultra—the highest across leading models. For context:

• For Gemini Ultra, a large chunk of the costs were for research and development staff salaries (including equity)—making up to 49% of the final cost

• Meanwhile, AI accelerator chips accounted for 23% of the total cost, followed by 15% for other server components

• By contrast, the DeepSeek-V3 reportedly cost $6 million to train

The pattern is clear: Raise massive amounts, burn through cash training models, hope for acquisition before the money runs out.

Meanwhile: The number of AI-related incidents rose to 233 in 2024—a record high and a 56.4% increase over 2023

What's Actually Happening (The Truth VCs Won't Tell You)

I've been inside these AI pitch meetings. Here's what's actually happening:

The VC FOMO Machine

1. "OpenAI alumni = instant unicorn" - Several of her former OpenAI colleagues have joined the new startup, including co-founder John Schulman

2. "We can't miss the next OpenAI" - If OpenAI managed to secure a recent $300 billion post-money valuation largely driven by the prowess of its team, it's reasonable to expect great things out of its early leaders

3. "AI is different this time" - Every bubble's famous last words

The Hidden Infrastructure Crisis

"AI servers use up to 10 times the power of a standard server, and companies are deploying them at an unprecedented scale"

• 1.6 million servers are related to AI out of 21 million total servers in 2020

• Electricity demand from data centres worldwide is set to more than double by 2030 to around 945 terawatt-hours (TWh), slightly more than the entire electricity consumption of Japan today

• Electricity is by far the largest ongoing expense for datacenter operators, accounting for 46% of total spending for enterprise datacenters and 60% for service provider datacenters

The Three Signs This Bubble Is About to Pop

Sign #1: Record-Breaking Rounds for Nothing

When a 6-month-old startup with unclear work raises $2B, you're at peak bubble. This represents a 142-fold reduction in parameters needed for equivalent performance over two years - yet somehow companies need MORE money than ever.