Hey Reader,

Hope you had a wonderful holiday season! ☃️ As we head into 2026, it's time to prep for another year of rapid change in the AI industry. The big picture:

• Models and software will improve and get better, faster and cheaper...
  ​
• But the core tech underlying large language models remains the same, so all the knowledge you've built so far is still valuable...
  ​
• And remember that half of Americans have heard "a little" or "nothing" about AI! 🤯

In yesterday's live training session for AI Consultancy Project students, we chatted about these big 2026 trends:

#1 → OpenAI will lose its lead as "best large language model"

ChatGPT is basically a household name and is likely to become the "Kleenex" or "Xerox" of AI – the product whose brand name stands in for a whole category.

However, most people don't know the name "OpenAI" (outside of our AI-enthusiast circles, that is). And OpenAI is fundamentally a pretty small company that hit on an amazing discovery. They'll always get credit for that, but they are not guaranteed to be the leader in the LLM space forever.

In fact, Google has already almost overtaken OpenAI with the release of Gemini 3. Elon Musk's Grok models are excellent. Anthropic (another small company) already outperforms OpenAI models at many tasks, particularly writing. Amazon is a very large and well-funded dark horse in the race. And I haven't even mentioned Apple and Meta, who seem like they're a little behind right now but obviously have the means to make a breakthrough.

In short, the 2023 warning that There Is No Moat around LLMs is coming to fruition this year.

That doesn't mean OpenAI's valuation will suffer. They might even IPO this year and take in a ton of cash. But if I were an OpenAI shareholder I'd be worried about my 10-year outlook, which is plenty of time for Google or another big competitor to permanently take over the top spot.

For consumers of AI tools and tokens (like you and me!), this competition is a good thing. Three or more major companies will always be competing for our money, which means we'll get more for every dollar we spend on tool subscriptions or tokens. And there will be a huge economic incentive for these companies to make everything better, faster and cheaper.

Contrast this with the 2010 startup business model. The goal of companies like Amazon (when it comes to e-commerce), Uber and DoorDash was explicitly to keep prices low while they were building market share, and then later to be able to act as a near-monopoly, raising prices only after they reached a point where it was hard for users to change. (Note that in the case of DoorDash and Uber, this is also a multi-sided marketplace, making it even harder for consumers to leave once the marketplace grows to a certain size.)

This is not the case with the burgeoning large language model market. It is trivially easy to switch between tools, there is no multi-sided element (e.g., connecting drivers and riders), and "AI tokens" are rapidly becoming a pure commodity with minimal differentiation between providers.

That means better, faster, and cheaper for us – but also means it is totally unclear which company will be "top dog" next year.

#2 → AI will drive local, national and international politics

When AI image models first came out, there was a lot of fear that they'd be used to generate realistic-looking disinformation. While AI financial scams are a big problem, I think we can safely say that AI imagery has not meaningfully affected any elections or "tricked" anyone in a long-term way. In fact, we can definitely say that "human-made" conspiracy theories are still outperforming AI fakes. 😅

However, we have reached the point where AI expansion will affect politics in three big ways:

1. At the local level, energy and environmental effects of data centers will be a major point of contention. This already happened in Georgia, where a usually-obscure election to a local energy commission turned into a major upset, in part because of issues around AI data centers and energy prices.
   ​
2. At the national level, the 2026 U.S. Congressional elections will be about affordability, which also plays into energy prices and job loss. There probably won't be proof that AI directly causes job loss – but most companies are holding off on hiring, with the expectation that some jobs will be automated by year's end.
   ​
3. At the international level, the chip war is well underway. Expect to see more hemming and hawing from the American government on whether and how to restrict advanced chip sales to China and other adversaries. Meanwhile, expect Europe to lead the way on AI regulation while the U.S. does the opposite. Finally, keep an eye on the United Arab Emirates and its neighbors, who are making huge investments in AI as a long-term hedge against shrinking demand for oil.

And as this all happens over the next year, more consumers and business owners will become more familiar with AI, and pretty much every tool will get better, faster and cheaper – which means a lot of cool opportunities for folks like us, who are already engaged with learning and experimentation every day.

Until next time,

– Rob
CEO of Innovating with AI
​