2026 AI : 10 Things Coming In 2026 (A.I In 2026 Major Predictions)

2026 could be a pivotal year in AI. That's why in today's video, I'll be taking a look at all the statements from the top AI CEOs and giving you guys every statement that they've said about the next year in AI so you can get some insight on what to expect for next year. So firstly, we're going to take a look at what Dario Amade says. Dario Amade is an AI researcher and entrepreneur best known for co-founding Anthropic, the company behind Claude AI. Previously he was actually the vice president of research at OpenAI where he played a key role in developing GPT2 and GPT3. So one thing that Dario Ammed says about coding in 2026 is that AI coding may match the best human coders by late 2026 and that they will reach a very serious level by the end of 2025. for winter break as you

know as I was looking at where things were scheduled to scale within Enthropic and also what was happening outside Anthropic I looked at it and I said you know for coding we're going to see very serious things by the end of 2025 and 2026 it might be you know everything you know close to the level of the best humans and I think of all the things that I'm good at right you know I think of all the times when I wrote code and you know I think of it as like this intellectual activity and boy am I smart that I can do this and you know it's like a part of my identity that I'm like you know good at this and I get mad when others are better than I am. Um uh and then I'm like oh my god there are going to be these systems that you know and even as the one who's building this even as what one of the ones who benefits most from it uh it's there's still something a bit threatening about it. Yeah. Um I just think we just we need to acknowledge that like it's it's wrong to tell people that is coming or to try to sugarcoat it. Now Dario Amade also says that in 2026 and 2027 is likely to get us to AGI. Of course, this is not set in stone, but he does say that this is something that is rapidly approaching us and we should be

aware of this. So 2026 2027 is when you effectively get to AGI across the board and it's the threshold moment. Whoever's ahead then is ahead forever. Is that what you're saying? Potentially. I mean we don't know these things but there there's a risk of this that's happening at a particularly uncertain time geopolitically but we'll get to that. I've heard those same years you know floated as no I'm I'm aware of how frightening this is. Daria Ammed has actually said more about the capabilities for 2026 and 2027. Of course, in this interview clip with Lex Freiedman, he does provide context saying that, you know, things can change due to many of the things that happen in the field of AI, but this is where he gives some more information on those capabilities. If I say 2026 or 2027, there will be like a zillion like people on Twitter who will be like, "Hey, ICO said 2026, 2020. " And it'll be repeated for like the next two years that like this is definitely when I think it's going to happen. Um, so who whoever's exerting these clips will will crop out the thing I just said and only say the thing I'm about to say. Um, but I'll just say it anyway. Um, uh, so, uh, if you extrapolate the curves that we've had so far, right? If you say, well, I don't know, we're starting to get to like PhD level and last year we were at, um, undergraduate level and the year before we were at like the level of a high school student. Again, you can quibble with at what tasks and for what. We're still missing modalities, but those are being added. Like computer use was added, like image in generation has been added. If you just kind of like, and this is totally unscientific, but if you just kind of like eyeball the rate at which these capabilities are increasing, it does make you think that we'll get there by 2026 or 2027. Now

Dario Amade actually further spoke about this in his blog post, Machines of Loving Grace. And one of the things he said is quoted here, "Time is short and we must accelerate our actions to match our accelerating AI progress. Possibly, and this was something that was widely quoted, he said, possibly by 2026 or 2027 and no later than 2030, the capabilities of AI systems will be best thought of as akin to an entirely new state populated by highly intelligent people appearing on a global stage. the main thing that was quoted several times was a country of geniuses in a data center with profound economic, societal and security implications that would bring. And so yeah, 2026 and 2027 looks like a key pivotal year, especially for AI development. Now, I will add some context here. One thing that there is to be known about these AI companies is the fact that these companies, they are constantly seeking new investment. So they do have to say that, you know, this technology is improving. It's constantly getting better. So, one part of me does think that these companies are saying all of this stuff happens next year is because they want to continually get your investment. But another part of me thinks that it's not just hype. It's not just them touting their own horn about their own products. It's actually them looking at what the technology they have and basically extrapolating it out into the future saying, "Look, we know where this is headed. " Now, if we take a look at Elon Musk, this is the guy that founded OpenAI, founded his own AI company recently that has been moving at an insane pace. And of course, he leads Tesla, which is at the forefront of AI self-driving technology. And this guy's been speaking about AI for quite some time now. He made a recent prediction where he spoke about the fact that it's quite likely that we get to AGI by 2026.

Grock's coming to Tesla. So, you'll be able to access Grock through Tesla, through your Tesla, and you'll be able to ask it to do whatever you want. You could uh ask your car to go pick up uh a friend or anything you can think of it. It'll the car will be able to do it. Yeah, you'll be able to ask a Tesla to go pick up groceries, pretty much anything. Optimus is really going to be next level. You'll be able to uh skin Optimus in a white, you know, pretty much with anything. So, people on the internet have asked me to make cat girls. And actually, you can make cat girls real if you have a robot. You have a robot catgirl. Yeah. Optimus will be able to pick up the kids from school and Optimus will be able to be school if you want. Be able to teach uh kids anything. Yeah, it'll support any language too. I think AGI will be next year probably. It's not next year. It's let's say 2026 at the latest for AGI at the latest. I hope it's nice to us. So when defines AGI as smarter than any human, I think it's we're less than 24 months away from that. Yeah, please be nice to us, AGI. Humanoid robots will usher in a level of abundance that was hard to imagine. There will be no shortage of goods and services. Now, Elon Musk spoke even more about the Tesla bolts. And he says that the company plans to build about 10,000 this year, saying that they're going to be working in Tesla factories by the year end, and that data from these robots is going to be deployed on the second model, which is expected to launch in 2026. So we're possibly looking at another version of the Tesla bolt in 2026 that's probably going to be deployed widescale. Don Musk speaks even more about AGI and he basically says it's going to happen within the next 1 to two years and it's going to be really smarter than any

human. When things are changing rapidly the ability to predict the future I think is uh becomes a lot harder because the rate of change is so great. Um but I think some things are fairly obvious to predict which is that we'll have um AI or AGI that's at a level that it can really do almost any cognitive I think really not almost really any cognitive task. It's just a question of when one could debate is it you know smarter than any human at the end of next year or is it two years or three years but it's not more than five years that's for sure. I think it's probably end of next year before AI can do better than any individual human could do. Um and then uh but there's it's a much higher bar to say well is it smart than um you know human intelligence collectively. uh but if the rate of change continues uh that that's why I think probably 2029 or maybe 2030 is where um digital intelligence will probably exceed uh all human intelligence combined. Now of course this video wouldn't be fit without talking about Samman. Samman essentially has made several accurate predictions from the past and I think his predictions are certainly the most insightful and I don't say that just because he's the CEO of OpenAI but he made predictions before about chat GBT at a time where there wasn't really AI that could do what he said. He was

talking about things in I think it was 2016 or 2020 where he was basically saying that AI is soon going to be like a virtual assistant. is going to be able to write essays. And it was pretty crazy how accurate that was for the future we're living in now. And September 2024, he actually made predictions about the next 18 months, which includes the next year, 2026. So, he says it won't happen all at once, but soon we'll be able to work with an AI that helps us accomplish more than we ever could without AI. And eventually, we'll have an AI that's going to be a personal team full of virtual experts in different areas working together to create almost anything we can imagine. And our children will have virtual tutors who can provide personalized instruction in any subject in any language at whatever pace they need. And we can imagine similar ideas for better healthcare, the ability to create any kind of software someone can imagine, and much more. So these are his predictions for the next 18 months. And I think they're rather accurate, and many of them are already happening today. Now, of course, it's time for Demis. This is the British AI researcher, neuroscientist, entrepreneur, and award-winning gamers. He's best known as the co-founder and CEO of Google DeepMind, a leading AI research lab responsible for the breakthroughs in AlphaGo and Alpha Fold, which are absolutely incredible. Now, DeepMind has basically said that in 2026, what they're trying to do is eventually combine the Gemini AI and via model. The reason they are trying to do this is so that they can actually get models that really understand the

physical world. Gemini was designed to be multimodal from the start, supporting Google's vision of a universal digital assistant that can help in real world context. And this aligns with the broader industry trend towards omni models that handle multimedia types like text, images, audio and video. And habish suggested that VO learns about the world by analyzing large amounts of YouTube videos. And Google had and Google has acknowledged that it models may be trained on some YouTube content under updated terms of service. So overall, it looks like in 2026 and late 2025, we could be getting a multimodal omnimodel that truly understands the physical world. As for his predictions about AGI and the future, he actually does say that it's quite like three to five years away. We've been working on this for more than 20 plus years, um, we've sort of had a consistent view about AGI being a system that's capable of exhibiting all the cognitive capabilities humans can. Um, and I think we're getting, you know, closer and closer, but I think we're still probably a handful of years away. Okay. And so what is it going to take to get there? So the models today are pretty capable. Of course, we've all interacted with the language models and now they're becoming multimodal. I think there are still some missing attributes. Things like reasoning, um hierarchical planning, uh long-term memory. Um there's quite a few capabilities that uh the current systems uh I would say don't have. They're also not consistent across the board. You know, they're very

strong in some things, but they're still surprisingly weak and flawed in other areas. So you'd want a an AGI to have pretty consistent robust behavior across the board all the cognitive tasks. And I think one thing that's clearly missing and I always had as a benchmark for AGI was the ability for these systems to invent their own hypotheses or conjectures about science, not just prove existing ones. So of course that's extremely useful already to prove an existing maths conjecture or something like that or play a game of Go to a world champion level. But could a system invent go? Could it come up with a new reman hypothesis or could it have come up with relativity um back in the days that Einstein did it with the information that he had and I think today's systems are still pretty far away from having that kind of creative uh inventive capability. Okay, so a couple years away till we hit AGI. I think um you know I would say probably like 3 to 5 years away. So you can see there that Demis' predictions are a little bit more conservative and like I said it's not that maybe that company's behind. I don't think there's any incentive for him to say that AGI is going to be next year. But overall, it's really interesting to see the different perspectives. Now, he actually does talk about the year 2026 on stage with Dario Ammed. So, this is an interesting discourse between them both about what happens next year and a few years after when we're at the point where we have a model, an AI model that can do everything a human can do, you know, at the level of a Nobel laureette like the one sitting next to me across many fields. um uh can do anything a human can do remotely, can do tasks that take, you know, minutes, hours, days, months. My guess is that we'll get that in 2026 and or 2027. Dennis, do you agree with that? I think you thought this was a bit further away. We don't disagree about uh too much. Uh I think the timelines are uh a little bit longer. The way I would define AGI is a system that's that can exhibit all the cognitive capabilities humans can. And that's important because the human mind is the only example maybe that we know of in the universe that is a general intelligence. Of course, how to test that is the big question. The one I'm really looking for and why I think it's a little bit further out, maybe 50% chance of in 5 years time, so maybe by the end of the decade. I think we don't have systems yet that could have invented general relativity. uh when Einstein did with the information that he had available at the time or another

example I give is can you invent a game like go not just play a great move like move 37 or build alpha go that can beat the world champion could you actually invent a game that's as um beautiful aesthetically and so on as go is so I think it's going to take a little bit longer to get that kind of capability now Mark Zuckerberg the American entrepreneur and programmer known as the co-founder and chairman of Facebook which is now called Meta is one of the world's largest tech companies. Now, he's actually spoken recently about some things that he expects to happen in 2026. He actually wants AI to do half of Meta's coding by 2026 at Llamicon with Sati Nadella, who's actually sitting down and having a chat about the future of AI and coding.

But the big one that we're focused on is um building an AI and a machine learning engineer to advance the llama development itself, right? Because I mean our bet is sort of that in the next year probably you know I don't know maybe half the development is going to be done by AI as opposed to people and then that will just kind of increase from there. Now of course it's time to talk about Eric Schmidt the American technologist entrepreneur and philanthropist best known for serving as the CEO of Google from 2001 to 2011. This guy played a key role in transforming Google, driving innovation in search, cloud computing, and AI. And recently, he made several statements on what he believes will happen next year in AI. Take a listen. We believe as an

industry that in the next one year, the vast majority of programmers will be replaced by AI programmers. We also believe that within one year you will have graduate level mathematicians that are at the tippy top of graduate math programs. So that's one year. Okay. What happens in two years? Well, I've just told you about reasoning and I've told you about programming math. Programming plus math are the basis of sort of our whole digital world. So the evidence and the claims from the research groups in open AAI and anthropic and so forth is that they're now somewhere around 10 or 20% of the code that they're developing in their research programs is being generated by the computer. That's called recursive self-improvement is the technical term. So what happens when this thing starts to scale? Well, a lot. One way to say this is that within three to five years, we'll have what is called general intelligence, AGI, which can be defined as a system that is as smart as the smartest mathematician, physicist, you know, artist, writer, thinker, politician. I call this, by the way, the San Francisco consensus because everyone who believes this is in San Francisco. It may be the water. What happens when every single one of us has the equivalent of the smartest human on every problem in our pocket? But the reason I want to make the point here is that in the next year or two, this foundation is being locked in and it's not we're not going to stop. It gets much more interesting after that. Now, of course, we have to talk about Jen Sen Huang, the Taiwanese American businessman, electrical engineer, the co-founder and president and CEO of Nvidia, the world's largest semiconductor company. This guy actually spoke about something that is 100% coming in 2026, hopefully if there are no delays. So, this is where the Nvidia CEO, Jensen Huang, talks about Nvidia

Rubin. It's a next generation AI platform consisting of a new GPU called Reuben and a custom CPU called Vera. Designed for advanced AI applications, it's named after Astravis Veron Rubin and is slated for release in the second half of 2026. Now, Reuben is expected to deliver significant performance improvements over Nvidia's previous Blackwell chips, particularly for AI training and inference. Now, he actually spoke about this here. It is a little bit lengthy. I mean, it's 3 to four minutes, but I will include it in the video so you guys can watch it if you'd like, which is the reason why I show you our road map a couple two three years in advance so that you we don't surprise you in May. you know, hi. You know, in another month, we're going to go to this incredible new system. I'll show you an example in a second. And so, we plan this out in multiple years. The next click, one year out, is named after an astronomer and her uh her grandkids are here. Her name is Vera Rubin. She discovered dark matter. Okay. It's the Yep. Vera Rubin is incredible because the CPU is new. It's twice the performance of Grace and more memory, more bandwidth, and yet just a little tiny 50 watt CPU. It's really quite incredible. Okay. And Ruben brand new GPU CX9 brand new networking smart nick MVLink 6 brand new MVLink brand new memories HBM4. Basically everything is brand new except for the chassis and this way we could take a whole lot of risk in one direction and not risk a whole bunch of other things uh related to the infrastructure. And so Vera Rubin MVLink 144 is the second half of next year. Now one of the things that I made a mistake on and so I just need you to make this pivot. We're going to do this one time. Blackwell is really two GPUs in one Blackwell chip. We call that one chip a GPU and that was wrong. And the reason for that is it screws up all the MVLink nomenclature and things like that. So going forward without going back to Blackwell to fix it. going forward. When I say MVLink 144, it just means that it's connected to 144 GPUs and each one of those GPUs is a GPU die and it could be assembled in some package. How it's assembled could change from time to time. Okay? And so each GPU die is a GPU. Each MVLink's connected to the to uh to the GPU. And so very Ruben MVLink 144. And then this now sets the stage for the second half of the year. The following year we call Reuben Ultra. Okay. So very Ruben Ultra. I know this one. That's where you should you go. All right. So this is Vera Rubin Ruben Ultra second half of 27. It's MVLink 576 extreme scale up. Each rack is 600 kilowatts, two and a half million parts. Okay. And obviously a whole lot of GPUs and uh everything is X factor more. So 14 times more uh more flops 15 exoflops instead of one exoflop as you me as I mentioned earlier is now 15 exoflops scaled up exoflops okay and

it's 300 what 4. 6 6 pabytes. So 4,600 terabytes per second scale up bandwidth. I don't mean aggregate, I mean scale up bandwidth. And of course lots a brand new MVLink switch and CX9. Okay. And so notice um 16 sites, four GPUs and one package, extremely large MVLink. Now just to put that in perspective, this is what it looks like. Okay. Now, this is this is going to be fun. So, this you are just literally ramping up Grace Blackwell at the moment. And I don't mean to make it look like a laptop, but here we go. Okay. So, this is what Grace Blackwell looks like. And this is what Ruben looks like. ISO is dimension. And so, this is another way of saying before you scale out, you have to scale up. Does that make sense? Before you scale up, scale out, you scale up. And then after that you scale out with amazing technology that I'll show you in just a second. All right. So first you scale up and then now that gives you a sense of the pace at which we're moving. This is the amount of scale up flops. This is scale up flops. Hopper is 1x. Blackwell 68x. Reuben is 900x. Scale up flops. And then if I turn it into essentially your TCO which is power on top power per and the underneath is the area underneath the curve that I was talking to you about the square underneath the curve which is basically flops times bandwidth. Okay. So the way you think about a very easy gut feel gut check on whether your AI factories are making progress is watts divided by those numbers and you can see that Ruben it's going to drive the cost down tremendously. Okay so that's very quickly Nvidia's road map. So the next one here is Aiden Gomez, a British Canadian computer scientist and entrepreneur known for his pioneering work in AI, especially natural language processing. Now, he actually co-authored the influential 2017 research paper, Attention Is All You Need, which basically introduced the transformer architecture and the foundation for generative AI models like Chat GPT. Now, in his prediction, he thinks that in 2026, we're actually going to get models that are no longer stateless, but actively learning. One of

the things and frustrations about these current models is that their knowledge cutoff is, you know, sometimes in 2023 or even early 2024. But he thinks this changes next year. There's this second thing which I think it'll come this year or next, which is continual learning or learning from experience. The status quo is that we build models, we spend hundreds of millions of dollars building them, and then they're frozen. And yeah, they're out there talking to people, doing things on behalf of people, but they're not learning, they're just static, and everyone talks to the same version. There will be breakthroughs coming soon which unlock the ability to learn from experience and you can imagine working with this model and asking it to do something and it fails. It's fine. You know interns fail like failure is not a bad thing. But the thing about humans is that when you teach them how to do it properly, when you correct them, they remember that forever. They'll never make that same mistake again. And so they grow with you over time. That's what experience is. So giving that to the underlying technology will unlock dramatic lifts in terms of value for the enterprise. So we're excited to push that forward. Then of course we have Imad Mustak, the British Bangladeshi mathematician, business executive and former hedge fund manager, best known as the founder and former CEO of Stability AI, the company popular open source AI image generator, Stable of Fusion. And he actually makes some interesting predictions about next year too. You could definitely run Deep Seek R1 on solar power panels. And if we look at the direction this is going because it's still not optimized, next year you should be able to get an 01 level model on your smartphone that pulls at most 20 watts of electricity and it's a less than a dollar per watt of solar power.

And this doesn't make sense if you look at what these models are capable of and we think about the cost of intellectual labor. Well, it makes sense when you think about how much energy your brain pulls. It's 20 watts and so we've got a we have a huge efficiency curve to ride to get there. And I think the thing is like by next year you will have these 01 level models on 20 watts which is our human brain level. And these are PhD level in so many areas and that doesn't compute because we've had these discussions of Microsoft is bringing back 3M island as a nuclear power reactor. You know Dyson spheres energy is going to use everything like 60 gawatt of electricity is coming on for data centers in the US I think over the next year or so. Yet when we get down to the actual numbers for a given unit of intelligence, it's a few watts. It's a few pennies. Before that, it would take entire teams using how many watts of energy in their brain, in their infrastructure, and we're not ready. Next, let's take a look at what Yan Lakhan says. The reason I've included this is because Yanukan is a AI scientist that has some incredible achievements, but he actually takes a different opinion than almost everyone in the AI industry. He actually says that advanced AI capabilities are not coming next year. you're not getting AGI for at least 3 to 5 years and it's

really interesting because he's quite controversial but is standing his ground and it's good to have different opinions. So how long is it going to take? So I think to have uh possibly a system that at least to most people feels like it has seal intelligence as humans if all of the plans that all of the things that we are imagining will work. Okay. So those JPA architectures and you know some other ideas that we're playing with succeed. I don't see this happening in less than five or 6 years. Okay. But now is it going to happen in five or six years? And I think the this a distribution with a tale that's very long and the history of AI is that people just keep underestimating how hard it is. I'm probably making the same mistakes right the same mistake right now. you know when I say five six years this is if you know we don't run into a major obstacle that we didn't foresee if all the things that we're planning to try out actually work if things kind of scale if uh computers you know accelerate and all that stuff like you know there's a lot of things a lot of planets that need to line up for this to happen so that's the best case right it's not going to happen next year like you might have heard from from you know some other folks yeah right yeah hopefully Alman you know various people uh or you know Dario Emily of yeah it's going to happen within the next two years or something. Uh no. So that being said let me know what you guys think if this stuff is going to happen next year or not. I think it's going to be interesting for sure. There's definitely going to be a lot of AI development. That being said, hope you guys enjoyed the