Gene Munster

I think $TSLA is up tomorrow on the NHTSA detail that Tesla is ready to roll out 1k Cybercabs in Austin. Last Thursday I estimated they would reach 300 over three months. That’s bullish for the stock because it shows Tesla’s confidence in safety is higher than most thought. And it’s particularly positive given Elon’s history of paranoia on the topic. ps: The fact that NHTSA is fighting Tesla is a speed bump for Cybercab. They’ll eventually give in, then it’s off to races.

NHTSA needs a wake-up call. The agency rained on the $TSLA Cybercab launch by opening a probe into Tesla’s self-certification — specifically whether a vehicle with no steering wheel or pedals meets federal safety standards. Question for NHTSA: why slow the rollout of autonomous vehicles when the best data we have says they’re 7-10x safer than human drivers? I suspect the answer is the march of nines. With all due respect, NHTSA needs to update the rules written for human drivers and move faster. About 37k people died on U.S. roads last year. If robotaxis delivered even a 7x cut in fatalities, that number would be closer to 5k.

Elon’s days of being “paranoid” over accidents appear to be coming to an end. My guess is $TSLA adds 300 Cybercabs to Austin over the next month. Thats just the start.

$TSLA just put Waymo, Uber, and Lyft on notice: https://t.co/DAZSo1BN05

Yesterday, PwC came out with their long-run global AI data center spend report. My take is that their estimates dramatically understate the staying power of data center (aka the brain of AI) growth. To put it in perspective, data center spend is still growing about 70%+ in the near term. PwC estimates annual growth between 2030 and 2050 will average 2.5%. That number feels too low. My best guess is that it probably approaches between 5–10% per year, as supported by PwC’s upside scenario of 6.5%. Those rates may look close on paper (2.5% vs. my midpoint of 7.5%), but they are not. Compounded over 20 years, 10% growth leaves annual spend about…

A humble suggestion to Ternus on his first day; make a splashy AI hire by year-end. Yes, it may cost $50m a year but the payoff would be immediate on multiple levels. 1. recruiting AI talent 2. injecting energy into product development 3. AAPL would go up. https://t.co/n7HxcgUxSq

In the Apple-OpenAI he-said-she-said trade secret dispute, the market is believing Apple, with shares up 3% today vs. the Nasdaq down 1%, on the news that $AAPL has more evidence against OpenAI of stealing hardware secrets. My take: In reviewing what both companies have said over the last 24 hours, it seems Apple has the upper hand given their bold claims of evidence destruction are likely true because if they weren’t true, it would be a reckless move from a company that typically plays a conservative legal hand.

Welcome John Ternus to your new role as CEO. Your first priority: reverse the brain drain. OpenAI says roughly 400 Apple employees have moved from Apple to OpenAI. AI is moving much faster than the smartphone era did. The biggest long-term risk for Apple is losing control of the interface between consumers and AI. Given Apple can lean on Google’s models, it’s unlikely that OpenAI can create an AI-powered consumer device family that compels consumers to switch because it’s 10x better than Apple. It’s easy for OpenAI to design one or two consumer devices. The difficult part is designing, manufacturing, and distributing a family of devices a…

Get ready for an AI tsunami. $NVDA’s CY27 revenue guide means the AI brain next year will be about 40% bigger than I thought 24 hours ago, which means the distribution will be greater than I previously expected. https://t.co/T4zRZUqKdf

Oh boy. $NVDA guided next year revenue growth to at least 70%. The street was at 44%.

$SPCX accounted for about 5% of overall revenue. Last quarter it was more like 3%. Looks like they moved $SPCX revenue from ACIE into Hyperscalers which makes sense because SPCX is now a hyperscaler with 8GW coming online next year, same as $META and $AMZN. If you put SPCX back into ACIE in July, ACIE would have grown 160% yy, vs. Street 130%. That would imply hyperscalers growth of 83% in July, inline with the Street. For the record, with the reported recast growth rates were hyperscalers up 101%, and ACIE up 138%. This does not change the fact, the the script has flipped to ACIE having higher growth rates going forward vs. hyperscaler.

$META is up 3% (Nasdaq is down 0.1%) on the news of the national tech addiction settlement which underscores investors value clarity, even if the news is bad. The company will pay $18B over the next 10 years. Last week, I had estimated the number would be below $25B, so these penalties are in line with expectations. The settlement has echoes of the Big Tobacco settlements in 1998 on two levels. First, social media is addictive and can be damaging. Second, the penalties will be chump change to the providers in the end. For example, the annual $1.8B in penalties compares to what will likely be $40-60B in annual FCF in 2029.

Here’s my $NVDA preview (they report next Wednesday). The three most important topics: 1) The pace that the growth driver is shifting away from hyperscalers. 2) Impact of SpaceX massive jump in AI spend on next year's overall revenue. 3) Will the component environment push gross margins below the Street exiting CY26? https://t.co/NmsHIio2sF

What’s going on with $META today (down 4.4% vs. the Nasdaq down 1.3%) around the states, teen safety, and social media addiction is just the tip of the iceberg compared with what’s around the corner. And by “around the corner,” I mean over the next year. https://t.co/UUMnB2iJyO

When it comes to reaching the top of the AI model leaderboard, it pays to be smart. https://t.co/zbZB5X67Vb

Elon just put OpenAI and Anthropic on notice. Grok 4.6 came out yesterday and jumped from 8th to 4th on the leaderboard (@ArtificialAnlys). I expect by January Grok will be at the top of the leaderboard because of 4.7 and 5.0. 4.7 comes out in a month, and will use Cursor data for the first time. 5.0 comes out likely in December and will leverage all of SpaceX engineering data. This model is going to be a monster.

Which AI model has generated the highest returns over the past year? Probably not the one you’d think. The IA 500 says it was Qwen, which was up 42.8% since 10/1/25 vs the SPY up 16.6%. Check out the IA 500, @iafunds latest benchmark for measuring the performance of frontier AI models for investing at https://t.co/yK1hVeiEcj. Disclaimer: The IA 500 is experimental and for informational purposes only. Past performance does not guarantee future results. Investments involve risk, including possible loss of principal. Additional performance methodology and disclosure information available on the IA 500 website.

We're still early: https://t.co/L736Dduu8U

$AAPL down 1% in premarket vs. the rest of Mag 7 up 0.5% on Jefferies downgrade to underperform from hold. I disagree and think the next year AAPL will outperform the rest of Mag 7. https://t.co/h3alevzFka

A thought on $SPCX CY27 Street estimates: The company’s guidance calls for about $8.5B in monthly revenue in December. The Street is currently at about $100B in revenue for CY27, which will be up an impressive 122% y/y. Before I get into this modeling exercise, it’s worth noting the unknowns around the company’s seasonality. For example, the launch business (8% of sales in Dec-26) is lumpy, and the Cursor business is taking off month over month. My basic thought is a majority of SpaceX’s businesses are rapidly growing month over month, which leads me to believe SpaceX’s revenue will grow month over month in CY27. Having fun with numbers, if…

The dust is settling on the $SPCX analyst estimates. As a reminder, the company only gave guidance for an exiting 2026 revenue run rate of $100B . That basically implies December quarter revenue of $22B, versus $7.8B reported in June. This means December revenue estimate numbers went up by about 17%.

$SPCX is taking Elon's first principles approach to building the worlds intelligence infrastructure.

Don't get caught up in capex noise or the concept that $SPCX is an AI and Space business. They're the worlds only sovereign AI company and early in building the what I believe will be the world's most valuable company. https://t.co/Lw0BqzNieM

Call just wrapped. More to come from me.

Starlink Version 3 (launches in next month) has 10x capacity and needed for sub growth and higher valued tiers. My take: This helps answer how they can keep growing Starlink at fast pace.

I'm not seeing any guidance. Call starts in 15 mins.

Capex came in 39% higher than the Street at $18.4B. It was $10.1B in March. My take: This is not as big as a negative as investors reaction to higher capex from the hyperscalers. The reason is investors know the buildout will be significantly over the next several years. For CY27, the Street is looking for 91% capex growth.

$AAPL investors will feel better about the September guide and new iPhones in the weeks to come. https://t.co/1j0s3QTNZH

On a personal note, I want to salute @tim_cook. He' s been making time for these earnings since like 2004. It used to be the Han and Chewie show with Jobs and Cook. Then he stepped into shoes that were impossible to fill and took off running. Bravo Tim.

The $AAPL is over. Stock down 6.5%, a slight improvement from the down 8% at after-hours session lows. The stock turned around slightly with the explanation of why supply constrains get so much worse in September along with a subtle comment from Cook suggesting it would take one quarter to fix it. That one quarter comment means that the impact from September to December will be neutral to a slight improvement.