GM. This is Milk Road AI, the newsletter that tracks the AI arms race so you don’t have to doomscroll it.
Here’s what we’ve got for you today:
- ✍️ Tesla’s most dangerous pivot yet.
- 🎙️ The Milk Road AI Show: Anthropic’s Rise: Is OpenAI Losing Its Lead? w/ Patrick & Duncan.
- 🍪 Apple’s big bet on on-device AI.
Summ (formerly Crypto Tax Calculator) is a tax software built specifically for crypto. Get started for free with Summ.
Prices as of 10:00 a.m. ET.
TESLA IS KILLING ITS PAST TO BUY THE FUTURE
In 2005, Steve Jobs did something that looked absolutely psychotic to the outside world.
He killed the iPod Mini.
At the time, it wasn't just a popular product, it was the product.
And it was printing money faster than Apple accountants could stack it.
Analysts screamed, shareholders hyperventilated.
Why would you murder your best-selling product at the peak of its powers?
But Jobs saw something the market didn't.
He realized the future wasn't a plastic click-wheel; it was a sheet of glass.
He replaced the Mini with the Nano but he knew the iPhone would make the entire iPod line irrelevant.
In 2026, Elon Musk is pulling a Steve Jobs.
Tesla just announced they are killing their original flagships, the Model S and Model X.
These are the cars that put Tesla on the map.
The Model S was the first electric car that was actually cool, the one that proved you didn't have to drive a glorified golf cart to save the planet.
It made EVs a status symbol.
But Elon Musk has decided he isn't in the car business anymore.
He is pivoting Tesla to become a physical AI company.
And he is willing to burn his own legacy and maybe a few short-sellers to do it.
The scoreboard: A very weird quarter
Let's look at the numbers from Q4 2025, because they paint a picture of a company in the middle of a violent identity crisis.
Revenue came in at $94.8B for the full year. That is down ~3% year-over-year.
For those keeping track at home, that is Tesla’s first annual revenue drop ever.
Since the day they went public, the line has only gone up. Until now.
Let’s translate this for the back of the room.
The thesis that Tesla is just a rapidly expanding car manufacturer is dead.
Automotive revenue fell 11% to $69.5B.
They delivered 1.63M cars, down from 1.81M the year prior.
If you just read the headlines, you'd think the sky was falling.
But if you look under the hood, the business is actually leaner and meaner than it has been in years.
While revenue dipped, gross margins engineered a massive comeback to 20.1% in Q4, the highest in two years.
Despite selling fewer cars, they made significantly more money on each one.
How? Ruthless discipline.
They tightened supply chains, boosted yields on their 4,680 batteries, and stopped the discount war.
They stopped trying to bribe people to buy cars and started focusing on printing cash.
They are now sitting on a war chest of $44.1B.
So, you have a company with shrinking sales but exploding efficiency.
Wall Street is confused. They want Tesla to build a cheaper $25k hatchback to compete with BYD and the flood of cheap EVs coming out of China.
They want Musk to play defense but Musk isn't interested in winning that game.
He’s interested in ending it.
And to prove he’s serious, he is starting at the top.
It’s the end of an era for the Model S and X.
Musk is retiring the veterans who put Tesla on the map, with production ceasing entirely by mid-2026.
The Fremont factory lines that minted the world's richest man are being gutted to make room for the new recruit:
The Optimus robot. The robot that dances better than your uncle at a wedding.
He is killing the products that built the brand to double down on a product that hasn't sold a single unit yet.
It is the business equivalent of free-solo climbing El Capitan without a rope.
But the scariest part isn't the climb itself. It’s how far the fall would be.
THE SOLUTION TO CRYPTO TAXES
Tax season is right around the corner.
And if crypto taxes are already stressing you out, meet Summ.
Summ is a tax software built specifically for crypto.
Why we like it:
- 3,500+ exchanges, wallets and crypto projects (with DeFi, NFT, staking and airdrop support)
- Direct TurboTax integration
- Official tax partner of Coinbase and Metamask
- Helps sanity-check your new 1099-DA form/s.
Stop doing your crypto taxes the hard way.
Get started for free with Summ.
TESLA IS KILLING ITS PAST TO BUY THE FUTURE (P2)
Let’s look at the price tag.
Tesla trades at $1.35T, sporting a P/E ratio of roughly 400x.
For a mere car manufacturer, that price is a hallucination.
The easy EV money has been made.
The market is crowded, subsidies are dead, and competitors like Xiaomi and Ford are fighting for scraps.
Investors are paying for a future where Tesla isn't selling cars, but selling labor.
Proof of this pivot is in the spending.
Tesla is deploying a staggering $20B+ in CapEx in 2026, double what they spent last year.
I know that might not sound like a lot compared to the hyperscalers (Meta probably spends that on snacks), but it's a lot for Tesla when compared to their free cash flow of $6.2B.
And they aren't spending it on more car factories.
They are spending it on three specific bets:
Bet #1: The brain (merging with Grok)
Tesla just disclosed a $2B investment in xAI. But they aren't the only ones.
SpaceX has already invested $2B in xAI.
This is the part everyone misses: Elon’s companies are becoming more and more interconnected.
Think about the long term.
The future will be powered by solar energy and AI satellites operating in deep space:
- SpaceX will launch them.
- Tesla will provide the energy systems and AI hardware to power and control them.
- xAI will be the brain that runs the show.
Tesla powers the body, SpaceX launches the body, and xAI tells the body what to do.
Bet #2: The body (Optimus production)
Tesla is launching the 3rd generation Optimus robot in Q1 2026.
Musk says the supply chain is officially ready for mass production.
They have already selected seven Chinese tier 1 suppliers to build the thing.
The targets are aggressive to the point of comedy: 50,000 to 100,000 units by the end of 2026.
The long-term goal? 1M robots per year by 2027-2028.
Musk thinks the market for humanoid robots is bigger than the market for cars.
He isn't the only one doing the math.
ARK projects that generalizable robotics represents a staggering +$26T global revenue opportunity.
To put that in perspective, that is almost the size of the entire U.S. economy today (~$30T).
He believes that in the future, labor will essentially become optional.
Bet #3: The fleet (Cybercab)
The Cybercab the steering wheel-free robotaxi, starts production in April 2026.
They are using a new unboxing manufacturing process aiming to spit out one car every 10 seconds.
Musk went on record saying that by the end of 2026, fully autonomous Teslas could be operating in up to half of the United States.
An analysis tracking over 94,000 rides between Nov 2025 and Jan 2026 just dropped a bomb on the industry.
The price war scoreboard:
- Waymo: Charging $5.72 per kilometer.
- Tesla: Charging $1.99 per kilometer.
This is a predatory pricing war. Tesla is undercutting Google's self-driving arm by nearly 65%.
While the average Lyft ride was hitting wallets for $15.47, Tesla’s robotaxis were clocking in at an average of $8.17, rarely exceeding ten bucks for a trip across the city.
The catch? Wait times. You are waiting longer for a Tesla (about 15 minutes) than for a Waymo or Uber.
But while everyone is hyper-focused on the robots and the taxis, there is a monster hiding in the earnings report that almost no one is talking about.
The energy business is quietly exploding.
While car sales dropped, energy revenue surged 26.5% to hit $12.8B for the year.
Even better? The segment operates at a massive 29.8% gross margin.
That is nearly double the automotive margins.
It's more profitable to sell a battery than a car right now.
They also deployed 46.7 GWh of storage in 2025.
Why does this matter? Because the AI revolution has a dirty secret:
It is incredibly power-hungry.
Data centers running models like GPT-5 and Grok 3 consume city-sized amounts of electricity.
The grid cannot handle it.
Tesla is positioned to be the battery backup for the entire AI economy.
While they figure out the robots, the battery business provides a massive, high-margin safety net.
It is the perfect hedge.
The verdict
This is arguably the most dangerous moment in Tesla's history.
Even our favorite bear, Michael Burry, couldn't resist weighing in, calling the valuation 'ridiculous' and pointing out the shifting narrative:
"The Elon cult was all-in on electric cars until competition showed up, then all-in on autonomous driving until competition showed up, and now is all-in on robots – until competition shows up."
He might be right about the volatility, but he might be missing the revolution.
The global economy is being rewritten. The companies that cling to the old ways are going to crash and burn.
In the short term, Tesla looks risky.
But in the long term, pivoting to physical AI is the only way to escape the car manufacturing rat race and become a $10T+ entity.
Musk sees the iceberg ahead, and he’s steering the ship toward open water.
He is doing exactly what Steve Jobs did in 2007.
He is killing his own golden goose (the Model S/X) before a competitor does it for him.
The skeptics sound a lot like the Blackberry executives who laughed at the iPhone because it didn't have a keyboard.
They didn't realize the game had changed until they had already lost.
The car was always just the Trojan Horse. The robot is the army inside.
Alright, that’s it for this edition of Milk Road AI. We want to hear from you.
Where do you stand on Tesla’s robot pivot?
- All-In on Robots: Musk is building the next trillion-dollar industry.
- Stick to Cars: This pivot feels way too early.
- I Just Like the Drama: Don’t stop, this is better than Netflix.
ANTHROPIC VS OPENAI: WHO’S AHEAD? 🤖
This past Friday, Patrick & Duncan sat down to talk about the shifting AI landscape and whether Anthropic is closing the gap on OpenAI.
Here’s what you’ll hear:
- Why Anthropic’s $20B raise changed the market narrative and boosted enterprise momentum.
- How Claude stacks up against OpenAI Codex and GPT models across UX, personality, and execution.
- What massive fundraising, political scrutiny, and labor concerns mean for frontier AI labs.
- How open source advances and global players like China are reshaping cost curves and competition.
Hit play and see for yourself 👇
YouTube | Spotify | Apple Podcasts
Bridge helps businesses send, store, accept, and launch stablecoins instantly. Learn how Bridge is powering stablecoin-backed money movement.
Mercuryo is the simple way to buy, sell, and spend crypto with low fees (yes, even with Apple Pay). Discover an easier way to use crypto.
Summ (formerly Crypto Tax Calculator) is a tax software built specifically for crypto. Get started for free with Summ.
BITE-SIZED COOKIES FOR THE ROAD 🍪
Apple bought Israeli AI startup Q.ai for about $2B to boost audio and on-device AI. The move supports smarter AirPods, Vision Pro upgrades, and future AI hardware.
Jensen Huang dismissed claims Nvidia’s $100B OpenAI deal is stalling. He said Nvidia still plans to invest heavily in OpenAI.
OpenClaw’s AI agents are now chatting on their own Reddit-style network called Moltbook. Creators warn it’s experimental and still has serious security risks.
MILKY MEMES 🤣
ROADIE REVIEW OF THE DAY 🥛
