2026.10.08 | GP | Gaming | Ubisoft: Open Worlds vs AI Production

Primary:

UBI:FP

Associated:

6758:JP

7974:JP

700:HK

MSFT

Electronic Arts Inc.

TTWO

MTGB:SS

EMBRACB:SS

Viewpoint:

Former Executive

Moderator: Sahaj Singh Saluja (SSS)

FrançOis Courset (FC), Former Global Market Intelligence Analyst - Ubisoft Entertainment

SUMMARY

Overview

The discussion examined whether generative world models could weaken Ubisoft’s advantage in large-scale games. The expert argued that AI will initially accelerate individual production tasks rather than create polished, commercially viable games from a single prompt.

Ubisoft’s established franchises, proprietary data, creative direction, and brand recognition remain meaningful advantages. However, lower development barriers and faster AI adoption by Chinese publishers could intensify competition.

Key Insights

AI-generated environments are already becoming usable within conventional production pipelines. Fully generated AAA games remain much further away because they require coherent design, narrative, systems, and creative oversight. Ubisoft’s main defense is not world size but its ability to connect gameplay, story, lore, and curated content into a cohesive experience. AI could materially reduce unsustainable AAA development costs across coding, art, localization, testing, live operations, and environment creation. Publishers are expected to retain most of these savings. Large back catalogs should become more valuable as AI lowers remastering and platform-porting costs. Brand owners are better positioned than studios imitating older gameplay without the associated intellectual property. Human-made positioning may persist among smaller games, but the expert expects mainstream publishers to adopt AI broadly as consumer resistance declines. Younger players continue to value exploration, but may favor mobile, user-generated platforms, and culturally native experiences over traditional consoles. The moderator pursued the expert’s particularly strong view that Chinese publishers have established a substantial AI-adoption lead over Western studios.

Competitive Landscape

10/12/2026 at 06:05 AM UTC

Company or segment

Positioning, advantages, and risks

Ubisoft

Benefits from major franchises, proprietary game data, creative expertise, and a deep catalog. Risks include high costs, fragmented AI deployment, and stronger AI-enabled rivals.

Chinese publishers

Lead in AI integration and rapid live-content production. Their centralized support teams and lower cultural resistance to AI strengthen adoption.

Tencent

Offers China distribution, co-development resources, and technology. Its infrastructure could also support further consolidation of gaming assets.

Roblox and Fortnite

Threaten traditional platforms by combining creation tools, communities, and younger audiences, although their users may later adopt premium games.

Remastering service studios

Face pricing and demand pressure as publishers internalize AI-assisted porting and remake work.

AI world-model developers

Provide environment generation, simulation, and agent-planning tools that can accelerate game production.

Quantitative Insights

Metric

Fully engaging AI-generated AAA games

AI-expanded existing worlds

Coding efficiency

2D art efficiency

Localization efficiency

Audio and 3D asset efficiency

Live-operations efficiency

Quality-assurance efficiency

Environment-generation efficiency

Expert assessment

Approximately 5–8 years

Within about 3 years

Up to 30%–40%

Up to 20%–30%

Up to 50%

About 10%–15%

About 15%–20%

About 10%

About 5%–10%

Chinese PC share of Steam game sales

Roughly 15%–40%

AI-enabled production pipeline

China about 80%; West about 20%–30%

Industry professionals viewing generative AI negatively

52%

10/12/2026 at 06:05 AM UTC

SSS: So hello everyone, I am Sahaj Singh Saluja, moderator at Guidepoint Insights, covering EMEA's software IT service and media sector. Today, we are spending an hour on Ubisoft and a question that has hung over the gaming sector since Google previewed Project Genie in January. So if AI can increasingly generate playable worlds, would that stay defensible for a publisher built on large world franchises?

Joining us today is François Courset. François is a lead consultant at Naavik and before that he was a market intelligence analyst at Ubisoft. François, anything that you'd like to add about your background there?

FC: Yes, so currently a consultant at Naavik where I kind of founded and I'm leading the AI consulting practice specifically around gaming. So monitoring actively all the developments around AI for video games specifically. And before Naavik, I was part of the strategy team at NetEase in charge of defining the AI strategy for the company, evaluating new tools that were released and deploying those tools across production pipelines. So making all the assessments of whether these new technologies would have impacts on the way games are created.

SSS: Thanks, François. A reminder for everyone, today's conversation is limited to publicly available information, industry observation and advisors' professional experience. Nothing confidential, no material non-public information or security price or investment recommendation will be done here. With that, let's get started. So I wanted to start this conversation by talking about Google's Project Genie, which previewed in January 2026, and it can generate an explorable 3D world from just a text prompt. But that is only for around a minute and at 720 pixels and 24 frames per second. So a game like Assassin's Creed, which is one of the leading franchises of Ubisoft, it combines a large world with systems such as combat, stealth, and progression, a long story, and it had years of polish. So which of these layers do you expect AI to reach first? And which could stay hard to replicate even for the longest time?

FC: Yes, so I think with Google Project Genie, what happened is basically we saw the first example of world models, what we call AI world models, capabilities for games. And so what world models are, are AI models that are trained specifically on spatial data and are able to derive environments and actions within 3D worlds, which differ largely from the typical large language models we've seen before which were primarily able to generate texts or actions within 2D spaces. And so the potential for these world models is huge and goes way beyond gaming. But for gaming specifically, there are a couple of applications that are interesting. The way to think about those world models is I think we need to go beyond what Google Project Genie has showcased so far. Google Project Genie was just a technical demo, but since then, a lot of development has happened around world models with companies like General Intelligence and World Labs in particular experimenting with it. And there are fundamentally 3 types of world models: what we call renderers, models that can take actions in and output visually realistic pixel observations for humans. That focuses on 3D structure or accurate physics.

This is basically what Google Project Genie was. Then there are simulators, models that output underlying world states like geometry, physics, dynamics, so both humans and programs can inspect it, compute on it, and interact with these environments. And then there are planners, models that take observation in and output sequences of actions in 3D worlds. That's what's being used specifically for robotics, but has potential in gaming. So all of these application areas address different parts of the gaming ecosystem that could all be relevant to this open world conversation, especially simulators. With simulators, you can produce with AI entire environments that you can then use for your production pipelines, and that's what World Labs specifically has been making, and they already have tools that are mature enough to reach market. And on the other hand, you also have planners, so those models that allow agents to take actions within 3D worlds. So on one side, you can basically generate the environments, the open world in which your game is going to take actions, that's the simulator. And then the planners are going to give your NPCs the ability to interact and evolve within the actual world. And so all of that basically are building blocks that will be key to ultimately replicate open

10/12/2026 at 06:05 AM UTC

world. And so I think to wrap up all of this here, Google Project Genie tried to showcase this. I type one prompt and generate an entire open-world game. That's not the approach that anyone is going to be taking ultimately. But with this technology, what we can do is generate and accelerate the production of different pieces that, combined together, can create an open world game and as such really accelerate the production pipeline for such games. That'd be creating the environment, the rendering also, of course, 3D objects and the actions that NPCs will take within those environments. So it's really about how those new world model capabilities are going to supercharge existing production teams rather than being able to really generate an entire game from scratch with a single prompt.

SSS: Okay, now what needs to happen if an AI-generated world is good enough to just sit inside a commercial game rather than just being a short demo the way it is being produced right now and do you see is that closer to us in like next 2 years or do you expect it to take almost 10 years for these AI-generated worlds to sit inside these kind of commercial games.

FC: Yeah, so I mean again it all depends on how you define these generated worlds. If it's generating parts of the worlds that are then kind of used in production and deployed in final games, then we're already there. We can already use world models to generate 3D environments and then deploy that in actual public-facing games. So for this part, well, we're going to start seeing this in 1 year. That's actively happening right now. If you're talking about games that are like open world 3D games that are entirely generated from a single prompt, this is largely unrealistic within a near future perspective, within 3-5 years. Mostly because these kinds of games will lack the editorialization that is required for making a game interesting and relevant for players. What we're most likely to see is these kind of tools being used to expand on existing worlds that have been created. You take an Assassin's Creed game and the AI will expand the available world you can explore and engage with and create new quests and create new content that will keep the game engaging for longer than before. I think this part we're likely to see within the next 3 years. Complex 3D games with AAA quality that are interesting and engaging for a large consumer audience, I would say is more of a 5-8-year kind of time.

SSS: So Ubisoft's open worlds are known for their size, but players often complain that the side activities sometimes feel repetitive. So AI could make those kinds of repetitive content cheap for anyone to build. So if that happens, does a big open world stop mattering to people? Or what does the player care about more? Or care about the shift to the story and characters instead?

FC: So, all in all, the key thing in games is it fun? That's always going to be the central question for everyone making games. And so in open worlds, the fun can come from having a large amount of activities, side quests, explorable environments to engage with and that's part of the fun of these open world games. So this will remain an important driver for some of the players. If you go back to the player motivation, exploration is one of the core motivations for players, but there are others as well.

To your point, yes, there will be lower barriers to entry in terms of producing open world environments for games. So it's going to be easier for a game with smaller budgets to create games that will be able to rival the depth or scope of a Ubisoft game. So I think there's going to be more competition. But inherently, I don't think there's going to be any change in the player motivations themselves. Players will still be interested in engaging with rich open world environments that offer multiple activities to engage with.

And we can see that with Grand Theft Auto 6, which will likely be one of the most popular games of all time, and that's one of those rich open world games. I think the difference then comes down to how differentiated is your open world. Because with AI tools, again, you're going to be able to create something that is likely average in the sense that, yes, you're going to be able to replicate all those gameplay loops, meta systems, repetitive tasks that all happen within this open world. But Then how do you tie this into something that feels cohesive, that is tied to your lore and IP, that is

10/12/2026 at 06:05 AM UTC

making sense with the story you want to tell, and that is curated enough that it doesn't feel like an empty world that has been autofilled by all sorts of quests that are just there to keep you busy.

Again, it comes down to this editorialization and that's where the differentiators are going to be most seen. And so all the creative engagements and teams that are working at Ubisoft there will still act as a defensible edge versus games that don't have those capabilities internally. So tougher competition, more games will be able to achieve a similar level, but companies, games and studios that really invest into those editorialization and differentiator points will be the ones that will keep standing out and Ubisoft can potentially stand out in that context.

SSS: Okay, and talking about from the flip side. So where could AI world generation help a publisher like Ubisoft rather than threaten it? For example, in prototyping ideas faster, building environments more cheaply, or creating varied content inside its existing franchises?

FC: Well, I think the answer is kind of in the question already, but yes, there's a huge upside to AI tools for game publishers in terms of optimizing costs. We've seen that game development costs have been multiplied by more than 10 times over the past years and are becoming unsustainable and highly risky even for AAA game publishers. And with all these tools, there's a potential to significantly reduce the costs of producing these games to a scale that is more manageable and that allows these publishers to create more games or rationalize revenues or rationalize spending overall. And so that can include creating environments and game worlds more easily. That could be by creating more quests, dialogues, and narratives for the games. Ubisoft, for example, has tools that automate bark creation, and barks are all the lines that NPCs can say to players that are outside of the core story and narrative. So that's one use case that significantly reduces production costs. And so, for a large publisher, the upside will be in rationalizing its production capabilities. And where these publishers might have the upper hand is these publishers sit on a vast amount of data, previous games that they've made, which allows them to potentially fine-tune existing models and have them be more performing than off-the-shelf models that they would otherwise use. And we can see that already all the AAA publishers are investing into proprietary technologies that can create an edge, a durable edge versus competitors, which is similar to owning your own game engines that rationalize costs and potentially allows for creating experiences that feel unique and differentiated from what's available out there. So overall, the ability to secure a unique edge versus off-the-shelf solution and more critically, most critically, the ability to rationalize costs radically.

SSS: Now moving ahead to remakes. So remake of Assassin's Creed Black Flag sold around 3.5 million copies in just 2 weeks. So what makes a remake sell to new players and not just the old fans? And which part of that could AI make cheaper? So in terms of marketing and planning out the new remake of those games based on data and player actions, can AI help them make it cheaper to make remakes in that case?

FC: Yes, actually this week we've seen a lot of viral videos on Twitter of players that are doing game mashups or porting games to new platforms. I think it's the Claude 3.5 model release that helped unlock this capability. But basically, AI models are now able to decompile games which allows you to reverse engineer games and potentially port them across new platforms. So this capability is going to become extremely cheap right now. Cheap to the point that any player will likely be able to do it on their own. So all of that makes porting games to new platforms radically easier. Now to your earlier question, that is what makes players engage with a remake, a port, or a remaster that they might have not played, I think it comes down to the brand and all the brand activities and investments that a publisher puts behind a franchise. Everyone knows about Assassin's Creed. Every gamer knows about Assassin's Creed. So if you're able to turn an old game into something that looks similar enough to an actual current Assassin's Creed game, people will be interested in it. And they also know the legacy behind the Assassin's Creed brand and know about why this particular game has been central to the IP at large. So the interest in the IP and in the brand is what is driving interest for these ports and these

10/12/2026 at 06:05 AM UTC

remakes and remasters. And so the key there is making sure that the newer product is appealing enough to draw that audience that doesn't have the nostalgia to engage with an experience that feels dated. The other reason is access.

In lots of cases, these older games are simply no longer available or were available on specific platforms that are not available on platforms anymore, and so simply making those games available again is a key driver for people to re- engage or to engage with these games that they haven't played before but know about.

SSS: Just to follow up on that, if let's say AI cuts the cost for remaking these older games, which publishers are likely to gain more out of it? Those with large catalogs of well-known titles or smaller studios which could now make remakes or imitate the classic format more cheaply?

FC: Well, as I was mentioning earlier, the value is in the brand and the IP in this case. Publishers that have a deep catalog are bound to benefit most from it because the cost-benefit equation here is going to become more and more appealing. Cost of porting will be lower and the upsides of porting games to more platforms or remaking those will be bigger and bigger. For smaller studios, it means that these studios that would previously co-develop these remasters or remakes for larger publishers like Ubisoft, those capabilities will either be internalized by Ubisoft or will still be outsourced but for much cheaper because tools allow to port those games at a much cheaper price. Those co-development studios like, historically, Virtuos and Keywords are going to be at risk on that one and for smaller studios, the value in these older games is not in the gameplay systems, it's really in the brand. If you're making a copy of an old game, but you're changing the brand because you don't have the rights to this brand, you end up with something that is just not modern, not appealing, and doesn't have any sort of draw from nostalgia or brand recognition. I don't think studios without the rights to these IPs will really be able to produce anything meaningful that would compete with brands /companies that actually own the brand and everything kind of attached to it.

SSS: And we have seen a lot of young audience or young players who grew up on Fortnite, Roblox or mobile games. So they are kind of now shifting towards these games and focusing a lot on spending less time towards open world games. Do you think these are also some kind of a threat to... these established games in future because Fortnite and Roblox in this audience only have the capability to make their own islands and make their own games which players can engage with?

FC: So first of all, I would say that the assessment that these players who are on Roblox or Fortnite are not as interested in open-world games is, I wouldn't say wrong in itself, but it would need to be proven because first of all, there are lots of open world games on Roblox. There's Blox Fruits, which is one of the most popular games on Roblox, is open world. Then this same generation is likely to be engaging with GTA 6 when it's going to launch because this generation we know it is very engaged with a GTA roleplay at the moment. And this generation, by the way, is also playing Minecraft, which is sort of open world.

So the question is not whether open world is still drawing player interest. As I was mentioning, exploration is one of the core player motivations and that's very unlikely to change in any way because it's really part of player psychology and it doesn't really change from one generation to another, but there's always going to be interest for exploring game worlds.

The question is more: which game worlds is this going to be and on which platforms? And I think behind your question is the question of whether those gamers that engage with Roblox right now ultimately engage with experiences that are much more premium and AAA-looking? And that's a real question. And we see Roblox already making moves towards ensuring that its audience is evolving in interest from the basic graphics that you can see on Roblox toward more and more higher-end experiences.

10/12/2026 at 06:05 AM UTC

And there are early successes there. We see games that are starting to perform on Roblox right now. One very strong argument to say that players are going to graduate to more and more polished experiences is what we're seeing right now with the interest in Souls-like games on Steam. Players that grew up with Roblox are engaging with genres and categories on Steam afterward with experiences that are more polished than what they would have had on Roblox. Souls-like and horror games are probably the two most popular ones that drew from all this kind of culture that comes from Roblox but that is adapted to needs of an audience that is looking for higher-end experiences. So if on one side Roblox is pushing towards higher-end experiences and on the other side we see that these audiences as they grow up graduate from Roblox experiences to higher-end experiences as well. I don't think there are strong reasons to believe that these players won't be engaging with AAA experiences such as the ones Ubisoft is producing. The risk that we are seeing, however, is first, platforms. There's likely lower interest in console right now for these newer generations that are most likely to play on mobile. And so instead of playing on console, they will likely engage with high-end titles that are available on mobile. We see that with Genshin Impact, for example, being extremely popular. Or a game like Where Winds Meet from NetEase. So that's kind of the first trend, existing platforms and ecosystems are likely going to be replaced by others. Instead of console, Roblox might be one. The second trend is the current development budgets that are unsustainable. And so, it's unlikely that... Continuing to produce the experiences that Ubisoft currently produces with the costs that is implied with it is going to remain relevant for the long run compared to games that will be supercharged with AI available on these competing platforms like Roblox. That will deliver an experience that feels more native to the culture of this younger generation while matching the ambition of some of these bigger budget titles.

So it's more a shift in mentality that is going to happen there rather than a shift in what players are looking for. And so it's more on publishers to change their approach rather than chasing an audience whose needs are radically evolving.

SSS: So Black Flag, these things sold especially well on PC in US and China. So what makes Chinese PC players buy Western franchisee games? And could AI translation and voiceover help Ubisoft sell more of its older games in China?

FC: So, the first thing I would say is that localization costs are relatively marginal compared to the upsides of being available in a specific market. In that sense, publishers are already investing in Chinese localization and voiceover to a large extent. Voiceover maybe not necessarily, but at least standard UI and dialogue localization is already done in Chinese. So it's not whether publishers should move to Chinese localization, they're already doing it. And they're already doing it because they understand the market potential there. On average, games on Steam can sell between 15%-40% of their sales in China on PC. And so this is a sizable opportunity. So the opportunity is there. the opportunity is understood by publishers already. And yes, AI will make this process easier. It already does. A lot of localization companies are using AI-enabled solutions. Allcorrect, for example, has a product called Alocai, which helps with this AI translation. So it's going to be less costly to localize in China. Publishers are already there and already localizing, so AI is not going to increase the amount of games that are going to be localized. It's just going to make it cheaper. And what's going to make the difference in terms of succeeding in China is whether you're able to access the official channels beyond Steam, whether you'll be able to have marketing support in the region to communicate around your games, if you have partnerships with local entities in the market, and that's what's really going to drive your success. So partnering with a publisher like Tencent, which Ubisoft is doing to maximize the visibility of your game in China is really what gives you the edge. It's not your localization capabilities because localization everyone is doing it. There is an argument to make around hyper localization and yes AI can help. But hyper localization is more the strategy rather than something that could be supported by technology. So to succeed in China, have a dedicated China strategy, have dedicated partners, have dedicated investments, sign deals with local Chinese entities. These are what's going to move the needle most for your games rather than shifting from traditional localization to AI-enabled localization, even if this is indeed going to reduce your costs.

10/12/2026 at 06:05 AM UTC

SSS: Okay, now many players are still skeptical of games that visibly use AI. So on Steam, all these games which are using AI have to mention that it has been generated using AI. And even Ubisoft apologized in late 2025 for AI generated artwork found in Anno 117. Apart from that, a recent survey by GDC also said that 52% of the industry professionals think that generative AI is kind of harming the industry. So if keeping this as a premise, could 'made by people' or 'made by real humans' become a quality signal that helps established publishers stand out?

FC: So the first thing there is the reality of players' perception around AI is actually more nuanced than the blanket pushback. Because if you're diving into players' perception, first you're going to see that from a trend perspective. More and more players are okay with engaging with games that leverage generative AI. Then you also see that all application areas are not seen as equal. Most players don't care if you've used AI for producing code. They will care only if you're using AI for artistic assets that are used in your game. So you can see already that there's some irrationality behind how players perceive and push back on AI usage. And we're seeing this kind of going more in the direction of players accepting AI progressively. And we've seen that in the past with other trends in gaming, like in-app purchases initially were very poorly perceived by gamers and now are used as a standard across most games.

And to keep the parallel and answer the second part of your question, yes, right now you have games that are premium games with no additional monetization on top. And that kind of have a very hard stance on that, saying that there won't be any form of microtransactions in those games. And these games coexist with the rest of the ecosystem. Now what we see is that the top performing games, they're all pretty much using in-app purchases. And there's no real pushback related to those in-app purchases. If you make the parallel to AI now, it means that yes, there's likely going to be a group of games that will maintain this AI-free positioning.

But they're likely going to be smaller games, indie games that cultivate this idea of human craft, human labor. But for the bigger titles, the mass market titles, they're likely to all use the same technologies, all incorporate AI and maximize the benefits that AI can bring here. So yes, there might be some attempts at making games that don't use AI, but it is likely that those games will remain a minority of the market versus games that embrace AI.

SSS: Got it. So basically, these reactions of pushback will fade away as AI outputs increase and players get more used to using and playing with that, right?

FC: Yes, and you can see that already in other industries. On social media right now, you can already see a lot of AI content, AI videos or AI posts that are natively integrated with the rest of the experience and players not necessarily complaining about it. So if it's shifting also across the adjacent categories, there's no real reason that games won't follow suit and follow the same path.

SSS: Okay, now big franchise games usually sell for about $70, an indicative figure. Now If AI makes games cheaper to build, will players still pay that much or could well-known franchises ask players to pay even more than the other games?

FC: The first thing to say with pricing here is that pricing right now, prices of video games have increased slower than inflation over the past 20, 30 years. So there's little reason for games to diminish their price overall, simply because production costs have increased way more than selling prices have increased, so that's the first thing. The second thing is there are many different reasons for prices to change, but those reasons aren't necessarily related to AI usage. So you see right now games that will push for lower price points because they have plans for different monetizations down the line, DLCs, subscriptions, battle pass, things like that. And so with multiplayer games in particular right now, you start to see a model with a base price that is more towards $40 or $50 that is then kind of complemented with other sources of monetization. Then for games that have the highest budgets, yes, we're likely still going to see a push for prices to increase the same way that we see price increasing across pretty much any other entertainment areas or hardware or everything.

10/12/2026 at 06:05 AM UTC

The thing here is that price elasticity is very constrained in gaming, meaning that it's very hard for players to accept a higher price and you start easily losing sales. And so the wiggle room there towards the top is very, very limited. Adding AI into the mix, which will make production cheaper, I would say that we're likely going to see price diminishing rather than increasing overall.

SSS: Okay, now Rainbow Six Siege has been running for approximately 10 years and was one of the successful open world games for Ubisoft. What keeps its players loyal? Is it ranked play and esports, playing with the same group of friends or regular new content? And which of these could be the hardest for an AI-made game, which is, again, an open world game, to copy?

FC: What supports Rainbow Six player engagement in the long run is the live strategy around it, meaning, on 1 side having a continuous communication plan that is tied to esports events and competitions that gives continued visibility to players and live service content. So new activities, new characters, new seasons, new weapons, new skins, new maps, new everything that is released very regularly on the game.

So it's that pace of adding content that keeps players coming back to the game. If you look at these 2 aspects, the first one, the communication, the visibility, the brand is something that can be supported by AI. But that will likely not be fully replaced by AI. You can't ask AI to create an esports ecosystem for you. What AI will support is the amount of content that you can produce for a live game. With AI tools, you can repurpose assets and make Christmas-themed versions of your assets very easily. You can create new characters quicker than before.

You can add new maps, new content, new IAPs, all of that live service content, the production will be made easier through the use of AI. Yes, it would be easier to maintain and retain an audience in the long run. Rainbow Six doesn't have much to do with open world AAA game though, so the learnings for Rainbow Six don't really apply to an open world title because open world titles tend to be single player and limited in time. The only learning that you can say is, yes, it's going to be easier to maintain engagement around the game, whatever type of game it is, by being able to produce content continuously over the long run. And it's also going to be easier to maintain a strong brand top of mind because... With the amount of content that is being produced by AI, all the AI slop trend, brand is very defensible and a very strong way to distinguish your content. And so brand is going to be a big asset for any type of game. All the work that Ubisoft has done so far around this brand is going to pay off in that AI-enabled era.

SSS: Now Ubisoft has made its teammate prototype which uses AI-driven squadmates that respond to voice command available on its internal team. So could AI companions help live games keep solo players or new players who leave just because they lack a full team? And what might stop players from accepting these AI teammates in competitive modes?

FC: Yes, so on the retention side there's definitely upsides to having these AI teammates in the sense that for a player that is trying to learn the game, it can be very frustrating to join a game, especially late. If you join Rainbow Six like 7 years after launch, while all the players that are still playing the game are very good at the game. So as a new player, it's very hard to learn the game. And so being able to play with bots that play properly and can help you kind of step up your level over time is definitely a huge plus. So that's one area that will be facilitated. The second part is... Are those AI companions fun or not? What's pretty clear with these AI innovations is that it can make bots and non-playable characters more tuned to your level, which will ultimately make the game less frustrating.

So that's a win. But engaging in natural language with NPCs, this is something that has been experimented across multiple games and hasn't really found any sort of commercial success because players don't see it as fun, because engaging in natural language with an NPC is slow, it's hard to get the information that you need, and it derails you

10/12/2026 at 06:05 AM UTC

from the core kind of gameplay activity. So that part doesn't really add value to the game at the moment.

So generally speaking, I think the upsides of these technologies are going to be much more around, yes, how do you manage to onboard the player? How do you make sure that his first weeks playing the game will be less frustrating than they were before? I do not think that from a social perspective, playing with bots will fully replace playing with your friends. And I don't think there will be a lot of additional value added besides that, just simply because engaging with NPCs in natural language isn't inherently fun, nor aligned with the core game premise.

SSS: Now moving ahead to the development cost part of a game. So across studios that you have worked or researched with, in which part of a big budget game development are AI tools already reducing the work needed? Art, code, testing, localization, or writing, and roughly how large are these gains today? Is it close to 5% of the project effort or closer to 20% in that sense?

FC: So, I mean, right now, the areas that are most mature in terms of application cases for AI are coding, where you can achieve up to 30%-40% efficiency. 2D asset production, 2D art, where similarly you can go up to 20%-30% efficiency. Then localization similarly up to 50% efficiency. And then you have emerging areas that are a bit less mature, like audio, where it would be more at 10%-15%. Live ops activities, I would say 15%-20%. QA would be more around 10%. What else am I forgetting? 3D assets. 3D assets you are within 10%-15% at the moment as well. And then emerging application areas like environment generation or things like that would be more at 5%-10% right now.

SSS: Now, if AI makes these big games cheaper to build for everyone, who is most likely to keep these savings over the next few years? Is it the publishers through their higher margins, players through the cheaper games or more content? Or is it the platforms such as Sony, Microsoft, Steam, or Epic through their share of sales?

FC: So players first, as I mentioned, I don't think there will be any meaningful change in terms of price. I don't think there will be any player-facing benefits here. On the platform side, platforms are kind of tied to the scale of the market. So the question is, are we going to see more players playing games than before? Because of AI, I'm not sure. There's little reason to believe that. The demand is likely going to stay similar to where it's at right now, maybe slightly growing, but nothing major. The biggest impact is going to be on reducing production costs for publishers, where significant gains will be made in terms of how expensive it is to make a game from 0 to 1. So publishers are really the ones that are going to benefit most from these emerging AI tools.

SSS: So Ubisoft now runs 5 separate creative houses each with their own games and budget while AI tools are built by a central team. So in your experience do independent studio teams pick up these central AI tools quickly or do they resist them because they were made keeping in mind the other franchises as well or other games as well and it might not 100% suit their game as well.

FC: So I don't think it's really a question of fit with games, but more a question of the way these companies ensure access to quality talents is by building AI labs that allow those people to work on very forward-looking applications. And so the risk that you run with this is having AI labs or central AI teams that work on developing concepts that don't really align with the team's primary needs, with the production team's primary needs or the studio's primary needs. And so it's more a question of communication and alignment between what these AI labs are doing and what productions are actually needing, that is complicated. And we see that time and time again where these AI initiatives are kind of siloed and built in isolation from the teams that actually need them, and so yes, there is a disconnect there. I don't think it's specific to Ubisoft, but Ubisoft has that problem because their AI team, Ubisoft La Forge, is kind of fairly research-driven and isolated from the rest of the organizations. And then yes, to your point, AI tools aren't one size fits all, and the closer it is to the actual needs and customized to

10/12/2026 at 06:05 AM UTC

the needs of a specific team, the better and more impactful. And so it's essential to have coordination from the studios to these AI labs to ensure that the tools being developed are as impactful as possible.

SSS: Mm-hmm. Now as we were discussing that games like Ubisoft have a lot of decades of game assets, animation, and gameplay recorded in their own engines. So could Ubisoft earn more money by selling its game data to these AI companies and what could stop it? So actors rights over their own voices and faces the risk of helping a future rival or player privacy.

FC: So, yes, the first limitation is by selling its data to third parties, Ubisoft puts its own edge at risk because that data is what's going to differentiate your internal AI tools from competitors. That said, right now there's a huge market for acquiring quality training data, especially for world models. There are companies like Origin Lab that allow publishers to sell their game data for world model training. So there's a clear financial opportunity there. There is definitely a player data concern, especially with a company based in Europe like Ubisoft with RGPD kind of constraints. But the constraint is more on publishers having built the capabilities before to gather, collect and work around that data. And so the potential associated with it really depends on the investment made around the data infrastructure that each and every company has made.

And so, yes, to your point, there are some limitations around privacy, strategic advantage that is also becoming kind of a potential issue if you start selling the data to competitors. But overall, I think what I'm seeing at least is more and more publishers considering selling their data because they are very runway conscious right now and are looking to find new ways to generate revenues and selling data is one of them.

SSS: So talking about Tencent, which we discussed holds about 25% of Ubisoft. So it is also building its own AI world model. So at a market structure level, how do Western publishers usually benefit from partnerships and large Chinese platforms in technology and distribution and where do their interests tend to conflict over time?

FC: Yeah, so the key benefits of an investment from a Chinese company are going to be centered around 2 things. The first one is market access, having a partner to launch the games in China, which we already discussed previously as a major upside for Western publishers. And the other upside is tied to having access to those partners' co- development capabilities. So there are art team, there is a QA team, there is an UA team, and companies like Tencent have very large kind of centralized team that can support second-party studios like Ubisoft. So this is what you typically get out of an investment from Tencent. What we are likely going to see is these companies that have invested in proprietary AI development tools acquiring IPs that they will then be able to iterate on. So it's not a question of competition or interest diverging, but it's more that at some point we're likely going to see companies like Tencent rolling up more and more gaming organizations so that they can maximize the benefits of all the investments they've made towards their internal development infrastructure.

SSS: Now just moving towards the last question of this consultation. So if I want to compare Chinese publishers with Western publishers or big budget studios right now. So how are Chinese publishers adopting AI in their game production compared to Western ones and could that let them release large and high quality games faster or more cheaply than publishers like Ubisoft?

FC: Yes, so to give you an idea, I would say that right now about 80% of the production pipeline for a Chinese studio is AI-enabled compared to maybe 20%-30% for a Western company. So the pace of adoption is way faster in China and that's driven by the way those companies are structured and all these kind of internal co-development centers that are very costly to maintain and so these companies have very strong incentives to rationalize internal teams. And also because Chinese companies are developing games where the amount of content you produce is a very strong differentiator and they also don't have the same kind of perception of creativity and its role in making the case.

10/12/2026 at 06:05 AM UTC

So fewer barriers and more upsides for Chinese companies which all contribute to a way faster adoption. And yes, and we're already seeing that. Games released by Chinese companies are able to sustain way more live content than overseas titles, meaning that these companies are able to retain way better their audience by providing way more content than their Western counterparts and that creates a divide between Western and Eastern companies that is favoring companies based in China. Bye.

SSS: So François, this has been a really insightful conversation and we have covered a lot which part of open world AI may struggle to copy and what AI means for Ubisoft's backlog and catalog, and how this raises a lot of new questions towards how it is framing up the gaming world today. I'd like to thank you for your time and candid perspectives today. With that, we wrap up this call. Thank you so much for joining.

FC: Yeah, thank you so much. Don't hesitate if I can help in any other ways in the future.

SSS: Sure.

FC: Okay. Thank you so much.

DISCLAIMER: The statements or opinions expressed today are those of the Advisor and not Guidepoint, who disclaims all liability for the content provided. The Advisor may not disclose material nonpublic or confidential information or any information that would cause the Advisor to breach any duty or obligations. Guidepoint is not a registered investment advisor and the information provided is not intended to constitute investment advice.

10/12/2026 at 06:05 AM UTC