Tesla Stock Long Term: Why I'm Not Taking Profits on Tesla

There’s one piece of advice I hear more than any other, and I can’t stand it: “Take it when it gives.” Sell when you’re up. With the 3x funds especially, people say, you have to take it when it gives.

I don’t know why it bugs me so much. Maybe because it assumes I’m here for a bounce. I’m not. I’m holding because I believe a moment is coming when this company runs, hard.

Believing it doesn’t make it easy. I’ve watched my gains climb past 50% and then slide back down to 10%. More than once. Each time, I tell myself the same thing: give up an arm to take the neck. Let the market take its piece. I’m staying for the big one.

I’ve even watched big gains go by without selling. I’m waiting for something far bigger, and I can see the jump coming. I’m not going to miss it to lock in a smaller number.

That’s the conviction. It’s also why this first post in Skyland’s Investing zone is just my notes: what I think Tesla actually is, what I’ve seen with my own eyes, and what the numbers say right now.

Full disclosure: I own Tesla shares.

The Tesla I see: a map, not a car company

When people say “Tesla,” they picture a car. I see a lot of businesses under one roof.

Cars and software. Model 3, Model Y, Cybertruck, and FSD. (My list had Model S and X too. Tesla ended both this spring, and their Fremont line now builds Optimus.)
Superchargers. V4 keeps getting faster, now up to 500 kW per stall. I also noted Tesla bought a wireless charging company. It did, Wiferion in 2023, then sold the business months later and kept the engineers.
Insurance. Priced by how you actually drive, and it grows as more Teslas hit the road.
Energy. Solar, Powerwall at home, Megapack for the grid.
AI compute. I wrote “Dojo, using Nvidia for now.” Tesla shut Dojo down in August 2025 and trains on Nvidia-based clusters, then restarted Dojo3 work in January 2026, aimed at compute in space.
Robots. Optimus for factories first, then everyone, like a smartphone. Robotaxi, then Robovan.
Factories and batteries. California, Texas, Nevada, New York, Berlin and Shanghai, plus its own 4680 cells and its own lithium refinery.

What Tesla has that others don’t: the Giga Press, the 4680 cell, Megapack, its AI training systems, the robots, and the software behind FSD and Megapack. And around it all sit Elon’s other companies: SpaceX, The Boring Company and Neuralink.

My energy thesis

Right now Nvidia is the one winning big. Fair enough. But I think the next giant win is Tesla’s energy business, and almost nobody is looking at it, because everyone’s eyes are on the cars.

My bet is this: by the time energy growth is obvious to everyone, the stock will already be somewhere else entirely. That’s why I’m not selling. Not a message to you, just the reason I hold.

My FSD diary

I’ve kept dated notes. A few of them:

March 2024. Bad news kept stacking up. On March 5, someone set fire to a power pylon near Tesla’s Berlin factory and production stopped for about a week. A far-left group claimed it. At the time, with everything else going on, it honestly felt like someone wanted the stock down. That was my feeling, not something I can prove.

April 2024. Tesla gave U.S. owners a free month of FSD 12.3, and I finally tried it. It was impressive, and it was not ready. At a No Turn On Red sign, it started to go, and I took over. When we arrived, it tried to enter an ATM lane the wrong way.

November 2024. My note says FSD 13 was out internally. Tesla started sending v13.2 to its first outside customers at the end of that month.

Each version, the mistakes got smaller.

Then and now: FSD 12.3.3 vs. 14.3.13

Two and a half years later, I’m on FSD 14.3.13. My honest verdict: it’s almost perfect now. These days it waits at No Turn On Red signs, and it pulls into parking lots the right way.

What changed in between is public. Tesla’s v14, which started rolling out in October 2025, was its first major FSD update in about a year. It added the option to choose exactly where to be dropped off, pulling over for emergency vehicles, and better handling of detours and unprotected turns. Then v14.3, released in April 2026, made reactions about 20% faster and improved how it reads traffic signs and picks parking spots. Sign reading and arrival, the two things that tripped up 12.3.3, are exactly where Tesla kept working.

This is the part of Musk’s thinking I like most. Back in 2005 he said, “Failure is an option here. If things are not failing, you are not innovating enough.” The SpaceX crowd calls it “test, fly, fail, fix,” and FSD went the same way: ship it, watch it mess up at a red-light sign and an ATM lane, fix it, ship again. I try to hold my shares like that too. A bad stretch is just a failed test on the way to the next version.

“Almost perfect” still isn’t “no driver.” It’s still FSD (Supervised), which means a driver still has to watch the road. But going from “I took over” to “almost perfect” in about two and a half years is the trend I’m investing in.

The future I picture

Here’s where I think this goes.

Robots become something you buy, like a washing machine. Cars stop being something you buy at all. They become transport, like an elevator. You don’t own the elevator in your building. You just ride it.

The human urge to own things won’t disappear. It’ll move somewhere else.

What changes is the ride itself. The apps you use in the car will be different for each person. Some people will sit and do nothing. Others will pay to turn travel time into useful time, and some won’t be able to pay. The same car will feel like a completely different car depending on who gets in.

But some people will still own cars, as a business. If one self-driving car earns something like $10,000 a year (my rough guess, not a Tesla number), people with money will keep adding more. Owning a self-driving car becomes owning income.

What the latest numbers say

Beliefs need a reality check, so I read Tesla’s Q3 2026 report.

Cars: revenue flat at $28.05 billion, but car margins rose to 18.1% from 15.4% a year ago. Battery supply, not demand, was the limit.
Software: active FSD subscriptions hit 1.74 million, up 67%. Cybercab Robotaxi rides started in Austin in September.
Robots: Optimus production began, but deliveries to outside companies are targeted around the end of 2027. Robovan is still a concept.
Energy, the part that stung: storage deployments rose 10%, but energy revenue fell 16% to $2.88 billion, partly from tariff refunds. For my thesis, that’s the number I’ll watch hardest.
Profit and price: net income fell 43% as research spending jumped, and the trailing P/E is around 439. A lot of the future is already in the price.

What would change my mind

Energy revenue keeps shrinking while deployments grow.
Robotaxi still isn’t real revenue in 2027, the year Musk pointed to.
Unsupervised FSD on customer cars, hinted for around Q2 2027, keeps slipping.
Spending keeps rising with the future always “next year.”

And one honest note about those 3x funds people tell you to sell fast: they reset every day, so in choppy markets they can lose ground even when the stock ends up flat. Holding one for the long run means accepting that.

I’ll come back to these notes after the next report.

This isn’t investment advice — just my notes.

Why I’m Not Taking Profits on Tesla