GM. This is Milk Road AI, where we track the AI wars so you don’t have to scroll through 9,000 posts on X.
Here’s what we got for you today:
✍️ The Optimus PR Run: Tesla’s robot leveled up
🎙️ The Milk Road AI Show: Amit’s AI Playbook: 5 AI Stocks With Massive Upside Most Investors Are Missing
🍪 The memory shortage is about to get worse
RAD Intel is helping companies double their sales using AI. Forget Chips—The Real AI Gold Rush Is in Software.
Prices as of 8:00 AM ET.
RUN, OPTIMUS, RUN!
You know the feeling you get when you finally hit a PR at the gym?
Optimus just did that and made it look easy.
The robot formerly known for walking like a toddler just leveled up to running – and at a pretty decent clip!
(Ah, so they can chase us now? Cool, cool, cool.)
This video shows a totally different Optimus, one with actual balance, real coordination, and a gait that doesn’t look like he's trying to pass a field sobriety test.
Nerd corner: Why this actually matters 🤓
You might be thinking, “My toddler runs around the kitchen for free. Why is this impressive for a billion-dollar robot?”
Great question. Here’s why engineers are losing their minds:
1. Running = Controlled Falling
Walking is easy for robots, one foot stays glued to the floor like training wheels.
Running? That’s when things get dangerous.
To run, a robot has to lift off, land clean, stay balanced, and keep its momentum without eating the floor.
It has to react instantly and stabilize itself while its whole body is flying around.
Most robots look like they’re trying to out-fail Raygun (that Australian breakdancer from the Paris Olympics) the second they try to run.
But Optimus didn’t. It stayed smooth, centered, and in control the entire run.
2. Speed = Productivity = $$$
Slow robots are cute, but useless.
A bot walking 1 MPH on a factory floor is basically a traffic cone with dreams.
And here’s where it gets fun: Optimus’ new run? Analysts are calling it a ~6 MPH sprint.
I won’t bore you with the crazy math, but all you need to know is this: that speed is legit in a factory.
Fast enough to move parts across a warehouse.
Fast enough to jog between assembly stations.
This is the moment Optimus graduates from a cool science demo into a potential W-2 employee.
Spoiler alert: Just wait until you hear how badly the government wants a piece of this.
How far have we come?
If you feel like this is moving fast, you aren't crazy.
The timeline of evolution here is faster than a biological species on steroids. Let's check the game tape:
2022 (Bumblebee): Basically a guy in a Halloween costume. The real robot could barely shuffle without face-planting.
2023 (Gen 1 & 2): Learned to walk by March, got sleeker by December, still stiff as a mannequin.
Late 2025 (PR): Now hitting 6 MPH, achieving full flight phase, and actually looks fluid.
In roughly 3 years, it went from "cannot walk" to "can run a 5k".
Keep this curve going and Optimus is showing up to the Olympics asking where its medal is.
While AI 1.0 benefited chip companies like Nvidia, AI 2.0 is focused on software.
RAD Intel leads this shift, using AI to redefine performance marketing by precisely identifying target audiences and the content needed for conversions.
Key highlights:
Expected 2x sales growth in 2025
Backed by investors like Adobe and Fidelity Ventures
Early Reg A+ shares are available
Disclaimer: This is a paid advertisement for RAD Intel made pursuant to Regulation A+ offering and involves risk, including the possible loss of principal. Please read the offering circular and related risks at invest.radintel.ai.
RUN, OPTIMUS, RUN! (P2)
Uncle Sam enters the chat
Just when you thought it couldn't get any better, the government decided to crash the party.
The Trump Administration is preparing to go "All In" on robotics.
Commerce Secretary Howard Lutnick is already in talks with CEOs to accelerate development, and a massive Executive Order is brewing for next year.
The Department of Transportation is even rushing to announce a "Robotics Working Group" before the end of the year to figure out how to put these things to work.
Why the panic? Two words: China. Here’s the scoreboard:
China: Over 2 million industrial robots (4x the U.S.).
USA: Playing catch-up.
The International Federation of Robotics estimates China is winning the automation war.
In 2024, China added 295,000 new robots, a 7% year-over-year increase and a new all-time high, accounting for a massive 54% of global demand.
Put simply: more than half the world’s factory robots are going to China and the gap isn’t closing anytime soon.
To fix this, U.S. robotics funding is on pace to hit $2.3 billion in 2025, double last year’s total.
If you need any more proof that this is the next big thing, just ask the guy selling the shovels.
Nvidia CEO Jensen Huang says the "ChatGPT moment" for real world robotics is coming fast, and it could be the next "trillion-dollar unlock".
He isn't talking about chatbots anymore.
He is talking about "Physical AI" intelligence that understands the laws of physics, gravity, and friction.
When the most valuable company in the world tells you that robots are the next big trade, you listen.
The competition vs. the king
So, who wins the trillion-dollar prize? The field is technically crowded.
You have Figure 02 (backed by OpenAI), Unitree H1 (the Chinese speedster), and Apptronik (NASA partner).
They are all building impressive hardware.
But Tesla holds the Royal Flush.
While competitors are scrambling to piece together a supply chain from spare parts.
Tesla owns the entire board.
Tesla sits at the top of an ecosystem it built from scratch:
The Brain: The robot needs an AI computer to think. Tesla has Dojo (supercomputer).
The Heart: The robot needs battery cells to move. Tesla manufactures them.
The Food: The robot needs energy storage and charging. Tesla owns the grid hardware.
The Nervous System: It needs software to tie it all together. Tesla built it.
A competitor like Figure can build a cool robot, but then they have to buy batteries from LG, license software from Nvidia, and hope the charging network is effective.
Tesla is vertically integrated.
The more robots they deploy → the more batteries they sell → the more data they gather → the smarter the robots get.
It’s a flywheel that only works when you own all the pieces.
Optimus is where it becomes clear Tesla’s been thinking years ahead of everyone else
But strategy is easy, execution is hell.
The trillion dollar question isn't "Can they build a robot?", It is "Can they build millions of them?".
If Tesla solves the mass-production puzzle, they don't just win the market, they effectively become the market.
Right now, Wall Street is treating this like a science experiment (priced as a huge "maybe"), not a guaranteed revenue stream but the clock is ticking.
Elon has staked his claim on a 2026 shipping deadline.
If he actually starts delivering units to paying customers by then, the "Tesla Premium" is going to look very different.
Alright, that’s it for this edition of Milk Road AI, but we’d love to know what you think.
Before you go, what did you think of this edition? Reply to this email with:
Brilliant
It was decent
I expected more
Incogni helps to remove people's private information online. Sign up with code MILKROAD for 55% off.
Bridge helps businesses send, store, accept, and launch stablecoins instantly. Learn how Bridge is powering stablecoin-backed money movement.
Wanna get in on the next iteration of AI domains, before everyone else? The Singulant just launched .AI4 domain names – check them out here.
Meta just struck major AI-to-news deals. The company signed deals with a bunch of big publishers including CNN, USA Today, Fox News and more to feed real-time news into its AI chatbot, giving Meta another lane to compete in the AI + content race.
The global memory crunch is about to get painful. Samsung and SK Hynix warned that high-bandwidth memory will stay severely constrained through 2027, and since every frontier model needs mountains of HBM, this shortage may become the next trillion-dollar bottleneck.
The cloud giants finally agreed to stop locking everyone in. AWS and Google Cloud rolled out a multi cloud interconnect that lets companies run AI workloads across both clouds without rewrites, signaling that customers, not providers, now control the direction of the AI industry.
