Welcome back to earnings season. Netflix and Taiwan Semi reviews were published this week. Alphabet, Tesla, ServiceNow and more earnings coverage is coming next week.
Over the next several weeks, subscribers will get 40+ earnings reviews, ongoing fundamental news coverage and real-time updates of my own portfolio/performance.
For the next 48 hours, we are running an annual plan discount to give you access to all of this for less. Lock in your lower rate below:
Table of Contents
1. AI Industry Developments – Moonshot Model launch
Last week, I published a piece on AI Industry developments. In that, we worked through the closing performance gap between open source and closed source models and the potential implications for research labs like Anthropic and OpenAI. This week, that became even more relevant, as Moonshot’s Kimi K3 model (based in China) is now performing on par with Anthropic’s and OpenAI’s latest products in many areas. Many believe the product’s strength comes from distilling leading U.S. models. Regardless of whether or not that’s true, the performance here is notably excellent and changes things. Kimi K3 took a massive leap forward compared to Kimi K2.6 that came before it. It’s also only about 40% of the blended token price of Anthropic’s Fable model and much cheaper than OpenAI’s new models too. All of these models are also a lot more expensive than lower performing products from Meta and Alphabet. In other words, there is significant cost-pressure in AI land.
It’s intuitive to think all of this will threaten the pricing power of the two leading research labs. On one end, Meta and others are aggressively undercutting leading frontier offerings on token costs to take market share on tasks and projects that don’t need the absolute best model. On the other end, Kimi K3 now provides a cheaper alternative with similar performance compared to Anthropic and OpenAI. Less expensive AI consumption is great news for infrastructure companies, as Jevons paradox tells us cheaper compute would mean a larger usage tailwind than the coinciding pricing headwind. This would also be great news for software companies that use this technology and even the cloud giants.
While those companies would also suffer from lower token pricing power, they’d also sell a lot more compute from the higher volume of apps and workloads that naturally stem from token deflation. Cheaper access to technology has been uniformly good for cloud giants for most of this century, and I think the same will be true in this AI age too. These companies can also offer Kimi K3 and other leading models to ensure their cloud environments are optimally compelling. OpenAI and Anthropic don’t have this luxury. I think they’re the losers here. They’re in the business of making money on token consumption, with pricing power obviously a big part of that equation.
2. Axon (AXON) – Competitor Stubbing Their Toe
Flock Safety, a major Axon competitor, is seeing a string of customer departures. The LA Police Department let its contract expire due to a 30%+ error rate in stolen vehicle identification. The department also cited the same improper data sharing practices, with Flock supposedly sharing sensitive data with other agencies without permission. According to Military.com, Flock has now seen more than 80 police departments cut their contracts since 2021, with over 40 of these coming in 2026. Whether it’s Dayton, Evanston, Austin or other major municipalities, this is opening the door for more responsible and more effective vendors to take more market share.
Axon is the company that is best positioned to capture this opportunity. I think this will push customers to companies that take privacy seriously and have pristine track records of blending safety and responsibility... and that means they'll pick Axon.
Axon’s fixed camera offerings like Lightpost and Outpost paired with its drone suite and real-time crime center create an end-to-end service for (successfully) identifying stolen vehicles and protecting cities in so many other ways that Flock can’t match. Axon is also obsessed with data sharing transparency and doing things the right way. Its Ethics and Equity Advisory Council acts as a compliance check to make sure its innovation is always responsible and by the book. This is why they’ve never lost a major customer contract while Flock is losing big customers on a weekly basis. In my conversations with Axon, this was cited as a small but real boost to demand and overall momentum in the past. And it sounds like the tailwind is intensifying.