Cocoa Jumps as Market Weighs Hershey Ask for Huge Exchange Buy

Bloomberg (1/09/2025)

Bloomberg reported that Hershey wants to take a position that will allow it to purchase more than 90,000 metric tons of cocoa on ICE Futures US, citing people familiar with the matter. Supply chain disruptions continue to impact businesses globally. Despite derivatives exchanges being hundreds of years old, contract volumes continue to grow with a wider range of applications for an increasingly globalized and complex world. It is remarkable to think of all the latent incremental volumes that could still accrue to the world’s various financial exchanges. 

The LA county wildfires could be the costliest in US history, early estimates say

AP News (1/11/2025)

First of all, my condolences and prayers go out to everyone that has been affected by the California wildfires over the last several weeks. As the headline shows, this has been a significantly destructive event, and AccuWeather opined that the total damages and economic losses could be more than $250 billion. These losses come not long after Hurricanes Helene and Milton on the east coast, which were also historically costly. The cadence of these costly disasters could slow the rate at which reinsurance capacity comes back into the U.S. and consequently extend the current hard market.  

What is DeepSeek, the Chinese AI company upending the stock market?

AP News (1/27/2025)

Monday morning, as AP News points out, the stock market was being “upended” by reports of new advances in AI out of China. Some people view this as a postured Chinese attack on American AI leadership. Some think that the details have been falsified. This thinking would be missing the forest for the trees. 

A few things: 

“Upended” is being a little dramatic. With increasing index concentration and strong performance within momentum assets, many investors have been exposed to one large factor bet. While NVIDIA and other AI capex beneficiaries were down upwards of 20%, a good portion of the market was left little changed. It was only one of hundreds of days, but it did show the potential benefits of being benchmark agnostic. We strive to ignore short-term distractions and keep our focus on the long term. Conversely, many other managers feel pressured to own these companies and were scrambling to provide explanations. Truth be told, I am not sure how well they truly understand the technology or the industry structure. (My undergraduate degree in math is probably more applicable to the subject than an MBA, but is still of little use in deciphering the whitepaper.)

And while I typically think that it is best to ignore the flashy headlines, DeepSeek may have serious implications. Was the Chinese LLM really developed for just $6 million? I am not sure. I will not pretend to be an AI expert. Rather, my opinion is that the main takeaway has very little to do with this specific DeepSeek model at all. If DeepSeek was banned tomorrow, I doubt it would change any of the long-term implications. The cat is out of the bag.

Instead, there is now the idea that there are algorithmic improvements that could be made to make model training less computationally expensive. OpenAI has largely led the world to believe that infrastructure is the largest bottleneck in AI. This perception was to OpenAI’s benefit! Imagine the incentives OpenAI has to say that they have done all that is possible on their end, and implying that competition is far behind. But now, you could reasonably question if the bottleneck is instead in the inefficient training architecture of the models, requiring less demand from infrastructure. (OpenAI looks less impressive in this scenario, but again, I am not an expert.) Preconceived notions for the investment path forward in AI development are being challenged. 

This shouldn’t be that big of a surprise. Every industry has a capital cycle; it’s basic economics. Some industries have short and violent cycles, while others have long and gentle cycles. New industries can have unpredictable capital cycles, as this piece of news has aptly demonstrated. Capital inflows typically drive down excess returns for incumbents, but this effect can also manifest as efficiency improvements through research and development. The market is clearly aware of this and reacted accordingly. As to whether it was too much or too little, I do not know.  

I found this memo from the Special Competitive Studies Project to be both very helpful and direct regarding potential implications.