Weekly TikTok Haul Analysis: 11/23/2025

Although Google has gotten quite a bit of love year-to-date, the stock still presents a fantastic opportunity for forward growth, despite being up considerably YTD.  I would consider Google to be a must-have holding to capitalize on AI growth in the future. Here’s why:

For the longest time, there was huge concern that Google’s crown jewel, Google Search, would go away.  However, a couple of years after ChatGPT's release, we’ve found that Google Search is stronger than ever.  There are countless ways to use artificial intelligence and LLMs, and seeking out information is just one of them.  Search queries and search revenue have continued to grow over time and are likely to continue to do so in the future.  

Although you might not think of YouTube as a streaming company, it’s got the majority of TV share when compared to all of the other streaming players.  YouTube continues to be the premier way to find incredibly niche, long-form video entertainment, and has been for just over two decades.  The moat that YouTube has built up over time is truly incredible.  As we all know, you can find everything from movie theater-grade short films to niche tech reviews, how-to videos, DIY videos, and more. If you want to watch a video about something, there’s nearly a 100% chance that you can find it on YouTube.  While there are plenty of places to find niche short-form video content (YouTube included), there is only one place to find niche long-form content, and it’s unlikely that this will change in the near future.  

For the longest time, I believe that Google wasn't showing all of their cards in the AI race, as it had a major looming antitrust case over its head. However, as soon as it received its “slap on the wrist”, they began shipping incredible improvements, like their image generation models, video generation models, and key AI improvements to forgotten product lines, like Google Home. 

We've also seen an incredible jump in capabilities with the release of Gemini 3. It's now the leading flagship model, beating out every single other LLM on almost every conceivable benchmark. The growth in popularity and capability of both the Gemini model and the Nano Banana/Veo models, combined with the failed release of GPT 5, have driven incredible adoption rates of the Gemini app on both the Apple App Store and the Google Play Store.

When you couple this with the fact that Google is shipping practically all new products with Gemini integrations and proactively integrating Gemini into all legacy products (Gmail, Gsuite, Pixel, etc.), Google can create one of the most substantial improvements to it's ecosystem that we've ever seen.

When it comes to frontier labs, there is only one that has the cash, brain power, and the compute needed to scale its models consistently, and that lab is Google. Google still has some of the largest cash flows in the world, giving the team all of the capital they need to fuel expansion, through building data centers and buying chips, without having to rely heavily on debt or equity markets. 

Google also has some of the smartest minds in the world on their teams, with employees of the company winning novel prizes for the past two years! Lastly, Google is not (entirely) beholden to NVIDIA, as it has its own custom silicon, the TPU, that it uses. Given its lower price point than NVIDIA chips, Google can get more bang for its buck in terms of compute per dollar spent.

Google is incredibly fortunate in that they have the largest captive audience in the world (i.e. largest pool of user data in the world). The company has billions of people using it's products on a daily basis, which means it can essentially drink from the firehose that is the entire world’s use of the internet at whatever rate it wants to. 

This is incredibly important when it comes to the AI race, especially given the fact that Google has a huge legacy cash flowing business to fund future expansion. Google is able to entice people into using their products by offering generous free tiers that allow them to collect massive amounts of data for training/fine tuning of models, which in turn allow Google to make huge improvements on their product lines. 

We've seen this recently with their image/video generation models, as well as Gemini - specifically when it comes to coding. Google offered insanely generous usage limits for both Nano Banana (its image generation model),  Gemini CLI (a terminal based coding tool for developers), and to a lesser extent Veo 3 (its video generation model). We then saw profound jumps in capability for Gemini 3 (most notably when it comes to coding) as well as Nano Banana with it's recent “Pro” release.

Google’s cloud business is also now growing at incredibly rapid pace, which is in part fueled by the fact that it is the only cloud platform where you can integrate Gemini. This line of business had been holding Google back for quite some time, as they invested huge amounts of money into capex and hiring, but it has finally reached is “escape velocity” and is growing at a very impressive rate, with Cloud revenues growing by more than 33% on a year over year basis in the most recent quarter. GCP remaining performance obligations (RPOs) also nearly doubled on a year over year basis in that same time frame.

Lastly, we can't forget about the line of business that gets the least amount of love at Google, Waymo. This incredible business is the only fully operational robotaxi company that's operating at scale, providing riders with more than 250,000 rides every single week. Currently the service operates in 5 US cities, and is planning expansion in several more. While many robotaxi company investors have thought that the limiting factor will be production, at the moment it seems like legislation and bureaucracy are the limiting factors, which is great news for Waymo.

Disclosure: We have a substantial Google position in our portfolio with an average cost basis in the ~$150s.