Top 3 Ways to Get Rich with AI Video in 2026 (Sora 2 vs Veo 3)

AI video has just opened up a new lane for making money online. You may have heard of Sora 2 or VO3. Using these latest video models, people are getting insane engagement and making real money online. Whether it's viral videos raing in millions of views across social medias, or AI influencers landing $10,000 per month, brand deals, all sorts of marketing content like product demos and UGC testimonials are being generated in minutes, and automated YouTube channels are pulling millions of views per month with scripts and B-roll that are completely generated by AI. This is a pivotal moment because these video generation models just got good enough that when used right, they can produce results that are nearly indistinguishable from reality. I run the world's largest AI business community and also guide companies through their AI transformation through my agency, Morningside AI, and I've seen firsthand how people can make bank with these kinds of AI videos. So, it's safe to say that we're at the right time to be at the leading edge of this new money-making path in the AI space. But the key is using the right tools for the job. So, in this video, we're breaking down three proven paths for making money with AI video and putting two of the latest models, Sora versus VO 3. 1, head-to-head to see which one is best suited to build these systems and ultimately make money with AI. Producing AI video can be simple using just a quick prompt for a quick video. Or you can produce longer and more in-depth videos that have a complete story arc, generating the script with an LLM, then storyboarding it with AI images, then editing generated clips together into a full story, including music, sound effects, and captions as well. But as it sounds, this can be pretty timeconuming, requiring a bunch of human work and a bunch of different tools as well. That's why AI video platforms like our partner in video that I've been using a lot lately are evolving to bundle these steps into autonomous video agents which handle hundreds of these decisions for you under the hood. They can behave like a director, an editor, a sound designer, and more, providing a unified workflow to produce polished videos with direct access to the latest versions of these AI video models as they come out. And on platforms like Nvidia, you can generate these videos without watermarks as well. So, the barrier to entry has just collapsed. You can now pump out tons of high-quality videos at scale and quickly test what is working and double down on your success. But again, you're going to need to know how to do AI video like a pro. So, let's break down the best business cases right now and the best tools to use so you can stop wasting credits and start earning some real money. For our first business use

case, we're going to learn how to create viral videos. I've carefully selected a few proven examples that you can easily replicate to generate millions and millions of views. People are literally using these formats to do this right now. So to start off, we're going to head over to Invido AI and you can either sign up to create an account or you can log in if you already have one. And once inside, you'll see this input where you can add in a prompt. This will trigger NVIDIA's video agent, which is going to make a bunch of creative decisions for us under the hood and is capable of generating complex longer videos for us. So we'll take a look at this later in the video, but for now, since we want to keep things short and sweet, let's head to the agent and models tab and put Sora 2 and VO 3. 1 headto-head to see who comes out on top for a lot of these use cases that we're talking about. So, these are the video models that we're going to be using in this tutorial, but there are more models available on the platform that you can also use if you feel like it. There's also this trend section, and I'm going to be going over that a little later in the video. So, to get started, we'll create a new project, and then give it whatever name, and then we'll be able to start prompting. We can also add attachments in here for the model to reference, select the aspect ratio, which will allow us to generate for vertical video, for social or horizontal for platforms like YouTube. And finally, the most important part is the model selector. In addition to video models, we can select image models to generate with too. We've got models like Nano Banana and Cream. For our first example, we're going to be creating a mesmerizing glass cutting ASMR video. The kind I'm sure like you guys have probably seen a lot of these on social media before, but as you can see, these types of videos often rack up like tens of millions of views and they seem relatively simple to make. So, we're going to be making an ASMR cutting video featuring a glass banana. So, I'm going to select VO 3. 1 fast and then I'm going to paste in my prompt. Use chat to help me generate these prompts. And as you can see, it includes a lot of crucial keywords that are going to make a big difference in the final result. So, you've got phrases like, "A razor sharp stainless steel knife poised just above the banana. " Or, "The first slice breaks through the glass surface cleanly, sending delicate translucent shards scattering around the object. " These kinds of vivid details are going to help to ensure a great generation for us. The default duration is 8 seconds, but you can produce shorter videos if you'd like. We could either generate this as a short for somewhere like Tik Tok or Instagram or horizontal if we wanted to publish a bunch of these onto YouTube as one big longer video. I'll select the horizontal aspect ratio. And finally, we're going to hit generate. Once it's generated, we'll find it over here in the interface. And let's see what we got. So, this is looking exactly like what we're looking at earlier that actually went viral. So, now we can head back into Nvidia and see what that same prompt is going to get us when we use the Sora 2 video model. Yeah, that's not too good. I'm assuming you'll agree with me here that the VO3. 1 video was the clear winner here while Sora 2 is clearly struggling with this, but maybe that was a fluke. So, let's try out another viral video prompt for a video like these crazy what kind of bed would you sleep in videos which are raking in millions of views at the moment. So, here's what we got with VO 3. 1. Not bad. And here is with Sor 2. I'm going to have to give this a VO3 again. But for good measure, let's just run one final test. And we're going to recreate this like which ocean would you rather swim in trend that people are using to get a lot of views. So using V3. 1, we can conveniently set the start frame and the end frame for our video. In other words, we can set the image that we'll see in the first and the last millisecond of the video. This helps us to much more clearly define how and where we want this video to go, which can be far more powerful than just trying to describe in this textonly prompt. So earlier I created these two images here using seedream here inside inv video. And I'll set these images as my start and end frames. Then I'm going to paste in my prompt. And by the way, if you're wondering how I actually came up with this prompt, I uploaded the same start and end images into chatbt and then asked it to construct a prompt that will guide the movement from that first to the last image. When it comes to prompting, there are of course best practices to getting the best outputs out of these models. An ideal prompt should have the S O R A ID or saw rate anatomy which includes firstly the subject. So who or what is going to be in the video? Secondly, the object. What action or movement is happening? Three, the realm. What's the setting or the environment that it's taking place within? Fourth is atmosphere. What's the mood or lighting or the feeling of the space? Fifth is imaging. This is the camera work like the pans and zooms. And finally, the details, which is like the style and the quality. If you make sure to mention all of these, you'll have a much better chance at getting exactly what you want from the AI video you're generating. For example, if you notice how I'm emphasizing the first person view, so you're going to want to think about the perspective your video is being viewed from. We're also directing exactly what happens at each second. So, just like those viral videos, after a few seconds, the point of view descends smoothly beneath the shell covered ocean surface, and then it's going to transition into an otherworldly underwater realm. Now that you have a better idea of how to create a good prompt, let's see what this one comes up with next. Okay, that one came out pretty well. Now, let's see what Sora 2 can come up with. So, notice that when I switch the model over here, the option for the last frame has disappeared. And that's because Sora doesn't have that first and last frame setup options, at least at the time of this recording. But that's okay. We can hit generate and see what it comes up with. Okay, that looks pretty good, but I'm going to have to give it to VO3. 1 again for this one. So, V3. 1 remains the undefeated champ. I think it's just a bit better suited for this kind of ASMR genre. And the control it gives with the first and end frame definitely helps, too. But, we don't want to solely determine which model is best based off some crazy ASMR videos. So, let's put them headto-head again on another viral video angle, which is the AI avatar. You've probably seen these characters on social media like J Monkey and this Yeti Bro. I'm constantly getting these kinds of videos in my feed. Let's see which model can generate better viral characters like this. So, we're going to start with generating an image of a character that we can actually use. And this is where your creativity has to actually come in. So, since I'm a New Zealander, I thought it'd be cool if I could come up with a Kiwi as a mascot for my agency, Morning Side AI. I'm just going to paste in this prompt that I wrote earlier. It's for like a surfer bro Kiwi. And since I wanted to wear a hat with my company's logo on it, I can attach my logo here. And I'll select the Cream image model and generate it as a vertical image first. I think that looks great. So, let's go and take this image and generate some videos and see how VO3 vers Sora is going to stack up. — Kod team, it's your boy Kiwi Kev coming at you from the best little island on Earth. Sun's out, Jandle's on. And I've already lost my surfboard twice, bro. If you ain't wiped out, you ain't trying, mate. All good, bro. Just testing gravity. E, no worries, mate. — Kr team, sun's out, jandle's on. If you ain't wiped out, you ain't trying to, mate. Kiwi Kev saves the day. All right. Well, now like and sub. — So, these are both pretty good. I think we're just going to have to call it a tie, at least for these AI character videos. So, viral video content like this is a good entry point because it's relatively easy to generate as you saw and it's something you can do yourself without relying on a team or on finding clients. You can literally just do what we just did there and come up with a character and start making videos and then be posting it yourself on social media and try to monetize from there. And if you're successful and create a popular enough character and a channel around one of these characters, you could start doing brand deals with them where companies are going to pay for your AI influencer like the guy we just cooked up there. And speaking of branded

content, that takes us to our next business case, which is creating branded content with these AI video models. So instead of creating content, you post directly on your own social channels. There's a growing opportunity for AI creatives to make branded videos for clients in terms of product demos, UGC style testimonials, and air creatives that look like they came from a full production crew. So, for example, here's a quick product demo for O Lollipop that was created entirely inside of Nvidia. So, how do we create videos like this? And which model is best suited for creating this kind of content? Only way to find out is to make some sales and put these models to the test. And since beauty is such a huge niche on social, let's say that I want to create an ad for a client in the space who is launching a new fragrance. Over in Invido, I'll paste in a prompt that I made earlier. And I'm going to add a reference image of a perfume bottle that I generated earlier with Nano Banana. And I'm going to see what comes up using Sora 2. Honestly, I think that came out pretty good. right down to the realism of the spray coming out of the bottle. Sora is known for having that realism grounded in actual physics which allows it to handle these kinds of moments really well. Now, let's see how V3. 1 compares. Again, we switch to the VO3. 1 model and we get the start and end frames that we can set. So, I'm going to add another image of the same perfume bottle as the end frame and see what it comes up with then. That was pretty good, but I think I'm going to have to give that to Saw two on this one. Even without the start and end frames, it ended up handling the task a lot better and really captured the realism of the product with that spray. Since that was pretty close, let's do another one, but this time for a different product. Let's say we're working with a client who's trying to market their energy drink. And they send us a static image of their product, which is just the drink can with this. Let's see what we can generate with Sora 2. That one was pretty good. It's got like the condensation on the can and like the fluid dynamics coming out. Again, Sor is pretty good with the kind of handling the physics of these product ads. But now we can switch over to V3. 1 and see how that does. We're going to add our N frame and then regenerate. Yeah. So, that was okay, but I've got to give to Sora 2 on this one as well. For this category, Sor 2 is pumping out high-quality content that really shows off the product and make it feel like it was filmed with a huge budget and a talented director photography or like a videographer as well. You remember earlier how I mentioned that Nvidia has these trends. So, these are basically pre-built viral templates which we can put our products inside of like when a holiday is coming up and you want to make festive branded posts for your business or your client. These are perfect for that. For example, we could place our energy drink in this jacko'lantern trend. Just note that these trends are using a different video model. I know this is a Sora versus a VO comparison, but I wanted to show you how you can make the most out of the NV video platform. There's tons and tons of features on here, so you can use this stuff for yourself or your current or future clients as well. — They locked the gates, but nothing stops this BT. — Now, let's talk about one of the most in demand services right now, which is UGC or usergenerated content, and how you can start offering them to paying clients. Brands are constantly looking for short form videos that feel authentic and like they came from a real customer or an influencer. But shooting those kinds of videos can be expensive or timeconuming as businesses have to establish relationships with influencers or their past customers. They have to make sure they have the product. They have to set up a paid partnership in the case of influencers and so on just to get these testimonial style videos out there. That's where AI video tools give you a massive advantage in supplying these types of videos to clients. So let's see how store 2 versus 3. 1 stack up for these kinds of videos. So, we're going to paste in a UGC prompt and attach a reference image of the product that we want to promote. Making sure to switch to a vertical aspect ratio, which is ideal for Tik Tok and Instagram. And then we're going to hit generate. — Okay, wait. This perfume actually gets compliments. It's warm, a little dreamy, and it lasts all day. The moment I spray it, people are like, "What is that? " It smells so good. Like, wo! — Perfect. I saw delivered a video that looks handheld, real, and relatable, which is exactly what brands want for social ads these days. Now, let's create a VO 3. 1 version. And I'll set a start and an end frame as we usually do to control the camera motion and use the still from the previous clip to keep it consistent with the same influencer. — Oh, okay. Wait, this perfume actually gets compliments. — So, in this case, both have turned out pretty well, but Sora gives you that raw and influencer style realism, while VO feels a little bit too cinematic and overly polished, at least in my opinion. But ultimately, I think in this case, Sora wins this category as well. So many businesses need these kinds of authentic short form ads for their socials and they need tons and tons of them to test so that the ad machine can do its work and find the winners. You can now deliver these to them as a oneperson AI creative agency. You've got no actors, you've got no shoots, no gear, just you and AI and tools like this. So you could even go and sell these as a monthly package like 20 UGC ads per month, $500 or $1,000 or offer them on a per video basis for like $50 or $100. And with a generation cost of only a few dollars per clip, your margins on this are going to be huge. Now for our third business

case, YouTube automation. Not everyone wants to be on camera like this, and I get that. But with AI, you don't actually have to be. You can be on YouTube and have a YouTube faceless channel. And you can do cinematic explainers. You can do motivational videos or storytelling compilations. These kind of videos are pulling in millions and millions of views and consistently making them thousands and thousands of dollars per month in AdSense. So for years, these kinds of videos relied on stock footage, which meant licensing a stock library and digging through all these generic clips in order to match your narration. But now that's starting to change because with AI generated B-roll, you can create custom visuals that match your exact scripts and scenes that look like they were filmed specifically for your video. It's a difference between piecing together a patchwork of these generic clips and directing your own mini film. And that's where Invidia's Copilot video agent comes in. Because instead of just generating a single short clip, the agent is going to build an entire sequence designed on the pacing, the transitions, and add music to bring the ideas to life automatically for you. You've already been watching it in this very video. So, not only can you use these AI generated B-roll for faceless channels, you can also do it on any kind of channel even when there is a person in the picture. So, instead of spending hours on handmade motion graphics, you can spend just minutes telling AI to generate the visuals for you, which is a massive time saver, of course. So, let's see which model Sor 2 or VO3. 1 is better suited for this creative workflow. So, let's say I'm creating a video about automation freeing people from repetitive work. And I want a quick B-roll sequence that's going to visualize this idea for my audience. I can describe exactly what I want the viewer to see. And here we'll specify the model in the prompt itself. And we'll keep this one in widescreen since it's for YouTube. And then we're going to hit generate. — A century ago, machines freed us from backbreaking labor. Steam engines roared, gears turned, and humanity stepped into the industrial age. But today, we face a different kind of labor. The endless digital grind. emails, spreadsheets, notifications. The mind trapped in repetition until now. AI awakens and everything changes. Papers become data. Chaos becomes order. Walls dissolve into possibility. Artists create, musicians compose, dreamers build. All enhanced by intelligence that amplifies human potential. Across the globe, barriers crumble. Students launch startups. Freelancers scale dreams. Small teams change the world. A century ago, machines freed our bodies. Today, AI frees our minds. It automates the boring so we can focus on the beautiful. No barriers, no limits, just imagination. This is freedom through automation. The new age has begun. — I'm pretty impressed by that. And if you switch over to this tab here, we can see that you can make edits to this video by just prompting the edits that we want. And we can even make changes to the music or the script individually. And if I switch here, we can see the individual clips that we got from Vio. So now let's ask for the same shot with Sora 2. — A century ago, the world trembled with the roar of steam and steel. Machines freed our bodies from backbreaking labor, launching humanity into an age of unprecedented possibility. But somewhere along the way, we traded physical change. — So, I'm not going to bore you with the whole 2-minute video there because video 3. 1 was a clear winner. And that's because the Sora video was more of like a slideshow of stills. I asked Nvidia about this and they let me know that it's because when Sora 2 detects a person in the frame, it blocks video generation entirely in order to prevent misuse of certain characters I guess which causes invido to fall back to using still images. So ultimately for this more long form use case for now the clear winner is VO 3. 1. Now imagine doing this for key moments in your video or in your entire faceless channel. You can generate all your supporting visuals, story intros, transitions, music without ever opening an editing timeline or hiring an experienced animator. So as you can see there's not a single winner of the best video model right now. Both Sora 2 and V3 excelled at some tasks and fell short on others. The trick is in knowing which models to use for which tasks, which was the whole point of like doing this experiment, and of course, how to put these into practice to make money with them, which leads us to the part that you've probably been waiting for. How do we actually take action on one of these business use cases and turn it into a way to make money?

So, but now you understand that there's these three different business cases that we can jump into, and you understand what tools to use to generate the kinds of videos that can go viral and that brands are willing to pay for and that you can use to scale your own YouTube channels. The first step is to

pick the lane that you want to go down. And this means not spreading yourself too thin. You need to pick the path that matches your strengths. If you love creating viral content or you're very kind of funny and creative, then trying to make viral content on social media by looking at what's trending and then putting your own spin on it might be the best option for you. When you get enough traction, then you can start experimenting and creating your own trends. And with enough engagement, you can start earning income off the platform itself. Or you could begin by building your own AI influence and become the next Jar monkey or the Bigfoot or the Yeti. But remember that the barrier to entry for this is super low. So the differentiator here is going to be your creativity, your humor, and your commitment to experimentation and trying to find something that works. See what gets engagement and then double down on it. Once you grow a big enough audience, eventually brands will want to do collabs with you and have their products show up with your influencer. If that doesn't sound like your thing, if you have business connections, then you can create several great examples of branded videos you can take to local businesses and your warm connections. In my free school community, I have a full guide on how to get your first AI client in 30 days using warm outreach, which is just talking to people that you already know or have some kind of relationship with. When you go to these businesses and offer some free video content in exchange for testimonials and ideally a case study, then use that experience and your portfolio of work to go and pitch higher value clients. Ultimately, this is a path to creating essentially an AI video agency or an AI creative agency where you run entire campaigns for brands and you with the trusted AI video expert. If those two options don't sound like you, maybe you could go down the route of starting a YouTube channel and you just need to pick a niche that's of interest to you and ideally isn't super saturated and launch your own faceless YouTube channel. You need to leverage AI for the scripting which an AI voice narrates or a character presents and then use custom animated B-roll for the storytelling or use a video agent like in videos which combines this all into one. This is a much slower burn since you'll need at least 1,000 subscribers before you can get your channel monetized. And since generating the content is going to cost you tokens, which aren't cheap with the current token rates on a lot of these platforms, you'll need to be smart about ensuring that you actually get a return on that investment. So regardless of what path you choose, you'll want to experiment, leverage your creativity, track your

progress, and once you see what's working, you'll need a plan for scaling. This is what separates the hobbyists from the professionals, which is systems and leverage. So, building templates of prompts you can reuse and tweak, creating automations that can help you analyze what's working, say, for your competitors or to analyze your videos and double down on your success. Over in my free school community, which is the first link in the description, I have full courses on building out fully automated systems, which can help you to scale your output much faster than doing everything manually. And after you've got consistent results and systems in place, then you can aim for that $10,000

per month, which is essentially like the freedom number for you to be able to quit your job and do whatever you want, travel the world, that's kind of the freedom number that people need to get to. for viral creators or those running your own YouTube channel or even a portfolio of faceless channels. That means building an audience big enough to earn revenue from the platforms themselves and eventually get brand deals and partnerships. For client work, that's just 2 to four clients per month at 3 to 5K each. If you scale through systemizing workflows, you'll be able to produce more content for more clients and eventually you'll need to hire a team under you. That's when you become more than a creator and you start building your own AI video agency, which is where the serious money and the freedom comes from. The biggest names in the AI video agency space are currently running sevenig agencies. So, if they can do that, then you can definitely make 10K a month with enough strategy, creativity, and persistence with these

tools. So, the whole point of this video is really that the AI gold rush has just started, and the models are getting better and better by the month. The tools, as we've seen with Invido, agents that can sort of do a lot of the hard work for you. They're getting more and more powerful with every release. So, if you want to get into it, you can sign up to Invido down below. They've got a bunch of free credits you get. And if you're serious about it, you can invest in a generative account, which is $100 a month. As I said in my recent video, I think AI video and AI creative generally is one of the best opportunities to get into AI in 2026. So, you've got everything you need. Now, feel free to join my free school down below, and you can access a whole classroom full of courses on AI and automation and how to start your own AI business. And also, you just got the biggest community with over a quarter million AI entrepreneurs who are sharing what's working, troubleshooting, and just celebrating the wins together. Plus, you can join my live Q& A with me every week, so you can ask me questions about how to achieve your goals. So, if you want to see my video that breaks down the top five ways for beginners to get into AI and build a business and make money, I'll cover that up here. and AR creatives and AR videos is one of those. So, there's another four that you're not aware of. Highly recommend you check that one out. But that is all for the video, guys.