Themes

A lot happened in 2025 when it comes to robotaxis, which makes it fitting to mark where things stand today between Waymo and Tesla. This year, Waymo moved beyond the pilot phase and entered the scaling phase, with a fully driverless service operating in five US cities and a fleet of around 2.5k vehicles. Tesla’s Robotaxi is running a fleet of around 50 vehicles, largely supervised in Austin and the Bay Area. While Waymo has the lead, the market remains nascent, likely exiting 2026 with about 1% of rides in the US, which means Tesla has plenty of time to catch up. For 2026, it will all come down to how the no safety driver Austin test goes. If that goes well, look out Waymo.

For the past month the market has been nervous that the AI bull market may be approaching an end. The latest pulse on that topic came this week from AWS re:Invent 2025. Amazon's message was clear, they are accelerating its AI buildout and agents will have a profound impact on the future of work. Unfortunately the market is not buying it, with shares of AMZN, GOOG, NVDA, MSFT, and META down an average of 0.5% in the day following the bullish comments from Amazon, while the Nasdaq is flat.

The market appears to be looking past the facts about how quickly AI is advancing, evidenced by NVDA shares being down 3% following better than expected earnings and guidance and the Nasdaq falling 2% over the past five trading days.…

One nagging question around AI is what the long-term business models will look like and whether these companies will be able to charge enough to become profitable. We looked at the ten most popular paid consumer subscription apps to get a better sense of historical pricing power. On average, their prices have increased 1.7% faster per year than the rate of inflation. The bottom line: this is good news for ChatGPT, Grok, Gemini, Claude and Perplexity, because it shows that if you get consumers hooked, you can gradually start raising prices.

The road to purchasing Meta Display led through a local Best Buy, where a mandatory demo was required to ensure proper fitting. In the end, I found the technology impressive, the use case still limited, and the fashion grade below average. The bottom line: Meta is making the right move by investing $20B annually into Reality Labs, but the return on that investment will take years.

Apple’s decision to shelve Vision Pro in favor of glasses underscores that the optimal AI-first form factor is still unsettled. Apple is now following Meta’s push into glasses, while Jony Ive joined OpenAI in May to pursue a different path: a screenless pocket companion. I believe the phone will remain the dominant AI-first device over the next three years. Beyond that, the shift will move toward pocket companions that work alongside phones and watches. Glasses, despite Apple’s pivot, are likely capped at a few hundred million units a year, limited by comfort, privacy, and fashion.

At the end of the month, the U.S. EV tax credit will sunset, which may explain why in August Ford and GM reset their EV plans and in September VW cut its U.S. 2030 EV target to 20%, down from 50% three years ago. This tempered EV outlook contrasts with increasing investments into autonomy, pushing these companies into a Catch-22: saving money by slowing the path to electrification dampens the autonomy reward.

At the surface: Oracle’s historic $300B contract with OpenAI cements Oracle’s position as an AI hyperscaler. Below the surface: The deal underscores just how early we are in AI, and how massive the infrastructure layer will prove to be. That foundation will power AI-native companies, which in turn will reshape industries and change the world.

For years, EV buyers fixated on range as the key buying metric in response to range anxiety. As Tesla enters the autonomous era, a new metric is taking center stage: the strength of the car’s self driving hardware. A Tesla’s processor and sensor suite (HW3 vs HW4 vs upcoming HW5) now determine how well it can run FSD software and whether it can access future autonomy upgrades. I believe HW5 will be needed for the fleet to go mainstream, and upgrading existing hardware will be cost prohibitive. The bottom line: Tesla is best positioned to be one of the biggest winners in autonomy, despite the reality that rolling out new hardware will be a near term governor on growth.

Every now and again, innovators share a behind-the-scenes look at their work. Sometimes these breadcrumbs lead us down a dead end; sometimes, however, they lead us down a path of transformation and revolution, a permanent path that forever changes the…

Jony Ive and Sam Altman are making a bold attempt to create an AI-native, screenless computing product lineup that could change how we interact with technology. Ive appears motivated in part to reverse screen addiction and envisions a family of ambient computing devices that fade into the background while putting advanced intelligence everywhere. If successful, the products collectively could equal a few hundred million units a year, compared to the iPhone at just over 200m. Apple’s core business will likely not be threatened, given Ive's vision appears to sit alongside of and beyond the Mac and the iPhone. And any success he might have will likely motivate Apple to enter the space. Google faces more disruption if AI-centric gadgets divert users from the screens they use for Google searches today.

There’s a shift underway in how people use search and it’s not going in Google’s favor. Last quarter, 56% of total revenue was from Search, and I estimate the segment accounts for 90% of profits. The company is still dominant, but their Golden Goose is falling ill. From Eddy Cue's comments at the Google trial, Apple is seeing declining Safari searches (Google) to rising AI competition, the headwinds are real. The stock may look cheap, but without a clear plan for navigating the post-blue-links world, it's likely going to get cheaper.