Waymo is operating in 15 cities with 4,000 robotaxis, reportedly expanding to 8,000 robortaxis by the end of this year and 20,000 robotaxis by the end of 2027. Turns out lidar can really help scale a robotaxi fleet. So much for Elon’s claim that it’s a fool’s errand. Who's that fool now?🤣 Geofencing and HD maps are two things $TSLA fans used to mock @Waymo for. Now they have to embrace them because Tesla’s Robotaxi service uses them too. Will Tesla add lidar? If so, I’d expect it to start with Semi, where longer stopping distances make additional sensing and redundancy even more valuable, especially at night. Otherwise, will the autonomous Semi be limited to day shifts only? 🤔

Wow, I almost forgot that American companies like Dell also make rugged computers lol. TBH, Anduril is starting to look increasingly sketchy with how many new products they're rolling out. 😂

$SPCX is a great business if you look at Starlink and its launch business alone. But the company's strategic focus and capital allocation are increasingly shifting toward neocloud, which could fundamentally change its cash flow profile. With neocloud in the mix, a great business is turning into a great story. $SPCX It could face significant cash flow pressure over the next few years. Starship continues to consume billions in development spending. The neocloud business requires massive upfront investment in AI infrastructure, while D2C adds another layer of satellite deployment costs. SpaceX's capex reached roughly $48B in the first nine months of 2026, versus just $13B in operating cash flow. Q4 capex is expected to hit $35B–$40B,, potentially pushing full-year capex above $80B. @elonmusk claims new AI compute investments can pay back in less than a year. But what exactly does that include? Just GPUs, or the full cost of land, data centers, power infrastructure, cooling and grid connections? If the payback calculation excludes some of these upfront costs, the actual return on total invested capital could be much lower. And even if new AI capacity pays back quickly, that doesn't mean the entire AI business—or SpaceX as a whole—generates positive free cash flow. If Musk's $500B–$1T AI ARR target for 2027 year-end is achievable, the real question is how much CapEx it will take to get there — especially when Wall Street estimates for 2027 capex are already ranging from roughly $200B to nearly $500B, with KeyBanc projecting a staggering $328B in free cash flow burn.

$OUST is no longer just a LiDAR company. It's building a full-stack perception platform for Physical AI. Today's launch of ZED Box 2 makes that vision tangible: LiDAR + stereo cameras + edge AI compute + sensor fusion + perception software + AI models. @ousterlidar is moving beyond selling individual sensors toward delivering an integrated perception stack that enables robots and autonomous machines to see, understand and navigate the physical world. And the business is already scaling. Ouster shipped 9,000 LiDAR sensors and 8,000 cameras in Q2, with LiDAR shipments up 70% YoY. Rev8 is ramping up, and Q3 shipments could be significantly higher. This morning I've updated my Q3 revenue estimate on Substack for paid subscribers. I'll share it publicly on earnings day. The market may still be valuing $OUST as a LiDAR company while the business evolves into something much bigger and stickier. Next Tuesday's Q3 earnings could be a major catalyst. 👀

I still don't understand why $TSLA didn't just use a Model Y as a robotaxi, with no steering wheel or pedals. CyberCab seems like just another one of Elon's toys, just like CyberTruck. Why not simply paint the Model Y gold, if that's what @elonmusk likes, to distinguish it from consumer Model Ys? Tesla must have wasted significant resources redesigning the vehicle and retooling production lines for CyberCab, which has only two seats. Tesla fans and executives will argue that most ride-hailing trips involve just 1–2 passengers. But do they seriously believe that a 5-seat Model Y Robotaxi would scare those passengers away? At least you'd have extra cabin space for your laptop bag instead of having to hold it on your lap when two people are riding together. 😂

Ouster has evolved into a leading perception platform for Physical AI — combining LiDAR, cameras, edge computing, software, and AI models. $OUST Yet the market cap is only $2B today.

Can’t fight the trend. $OPEN needs to wait for a tailwind. @nejatian

2~3 Elon years = 4~5 Earth years, if you’re lucky. $SPCX $ASTS

$OPEN is tanking even harder today. The departure of the company's president just two weeks before earnings is extremely concerning. “Personal reasons” sounds like a convenient excuse. What could possibly be so urgent that he couldn't wait until after the earnings call? The fact that the company will continue paying him through July 2027 suggests this may have been a negotiated exit rather than a voluntary departure. Meanwhile, CEO @nejatian previously said the last two weeks of August had pushed out the company's path to ANI breakeven by 6–8 weeks. September was even worse. And we know winter is traditionally the weakest season for real estate. So how much further has breakeven been pushed out — a few more weeks or several months? Kaz has been awfully quiet about that. It looks like Kaz built his model around a rate-cut cycle that would revive the housing market. But the war in Iran may have thrown that entire thesis off course.

To be clear, I think the $4.1B valuation for @BedrockRobotics is fair. It’s just that $OUST is way undervalued because the market still hasn’t realized the potential of the leading perception platform for Physical AI — the only one in the Western world.

屠龙少年变成恶龙。 川普和马斯克都是这样，让人无比失望。 我昨天发文说，这次中期选举给民主党投票不是因为民主党变好了，而是共和党变烂了。