50 AI Predictions for 2026 - Part 2

Today on the AI Daily Brief, part two of my AI predictions for 2026, the AI Daily Brief is a daily podcast and video about the most important news and discussions in AI. Alright friends, quick announcements before we dive in. First of all, thank you to today's sponsors KPMG, Blitzie, Super Intelligent and Robots and Pencils. To get an ad-free version of the show, go to patreon.com/airdailybrief, or you can subscribe and Apple Podcasts to learn about your sponsorship options. Set us a note at [email protected] Alright, so yesterday we did the first part of my AI predictions and because it got so long we had to split it into two. Today we are jumping in in the competition section where the AI race is headed into 2026. For our next set of predictions, let's talk about competition. My first prediction is that I think it's going to be very hard to shake and throp it off its coating lead. They are incredibly focused on it, they have held that perceived lead for more than a year and a half, which is approximately 100 years in AI time. And you can absolutely tell that an increasingly large number of developers are so comfortable with the cloud models that even when open AI or Google or in the future groc release something that is comparable or ahead on the benchmarks, it will be very unlikely that they switch. Now that won't be universally true, there are still lots and lots of room and so I think it makes sense for everyone else to compete for coating still. I just think it's going to be hard to shake and throp it off their lead. One specific prediction is that I think that this will lead ultimately to Microsoft doing a big deal with them, expanding that relationship even from the very nascent ways we started to see them get together this year. I think Microsoft will want to bring even more aggressively and tropics coding tools into their enterprise suite and I wouldn't be surprised to see that happen in 2026. Now for open AI, I think their big challenge is going to be fragmented attention. How much do they focus on consumer usage versus enterprise usage versus research that gets them to AGI? We've recently gotten a bunch of reports from inside open AI that not everyone there is super happy with the allocation of resources and are worried that all of these consumer things like Sora to are side quests that are actually distracting them from fundamental AGI research. Certainly open AI itself has talked about the difficulty of having to make hard decisions about where to allocate compute. I do not believe in 2026 that open AI is going to seed any of these things, however I do think that there will likely be some recognition that where they are dominant is on consumer. To many people, chat GPT still is the same as AI. These are synonymous terms. And even as Gemini has searched, chat GPT still has a commanding lead. My guess is that to the extent that open AI has to defend one flank, that'll be it. Now one additional specific prediction is that I just believe that it's inevitable that ads will come to chat GPT. If for no other reason, then if you look at charts of intent and how users that are sourced from an LLM interact as opposed to people coming from Google search, it's just a really good medium for advertising. I think open AI is going to continue to face questions in the market as to whether they can actually reach their heavy revenue targets and given that they have so much tied up in whether the market believes they can when it comes to funding all of this infrastructure, I think they're going to have to do some version of advertising. How well they do that and whether it causes a user rebellion will be key things to watch for in the coming year. Now let's talk about GROC. I have occasionally bit accused by some of you of not giving GROC its due. Some of you have even secretly suggested that I am an Elon Muskator, which is not true. The reason that GROC has come up less is that across all of the different use cases where I try all the different models on basically everything I'm doing, I have yet to find a single use case where I regularly prefer GROC's output compared to Gemini, ChatGBT or Clawed. Every other one of those models has something I prefer better than the others. However, I still pay for GROC's most premium subscription and I think if you judge GROC on the speed with which they have become a contender and are putting out models that are very near state-of-the-art if not there exactly, you cannot write them off as a contender in the AI race. That's more, Elon is in a unique position to raise capital and apply compute and you could see a compute investment-based leapfrogging. It's also possible that external factors like political pressure and constraints on others that Elon just decides to ignore could also create a raise on that for GROC in a way that I don't see right now. I basically just don't know who the natural user is, other than someone who's already on Twitter, which to be fair is not an insignificant number of people. One good thing in GROC's favor is that if you look at OpenRouter who GROC have partnered with to offer free tokens at various points this year, it is very clear that users have shown that they will use GROC if it's free or cheap, which by the way is not a knock on GROC. It's still early enough that people are willing, in many cases, to use a more expensive model than a less expensive one if it actually gets what they want done. So the fact that GROC has done so well on OpenRouter with those promotions is actually indicative of where it could go. But ultimately, if over the course of not just 26, but more like 26 and 27 and maybe even a little into 28, if it can't really differentiate itself and find some breakout that gives it a space relative to these others in the competition, I wouldn't be surprised if we saw some mass absorption of the Elon Empire all under, for example, the banner of Tesla. Next up, I think in 2026, we are going to see Meta re-enter the conversation, although I genuinely am at a loss for exactly what their direction will be. We have heard rumors that their next model will be closed source, but I have no idea what their spot in that is. I think what is clear is that where they will have strength is one, applying AI to social networks, particularly in terms of how it helps their ad products. So actually monetizing AI. And two, the success of the meta-ray bands is undeniable. As of this moment, Meta has the only AI-related wearable that anyone actually likes and people really like it. That could create momentum for an entire platform play that moves into other form factors as well. Next up, I think in 2026, the growth in Chinese open-weight models that we saw throughout 2025 is going to continue in a massive way. Already you have tons of US and Western startups turning to Chinese models for certain key workflows where what they need is efficiency not just the purest state of the art. If China stays on the trajectory that they are on and especially if they take advantage of these new H200s that the US is letting them purchase, I would expect to see them make up an even bigger share of a production tokens consumed in 26 than they did in 25. Next prediction is that one of the most interesting competitive battles of the next year will be the agent labs versus the model labs, which I articulated a little bit before. Now, as I mentioned, I think the line between these two types of companies is getting a little bit blurrier and where we're going to see this competition play out, maybe in the context of new models. This one is one I debated. Of whether any of these big agent labs like cursor cognition, rep litter lovable, will get bought this year. The reason to think no is that these companies to an insane degree control their own destiny. They are all making a ton of money, their high growth, their beloved by their users. So none of them have to sell. I also don't really see among any of the leaders of these companies, people who are super keen to take their chips off the table rather than keep building. However, I feel like at least one of these agent labs gets an offer that is too big to ignore ultimately. With Microsoft in my mind being the most likely acquirer, another area of M&A that I feel like is more high probability, I think both Gen Spark and Manus get scooped up this year. Like I said, I think these labs are going to realize that they need better interfaces for agents and they're going to look over and see that both Manus and Gen Spark have built really compelling starting points for that that are more performant than the things they have internally, even if they're using those companies models, and they're bringing with them a nice chunk of revenue. The reason I think these companies are going to be willing to be bought is one, the astronomical price that will be paid, but two, a concern that they inevitably are going to get competed out of the way by the model labs over the course of the next few years. Next up, we turn to markets, especially in the second half of this year, public markets really started to interact with AI development in a major way. We moved from the unbridled enthusiasm from basically November 2022 when Chatshipy T launched all the way up to honestly around September of 2025 and then saw a mass repricing as public market investors just started to really rethink the fundamentals of just how much growth there could be. Indeed, I think that that recalibration sets us up for the most important question for AI and markets in 2026, which is, will there continue to be private credit appetite for data center and AI infrastructure build out? So far, as expensive as the AI build out has been, it's all been financed or mostly been financed by the hyperscalers balance sheets. Now market analysts might have had to reset their brains when companies started spending their free cash flow on anything other than stock buybacks, but one of the best dampers to the AI bubble narrative so far for a long time had been the fact that this was all being financed by the companies themselves. That has, of course, shifted over the course of this year. A growing portion of the infrastructure build out is being financed and at the moment, private credit markets certainly have enough to keep that party going for some time. The question is whether they want to. We recently saw Blue Owl pull out of an Oracle financing deal, although Oracle has a different narrative for it, and that has made many wonder if we're going to see more of that next year. I predict that markets are going to be extremely on edge about absolutely any wobbles and data center financing, but I also tend to think that we're coming into 2026 in a much healthier spot with markets than we were last year. The repricing that has happened, the new appreciation for risk that is prominent, has a strong likelihood of tamping down enthusiasm before it gets over exuberant in a way that I think is likely healthy. Which is not to say that things couldn't rip in one direction or another. I think in fact that a lot of AI's market destiny is going to be tied to broader macro factors. This has always been the story with AI, even when we haven't recognized it. This is perhaps best expressed of the virality of this chart showing the K-shaped economy. The chart has two lines, one is for the S&P 500, the other is for total job openings. From 2005 up to around 2023, the lines move in pretty lock step. Since that inflection point, however, total job openings have gone down, while the S&P 500 is continued to go up. Now, most versions of this chart have a dotted line with some version of the text, what happened on November 30, 2022, with of course the implication being that the launch of Chatchy BT is the reason for those total job openings declining. However, as any macro commentator can point out, where the line actually shifts is not November when Chatchy BT launches, it's a few months earlier when the rate hiking cycle begins. This huge spike you see in 21, coming after the big dip you see in 2020, is the COVID era zero interest rate period, where everything just ripped up to the right. There was significant overhiring during that period, and a big part of the story of the job openings declining is a recalibration of that, and another part of the story, which you can see where the lines actually dipped, is rates going back up, and inflation soaring off to the moon. Now, this is not to argue that AI has had no impact on job openings, but it is a good reminder that AI's market story has always been tied up with the macro in complex ways. For so much of the 2022 to early 2025 period, it was AI enthusiasm versus everything else. AI versus the hiking cycle, AI versus the tariff tantrum. You name it, AI just kept winning ultimately. It wasn't going to last forever and it didn't last forever. And that is led to where we are today. If you've been enjoying what we've been discussing on the show, you'll want to check out another podcast that I've had the privilege to host, which is called You Can With AI, from KPMG. Season one was designed to be a set of real stories from real leaders, making AI work in their organizations, and now season two is coming and we're back with even bigger conversations. This show is entirely focused on what it's like to actually drive AI change inside your enterprise, and as case studies, expert panels, and a lot more practical goodness that I hope will be extremely valuable for you as the listener. Search You Can With AI on Apple Spotify or YouTube and subscribe today. This episode is brought to you by Blitzie, the Enterprise Autonomous Software Development Platform with Infinite Code Context. Blitzie uses thousands of specialized AI agents that think for hours to understand enterprise scale code bases with millions of lines of code. Enterprise engineering leaders start every development sprint with the Blitzie platform bringing in their development requirements. The Blitzie platform provides a plan that generates and pre-compiles code for each task. Blitzie delivers 80% plus of the development work autonomously, while providing a guide for the final 20% of human development work required to complete the sprint. Public companies are achieving a 5X engineering velocity increase when incorporating Blitzie as their pre-IDE development tool, pairing it with their coding pilot of choice to bring an AI native STLC into their org. Visit Blitzie.com and press Get A Demo to learn how Blitzie transforms your STLC from AI assisted to AI native. Today's episode is brought to you by my company Super Intelligent. Super Intelligent is an AI planning platform, and right now as we head into 2026, the big theme that we're seeing among the enterprises that we work with is a real determination to make 2026 a year of scaled AI deployments, not just more pilots and experiments. However, many of our partners are stuck on some AI plateau. It might be issues of governance, it might be issues of data readiness, it might be issues of process mapping. Whatever the case, we're launching a new type of assessment called Plateau Breaker that, as you probably guessed from that name, is about breaking through AI plateaus. We'll deploy voice agents to collect information and diagnose what the real bottlenecks are that are keeping you on that plateau. From there, we put together a blueprint and an action plan that helps you move right through that plateau into full scale deployment and real ROI. If you're interested in learning more about Plateau Breaker, shoot us a note, contact at bsuper.ai with Plateau in the subject line. AI changes fast. You need a partner built for the long game. Robots and pencils work side by side with organizations to turn AI ambition into real human impact. As an AWS certified partner, they modernize infrastructure, design cloud native systems, and apply AI to create business value. And their partnerships don't end at launch. As AI changes, robots and pencils stays by your side so you keep pace. The difference is close partnership that builds value and compounds over time. Plus, with delivery centers across the US, Canada, Europe, and Latin America, clients get local expertise and global scale. For AI that delivers progress, not promises, visit robotsandpensals.com/AIDailyBrief. Now, next year, there's going to be a lot going on. We're going to have a new Federal Reserve Chair and the White House presumably exerting a lot more influence on the Federal Reserve and rate policy. We have all the dynamics around the midterm elections. So almost outside of anything that actually happens in AI, a lot of what the market thinks about AI is going to be shaped by totally unrelated factors. Now, let's talk about the IPO market. I have such a hard time making up my mind for what I think is going to happen. My base case would be that we do not see any IPOs in 2026. I think we see both OpenAI and Anthropic in 2027. I think private markets are going to be strong enough, and the other pain in the butt that is public markets is going to be a big deterrent. Startups have for some time wanted to stay private for as long as humanly possible, and I see no reason why it would be any different in this area. In fact, I see even more reasons why it wouldn't be different in this area. However, there are a lot of factors that could change that. One is OpenAI's massive capital requirements. That company is already basically shaking the couch cushions of the entire world, and it is not impossible that they get to a point where they feel like they just have to go pursue a public market strategy at the same time as everything they're doing privately. Indeed, if they get even a whiff that private markets are getting less enthusiastic or might not have enough capital for their needs, they might try to get out ahead of that shift before it becomes a broader public narrative. There's also the fact that if current reports are to be believed, OpenAI is going to IPO at a trillion dollar plus valuation. We don't really have precedent for that, and it's not clear how much higher they could be valued in private markets and still have a successful public offering, so that might also force their hands sooner. There is also the possibility that if they think that Anthropic are trying to lap them, in other words, to get out ahead and take advantage of public enthusiasm and retail investor enthusiasm for AI, then it might move up everyone's time scale. I kind of think that the inverse is less true in the sense that I don't think that OpenAI moving faster would necessarily change Anthropics plans. However, to the extent that they want to get in ahead of OpenAI, which has a lot of reasons to consider, that could push OpenAI to move faster. So like I said from a prediction standpoint, my base case is 0IPOs in 2026 and 2 in 2027, but there are so many things that could change that and make me wrong. Another kind of simple prediction, I think that sometime in 2026, Alphabet will become the biggest company in the world. I'm not sure for how long they have that title and whether it's sustained, but I think especially if the growth trend in Gemini continues, growth in Google's cloud business continues, and especially if they actually start to sell TPUs broadly outside of the Google organization, we'll see them jump to the top. This could also be caused by a stumble from Nvidia, but I think that's less likely than a surge from Google. I think we're going to see the rise of artisanal anti-AI next year. In other words, we're going to see explicitly AI-free products, networks, and services. We're going to see the human-made label become a luxury symbol. We might even see certifications for 100% human-generated art writing. You name it. Now, the big question will be whether these things are small or whether there's actually some major network that decides to go this way. Could the next social network be one that aggressively works to prevent AI bots? It wouldn't shock me. Last category of predictions is politics and policy. Right now, I think it's a little bit easy to overstate and overestimate how big a deal data centers are to people. In fact, outside of our little bubble in the AI world, when you look at polling of how much people actually have opinions on data centers, it's very low. However, in the areas where data centers are being built or have been considered, it is a very big issue. It is also an issue where it's very clear to me that politicians on both sides of the aisle, with a particular bent towards populace on both sides of the aisle, see it as a winning issue and specifically being against it as a winning issue. It is very clear every political pundit will tell you right now that the number one issue heading into the midterms is going to be affordability. This is the word that we've chosen instead of economics to reflect the fact that what we're talking about is the specific difficult economics of living in America today. The anti-AI narrative writes itself, the robots are taking your jobs, your energy, drinking your water and raising your power bill, let's fight against them and your lives will be better. It will be extremely telling, I think, in 2026 how resonant this message actually is. Is it a winning political point in general or only in those areas that are specifically dealing with data centers? I think the answer to that question will have a huge impact on the 2028 presidential elections. I think regardless of the answer to that question, however, pretty much all layoffs in 2026 are going to be attributed to AI. Now in some cases that will be true, a thing that humans used to do will be more easily or cheaply done by automation that will come with restructuring and job loss. In other cases, AI layoffs will be attributed to AI, but it will be more indirect. It will be presented as a company reorganizing and repositioning for an AI future, where they need fewer people to do the same amount of work. We saw Amazon use this type of language a lot this year, and this is where it starts to get a little bit blurry. It's not crazy for an organization to be rethinking its broad organization structure in the context of AI. In fact, it would be crazy not to, but I think that you're going to see a lot of layoffs that would have happened anyway, be lumped under that or getting ready for the AI future kind of banner. And then, of course, there will be other layoffs that have nothing to do with AI that are just blamed on it because it's a very convenient scapegoat. All of this will, of course, fuel that political narrative that I was just talking about. And I think from a public perception perspective, 2026 being kind of rough. Now speaking of the 2028 presidential election, even though next year is just the midterms, and even though you will likely not get most campaigns formally announced until after the midterms, I think as part of the whole data center politics, as well as AI layoffs conversation, you're going to start to see a lot of policies that would naturally fall out of the AI transition, field tested as language in these campaigns. I think you're going to start to see more conversation about things like UBI. You're going to see more social safety net conversations for the democratic side of the aisle who are already pro safety net. This will be a little bit about figuring out which version of these things both makes the most sensitive as well as is the most politically palatable to independent and swing voters. Republicans in the right have a more interesting and challenging discovery process. It's very clear that the AI antagonist position will not just be the domain at the left, but will the right actually have an alternative solution? Or will they focus on trying to stop AI in their tracks? What's more, how does the protect versus the economic populist side of that party reconcile heading into the post Trump era? Moving off of mainstream politics a little bit, I think we're going to see basically the equivalent of fair trade AI labels start to become a thing, particularly in the entertainment space. As much as many artists would like to put AI back in the box, there is a tension. Lots of creatives are using these tools and creating incredible things and getting excited about the possibilities. And even if they're not, there is a sense of inevitability here. So how do you resolve all that? The answer of course is some sort of ethical certification around how AI was trained or how AI is used. You're already starting to see this in Hollywood, particularly I'm thinking of Moon Valley, which promises AI video without the same ethics and copyright infringement concerns. And I think you're going to see more of this now, it won't be a straight line path. If you read the response to Moon Valley and Asteria and the artist and filmmakers who have chosen to go this path, it is still a tough road to hoe. You're taking on an additional burden and challenge and not using the state of the art, while also still having a big part of your industry treat you like a scab. And I think that in 2026, you're going to start to see these types of projects carve out an important space that creates a path forward that is neither rejectionism nor head in the sand ignoring the challenges. On the China side, as much as they're heading and hog about whether to accept H-200s, I think China's going to accept the H-200s. Now, maybe that the CCP completely controls the flow in and works really hard to create incentives for domestic chip manufacturing regardless. But I just don't see any way that they're going to deny themselves access to the most premium chips that they can get their hands on right now. That is unless Congress passes some limiting legislation on exports to China. This wouldn't surprise me. I could see this being a politically palatable way in a very short term to reject a specific Trump AI policy in a way that can be turned into a positive economic narrative for constituents. Now, we've already got some legislation in the works around this area that doesn't necessarily seem like it has huge momentum, but we're also in basically the quietest period of the congressional cycle, and so I would watch in January to see if that starts to pick up. Lastly, on the politics side, I think that the big, bombastic policies like Bernie's data center moratorium doesn't really go anywhere. There's just too much against it, and I don't even necessarily think that Bernie thinks it's possible. I think that he's trying to shift the Overton window way back to the other side so that what does come out of it looks a lot closer to what he and the people who think like him want than it would have had he not proposed a dramatic policy like that. I think that the federal preemption that we got via executive order is going to be extremely difficult to put into practice. In fact, I think it's probably going to have the exact opposite effect where basically states now have even more incentive to go past AI rules as a way to assert their independence. Until you have a situation where all of these states are writing more rules and making headlines because of it, which my guess is actually shifts the balance a little bit among the AI lobby, and instead of enjoying the conspicuous lack of federal regulations, my guess is that you'll see the Washington offices of the big labs start to shift and support some basic national level rules. As much as we pretend there's not, there was a ton of space between the safety of extremes on the one end and the acceleration is to extremes on the other where a huge mass of people can agree on some common sense starting points. And I would expect those not to necessarily be put into practice in 2026, but to start to see the labs actually support a version of them. So that is my AI predictions for 2026 with just one last one to go. I forgot to put this one in the list. It's probably the easiest one that I'll make. Both Chatchee BT and Gemini will claim a billion active users in 2026. For Chatchee BT, it'll be in Q1. For Gemini, I think it'll be in Q2, but it might be a little bit blurrier because of how distributed and all over the place, Google's AI products are. But in either case, I think that in the first half of 2026, latest Q3 for Google, both of those companies and AI products will claim a billion users. So that is my list. Hopefully precise enough for you guys to have some debates about. And if you stuck with me through both of these episodes, you are a true AIG. For now that is going to do it for today's episode, appreciate you listening or watching. As always, until next time, peace. [BLANK_AUDIO]