🚨Opening Playbook is live at 10 a.m. ET🚨
I’m running the “Why I don’t use stop losses” training today. In my opinion, this is one of the single most important lessons any options trader can learn [tap to join us for Opening Playbook]
You know what’s funny about the whole AI-is-getting-cheaper narrative?
On paper, it makes perfect sense. Companies roll out newer, smarter models, then leave the older, less expensive models available for people who want to save money.
Perfect solution, right? Cost efficiency unlocked.
Except I’ve yet to find myself in a situation where I want to use the older model. I get access to the new model, and I want to use that. That’s the instinct. And it’s the same for most people.
What makes this even more interesting is how these rising AI costs are now showing up in broader economic conversations.
They’re becoming part of the inflation story alongside geopolitical uncertainty because, as demand shifts toward the highest-performing models, the real-world cost of running them pushes higher across entire industries.
And we’ve already seen where this behavior leads. A few companies found out the hard way that giving people AI tools without usage limits can be dangerous for the bottom line.
Some developers even ended up with multimillion-dollar bills because their applications kept calling the most powerful model by default.
The technical path to lower costs exists. The behavioral path doesn’t.
The Real Cost Problem
There’s undeniably a progression toward lower costs. You can see it in model updates, infrastructure improvements, product releases and development roadmaps. The industry keeps introducing more efficient options.
But at the same time, pricing structures are becoming more layered and more complex.
Companies are introducing different model tiers, context window pricing, new billing categories, usage-based pricing, fine-tuning costs and different inference modes — all designed to balance demand for the highest-end models with the need to protect margins.
The problem is that even with all this complexity, people still want the newest version.
They want the highest accuracy, the strongest reasoning, the best results — and pricing follows that demand.
That’s why cost savings often show up only in theory, not in practice. The cheaper options exist, but the appetite for them doesn’t.
What This Means for Traders
If you’re trading AI names or anything tied to that ecosystem, this behavioral mismatch matters.
The bullish narrative often leans on the idea that costs will naturally fall as the technology matures. And technically that’s true.
But if users keep gravitating toward the most advanced models regardless of price, those savings may never show up where investors expect them.
That means companies remain tied to expensive compute, costly inference and premium infrastructure because that’s what users demand. Instead of reducing expenses, greater adoption can actually amplify them.
So when you hear that AI costs are plummeting, the real question is whether they’re falling where it actually counts.
There’s a big difference between models becoming cheaper and people choosing to use the cheaper models.
Right now, most of the progress is happening on paper, not in behavior.
Now don’t forget to join us at 10 a.m. ET weekdays for Opening Playbook, and at 3:30 p.m. ET Closing Playbook!
Nate Tucci
Tucci Trades
You can also follow along and join the conversation for real-time analysis, trade ideas, market insights and more in my official Telegram channel!
Important Note: No one from the New Money Crew team or Tucci Trades will ever contact you directly on Telegram.
*This is for informational and educational purposes only. There is inherent risk in trading, so trade at your own risk.Â
