AI and Lottery Tickets - Is it Time to Gamble?

Humans love gambling - on sports, elections, wife carrying, and even cheese rolling. Although there is a darker side to gambling, it’s big business…and unfortunately growing. Draftkings is a $11B company, several startups have raised material rounds (to bet on new things), and casinos are increasingly opening in new states. 

I can count the number of times I’ve gambled in a casino or online on one hand. From winning a couple dollars on penny slots with my dad in Wendover, NV many years ago, to losing $20 in a few minutes on slots in Biloxi - I’ve had very little interest in gambling. And, I only lost $20 in Biloxi because it was lunch time, and there were few non-casino food options. In fact, I was happy to lose that $20 so fast…it was an incredibly depressing place (the food was not bad though).

While traditional gambling holds no allure with me, I was a huge trading card collector as a kid. I regularly dragged my parents to card shows, religiously studied Beckett, and even had a nice side hustle selling my cards at yard sales. Since I was mainly using allowance money (and any gains from good sales), I would often go table to table and try to negotiate on packs. I was also fairly cheap even back then, and preferred the $.50-$1 packs. If I had more pocket change, I’d splurge on a Fleer Flair, Topps Finest, or Fleer Metal pack. Although I was in elementary school, I can still vividly remember the card shows - and the good memories. 

I heard about 2024 Topps Chrome Update in October, and figured I’d buy a box - for nostalgic value, and the (incredibly remote) chance of pulling the hottest card since the Lebron logoman. At ~$200 per hobby box, it seemed like an entertaining half hour - with the chance of a huge hit…and likely losing most of the $200. Sales opened at noon, and I figured I’d buy the box around 11 PM…I was 10 hours and 58 minutes too late. All boxes sold out in less than 2 minutes. 

From October 17th to mid November, the hype train really picked up momentum. The Pirates themselves stoked the flames, and online breakers quickly ramped up the hype train - triggering a lasting spike in prices. Hobby boxes quickly rose to $300, then $400, then $500, and peaked around $1100 in January 2025. In three months, box prices jumped by 5.5x - even after most hobby boxes yielded generally terrible results. But, the hunt for the chase card grew - and people paid up (big time) for their lottery ticket. 

I could write about trading cards in every newsletter, and never run out of interesting topics (at least to me). However, in this case, I’ll finally make the tie-in to venture investing:

In the 2020-2021 bubble, I often saw companies (especially later-stage ones) raising an overvalued, hyped round - then turning around and raising an even crazier round 6 months later. It was almost like they were on a conveyer belt that steadily marched upward, with no break and no chance of anything falling backward. And in many cases, the companies were burning ungodly amounts of money to grow - but investors only cared about the growth (not about the unit economics, efficiency, or eventual exit). Predictably, most of these companies faced enormous headwinds once the capital markets slowed down, and they had to kick their addiction to easy money. I thought the rough realities of write-offs, massive write-downs, and bad fund returns would sober up investors. However, it feels like sobriety was short-lived. 

I’m getting inundated with emails enticing me to invest into the hot unicorns again, alongside many of my own investors. While companies like Databricks have actually hit their own projections, and some unicorns are interesting, many AI-specific companies are a different breed. Almost half of all new unicorns in 2024 were AI companies, and the trend/momentum accelerated in 2025. In most cases, it follows the 2021 playbook (this is just a representative example):

• Raise $400m at a $2B valuation in September

• Early investors (from a year or two ago) sell preferred shares (lower in the preference stack) at a $2.6B valuation in November

• Employees sell common shares at a $3.2B valuation in December

• The Company raises $1B more at a $4B valuation in January

Has anything materially (enough to double the valuation) changed with the business in 5 months? Doubtful. Have they landed multiple monster customers, created even $500m (of the $2B valuation change) in value, etc? Also doubtful. Are common shares (below $1.4B+ in preferred shares and management bonuses/earn outs) worth more than preferred months after a raise? Dubious at best. 

The same forces that propelled many companies into unicorn status - and also propelled many unicorns into decacorns in 2021 is back again in AI startups. While it’s true that valuations only feel like they’re rising, and may actually pick up more steam in 2026, these companies eventually need to exit. And unlike a hobby box of Topps Chrome, you cannot easily sell to one of the major online card shops (or on eBay).  

While Topps Chrome had its own conveyer belt moment, the Skenes card was pulled by an 11-year old — and the chase ended. Hobby boxes are still somehow above the pre-release price, but are down roughly 50% from the peak. Although many of these private unicorns likely have more time to run upward, many will eventually crash back down to reality. Chasing momentum and gambling is a fun game - but an inherently risky one too. In most cases, the house wins - and you’re left with losses and lessons learned. I thought more investors learned their lessons in 2020-2021, but once you catch the gambling bug, it’s hard to stop.