Tesla dropped its Q2 report. The stock got punched in the mouth. I filmed this the next morning in the car — no script, no slides, no “content strategy.” Just a thesis check.
Four times a year you get to ask the only question that matters if you run a concentrated book: did anything break?
Short answer: no. Timeline slipped. Thesis didn't. I'm not selling.
Not financial advice. Personal opinion from an operator who uses the product and owns the concentration. Do your own work.
The only thesis that matters
I've said this for a long time: Tesla is an AI company that makes cars right now.
Not a car company dabbling in software. Not a meme. A company that already builds physical machines, makes money doing it, and is wiring real-world AI into those machines.
That framing is why I own it in size. It's also why earnings weeks are chaotic. The market still doesn't know how to price “physical AI on a delay” versus “chatbot with a subscription.”
LLMs are becoming commodities. Physical AI is not.
If you actually run AI in a business — not just play with demos — you can feel the divergence.
Large language models are incredible. They're also racing toward commodity pricing. New model every week. Open source a week later. Token costs in free fall. The “moat” on pure chat is measured in days, not decades.
Physical AI is a different animal. You're not asking a model to draft an email. You're asking software + sensors to move a 2,000+ pound object through the real world with people inside it and people outside it. You don't get to Ctrl-Z a bad merge onto 195.
That's the moat. Hard problem. Slow problem. Valuable problem.
My proof isn't a slide deck. It's my commute.
While I was recording, Full Self-Driving was handling about 97% of the miles in my 2024 Model Y. I drive the other 3%.
What is that 3%? Backing out of the garage when it wants to kiss the wall. Parking lots when I don't like the spot it picks. The occasional pull for my kid who wants to feel the motors.
That's it. That's the hard edge of the problem. And that's what has to get to “no steering wheel, no pedals, no human takeover” before robotaxi is a real balance-sheet story at scale.
Blue Cruise and the other “hands-free on the highway” systems are not the same sport. Automated cruise is not camera-driven AI driving a car in the messy world.
What Q2 actually did to my head
It didn't make me a bear. It made me honest about the calendar.
- Thesis: unchanged — real-world AI + manufacturing muscle + long runway into robotaxi and humanoid labor.
- Timeline: longer than the hype machine. “End of this year” energy is not how I'm underwriting it. Think longer. Let it cook.
- CapEx: spending to build the future is not a scandal at a growth company. Google is lighting hundreds of billions to catch up on physical infrastructure. Growth costs money before it prints money.
- Cybercab: cars in the wild without full network deployment is not failure. Calibration, miles, edge cases — the boring work — is the work. Shoutout to Nick Gibbs for the camera-calibration framing; it tracks with how FSD actually behaves across different bodies.
Polarizing CEO. Messy communication. Belichick press-conference energy on timelines. That's part of the package. If you can't underwrite the leadership, you can't underwrite the stock.
Sentiment is a weather system
A month ago Bitcoin was “over.” Then the tape and the narrative flipped. Now the cold air is sitting on Tesla and pieces of AI.
I've watched this movie before — PC, internet, smartphone. This AI cycle is bigger. The companies that win will be the ones that can build and run intelligent systems in the physical world, not just rent tokens in the cloud.
Total addressable market talk gets abused. Still: transportation + labor is not a niche. If they execute, the numbers get stupid. If they don't, concentrated ownership will hurt. That's the trade. I know which side I'm on.
What this has to do with Fall River real estate
I don't post Tesla takes on a property management site for vibes. I run Fortified Realty Group the same way I run a concentrated book:
- Few bets. Deep knowledge. Daily thesis checks.
- AI is not a slide in a pitch deck — it's how we operate.
- Physical world first. Buildings, tenants, vendors, cash. Software that doesn't touch reality is a toy.
- Long horizon. South Coast multifamily is not a weekend flip mentality if you want durable income.
Same brain. Different asset class. If you own property on the South Coast and you're still managing blind — spreadsheets, hope, and a phone that only rings when something's on fire — that's the version of “the last 3%” in our world. Property management is how you close it.
Boots on the ground. You can't manage what you can't see.
Related read: SpaceX IPO — the AI lesson behind how I'm building Fortified.
Bottom line
Q2 was a gut check. Stock price is noise on a long clock. My car is still driving most of the miles. The company is still pointed at the hardest, most valuable version of AI. My timeline moved. My position didn't.
If you want the full unfiltered version — Cybercab, Optimus, SpaceX chatter, the sentiment rant — watch the video above.
David M. Ferreira · MA Broker #9537412 · Fortified Realty Group · Fall River, MA · 508-671-7228
FAQ
Did Tesla's Q2 change your investment thesis?
No. It adjusted my timeline. I still underwrite Tesla as an AI company that manufactures physical products at a profit and is pushing real-world autonomy and robotics. The hard last miles of FSD and deployment take longer than bullish calendars want.
Why talk about Tesla on a Fall River property management site?
Because the operating system is the same: concentrated bets, real-world constraints, and AI that has to work in messy physical environments — not demo-ware. Fortified Realty Group is being built as an AI-native operator on the South Coast.
Is this financial advice?
No. It's one owner's personal reaction and process. I'm not your advisor. If you buy or sell anything based on a guy in a car with a hat on, that's on you.
What's the “97%” number?
On the drive when I filmed, roughly 97% of miles were FSD-driven on my 2024 Model Y. The remaining stretch was human — garage, lots, preference. That's illustrative of progress and of the edge cases left — not a lab benchmark.