
Paradox splits from Colossal Order and hands Cities: Skylines II to Iceflake. Can the sequel recover from its technical and business missteps?
Why the sequel stumbled
and how a new studio might save it
Announcement: “An Update on Cities: Skylines II” (Paradox Interactive).
In mid-November 2025 Paradox Interactive and its long-time partner Colossal Order announced a quiet but monumental shift. After more than fifteen years together, the companies would “pursue independent paths”. The Cities: Skylines franchise – Paradox’s flagship city-building series – would be handed to Iceflake Studios, an internal Finnish team. Colossal Order (CO) would finish one last “Bike Patch” and an asset-editing beta, then move on to other projects. The announcement formalised a split that players and critics had anticipated for months. Cities: Skylines II (CS2) had launched in October 2023 to technical issues, design missteps and a conspicuous lack of mod support. A year later, many of those problems persisted, and Paradox’s patience wore thin.
In this article I attempt to disentangle the facts of that breakup, to understand why CO floundered, why Iceflake has been given the keys, and whether the sequel’s underlying issues can realistically be fixed.
Cities: Skylines (2015) emerged from the rubble of Maxis’ SimCity reboot, combining approachable city-planning mechanics with modding openness. Developed by the Helsinki-based Colossal Order and published by Paradox Interactive, CS1 quickly became the dominant city builder. Its success spawned dozens of expansions and thousands of user-made mods via Steam Workshop. CO – a studio of around thirty people – became a darling of the simulation genre.
Technical sources: Launch performance warning (GameSpot); CS2 performance analysis (Paavo Huhtala).
In 2023 CO attempted to leap ahead with a sequel. Built in Unity’s High Definition Render Pipeline (HDRP) and promising per-citizen simulation, a dynamic economy and cross-platform modding, CS2 launched on PC in October 2023. Even before release, Paradox warned that performance might not meet players’ expectations. The warning was prescient: the game shipped with heavy GPU bottlenecks, slow simulation speeds and a bare-bones economy. An autopsy by developer Paavo Huhtala found that every pedestrian model had 6,000 vertices (complete with fully modelled teeth) and that props such as pallet stacks were rendered in full detail even when invisible. The engine lacked occlusion culling and relied on high-resolution shadow maps, causing “an innumerable number of draw calls”. The result was a city builder that taxed even high-end GPUs while leaving CPU cores idle.
Player critique: “One Year Later – Cities: Skylines II Is Still a Broken, Lifeless Mess” (Paradox Plaza forums).
Alongside the rendering problems were deeper simulation issues. A year after release one forum thread titled “One Year Later – Cities: Skylines II Is Still a Broken, Lifeless Mess” complained of mindless citizens, dead public spaces and traffic AI that took nonsensical routes. The poster wrote that the sequel’s touted dynamic economy was “nonexistent”. Such criticisms weren’t isolated; they reflected a broader perception that CS2 had shipped as an unfinished Early Access game. CO acknowledged the problems and postponed the console release and paid DLC to focus on patches. Despite multiple updates, players still reported simulation slow-downs and path-finding issues in 2024 and 2025.
Modding coverage: Paradox Mods FAQ (Shacknews); Hallikainen on missing mod support (Game Rant).
Modding – a pillar of the first game – was largely absent. Paradox and CO announced that, unlike CS1’s open Steam Workshop, CS2 would use Paradox Mods, a centralised platform to ensure cross-platform compatibility. In October 2023 Shacknews quoted an official FAQ explaining that mods would be “confined in official capacity to the Paradox Mods platform” because the publisher wanted a single hub accessible on both PC and console. The FAQ went further: “We won’t support other platforms such as Steam Workshop”. This business decision frustrated PC modders and delayed many of the quality-of-life fixes that CS1 had enjoyed through community mods. In February 2024, CO CEO Mariina Hallikainen admitted that the team’s “biggest regret” was launching without mod support; Gamerant summarised her comments, noting that she acknowledged community frustration over the missing Editor and inadequate mod tools.
Paradox’s November 17 2025 update sets out the formal arrangements. The post states that Paradox and Colossal Order “mutually decided to pursue independent paths” and that the decision was taken “thoughtfully and in the interest of both teams”. The Cities: Skylines franchise will move to Iceflake Studios, one of Paradox’s internal management-game teams based in Tampere, Finland. Iceflake will take over “all existing and future development” of CS2, including free updates, ongoing work on the in-game Editor and console versions, and future expansions. CO will deliver one final update, colloquially called the Bike Patch, adding bicycle infrastructure, Old Town buildings and bug fixes. A beta of the asset-editing tools will be released before year-end, after which Iceflake will assume full development duties from the start of 2026.
Statements from the principals frame the split as amicable. Hallikainen thanked Paradox for fifteen years of collaboration and said CO was “excited to channel our experience, creativity, and passion into new projects”. Paradox deputy CEO Mattias Lilja expressed gratitude for CO’s achievements and emphasised Paradox’s commitment to “provide [Cities players] with more content and new experiences”. Iceflake studio manager Lasse Liljedahl called taking the reins “an immense honor and a great responsibility” and said the team sees “a strong foundation and so much potential waiting to be unleashed”. Together, the statements project optimism: the old guard departs gracefully, the publisher pledges continued support, and a new studio vows to unlock the game’s latent promise.
On launching early: Paradox on Cities: Skylines II and iteration regrets (Kotaku).
Yet hidden between the lines is a tacit admission of failure. In an October 2024 interview discussed in Kotaku, Lilja conceded that launching CS2 in October 2023 was a mistake, saying that Paradox and CO were “actually in agreement that iterating this live was probably the right way to go” but that, in hindsight, they “should probably not launch that early”. In other words, the game was knowingly released unfinished with the hope that post-launch patches would complete it; the strategy backfired. By late 2025 the sequel remained tarnished, and shifting development to an internal studio gave Paradox a way to reframe the narrative without cancelling the project.
Several interlocking factors contributed to CO’s struggles with CS2.
The team aimed high: a next-generation city builder with per-citizen simulation, realistic economies and cinematic visuals. But CO was still a 30-person studio – tiny by AAA standards – and Unity HDRP proved unforgiving. The engine’s GPU bottlenecks weren’t the result of exotic path-tracing but of ordinary models being rendered at absurd detail. Buildings and props lacked lower-detail meshes and proper occlusion culling, so millions of polygons were drawn even when off-screen. Shadows were computed at high resolution for every object. These problems could theoretically be solved through asset rework and rendering optimisations, but doing so required months of drudge work and careful pipeline changes – hard tasks for a small team already firefighting bugs.
On the simulation side, CO promised a dynamic economy and deep agent-based behaviours, but the implementation lagged behind the ambition. Players complained that citizens moved like drones, parks were empty and emergency services were purely decorative. Traffic AI took nonsensical routes, and public transport usage barely affected congestion. Economic interactions between industries were shallow, and the employment model produced bizarre labour shortages or surpluses. Fixing such systemic issues often requires redesign rather than quick patches; CO did release an Economy 2.0 update in mid-2024, but by the time of the split the simulation still felt off.
CO was simultaneously developing the PC release, console ports and multiple DLCs while also building an entirely new modding platform. Paradox’s decision to use Paradox Mods for cross-platform compatibility meant that CO had to engineer modding tools that worked on PC, Xbox and PlayStation while meeting console platform security requirements. As the Shacknews article notes, Paradox and CO confirmed that mods would be “confined in official capacity to the Paradox Mods platform” and that there would be no official support for Steam Workshop or Nexus Mods. The rationale was to provide a “centralized, cross-platform hub”, but it removed the de-facto modding infrastructure that had empowered CS1. Building a secure, cross-platform modding system is a multi-year effort; CO underestimated the work and ended up shipping the game without modding tools at all. Hallikainen later called this omission their “biggest regret”.
At the same time, Paradox wanted a steady flow of revenue from DLC and console versions. Lilja’s comments reveal that the publisher deliberately chose to release early and iterate publicly. That strategy might work for small indie games, but CS2’s player base expected a polished sequel, and paying customers became unwilling beta testers. Patches that fixed one issue often introduced new bugs, and repeated delays of the console release eroded trust.
CO’s team had been working on city-builders for over a decade. Burnout and fatigue likely played a role. The company’s history is entwined with the Cities series; moving on allows them to avoid being perpetually defined by “the team that broke Cities” and to experiment with new projects. Their public statements emphasise gratitude and optimism, suggesting that leaving the franchise was as much a relief as a dismissal.
Past work: Surviving the Aftermath review (Screen Rant).
Iceflake Studios isn’t a household name, but it has relevant experience. Founded in 2007 and acquired by Paradox in 2020, Iceflake developed Surviving the Aftermath, a post-apocalyptic colony-builder that entered early access in 2019 and reached full release in November 2021. Screen Rant’s review described it as an “entertaining city-building game” and praised its blend of survival mechanics and management. The game launched rough in early access but steadily improved; by 1.0 it was viewed as “mixed or average” by Metacritic (around 69/100) and maintained a consistent player base. Unlike CS2, its challenges stemmed more from content depth and pacing than from catastrophic performance problems. Iceflake therefore has experience iterating a complex simulation into a stable product.
As an internal studio, Iceflake is directly accountable to Paradox. The publisher can allocate more resources, embed technical specialists and control the roadmap more closely than with an external partner. Iceflake also inherits CS2’s source code, toolchain and documentation. Without the emotional investment that CO had, Iceflake may be more willing to prune systems, simplify mechanics and cut features that don’t work. Liljedahl emphasised that Iceflake sees “a strong foundation and so much potential waiting to be unleashed”. The foundation isn’t nothing: CS2 has larger maps, improved road tools, realistic topography and flexible zoning. If Iceflake can optimise assets, implement proper level-of-detail and occlusion culling and iteratively rework the simulation, the game could reach a state where it’s enjoyable for mainstream players.
However, expectations must be managed. Iceflake cannot rewrite the engine from scratch. The Unity/HDRP foundation, the cross-platform modding constraints and many of the simulation patterns are baked in. The studio will likely focus on performance optimisation, bug fixing and incremental economy/traffic improvements rather than grand redesigns. The Paradox Mods platform will remain the only officially supported mod hub, so deep code mods akin to CS1’s may never return. That’s a business decision that Iceflake cannot overturn.
The publisher’s response to CS2’s troubled launch reveals a broader shift within Paradox. Kotaku’s October 2024 piece notes that Paradox executives have been on an “apology tour” addressing missteps across several projects, including Bloodlines 2, Prison Architect 2 and the cancelled Life By You. Lilja admitted to PC Gamer that they misjudged hardware compatibility and that releasing early was a misstep. By moving CS2 to an internal studio, Paradox signals a desire to control timelines, budgets and quality more tightly. It mirrors similar decisions: Paradox previously shifted development of Bloodlines 2 to a new studio and delayed Prison Architect 2 indefinitely due to technical problems. The company appears to be prioritising quality over rushing sequels out the door.
Paradox has also been transparent about what the short-term roadmap entails: the Bike Patch, asset-mod beta and ongoing console work. After Iceflake takes over, the studio will share its own plans. The messaging emphasises continuity rather than abandonment. There’s no talk of a Cities: Skylines III, and Paradox continues to encourage players to connect their Paradox accounts for cosmetic rewards. Whether this rebuilds trust depends on execution.
Ultimately, Cities: Skylines II is a cautionary tale of ambition outrunning capacity. Colossal Order set out to deliver the most realistic, detailed city-builder ever made but underestimated the technical and design challenges. A small team built an engine that rendered thousands of hidden vertices, shipped without proper mod support and relied on patches to finish the simulation. Paradox, eager to capitalise on the success of CS1, allowed an unfinished game to launch, hoping to “iterate live”. Players rightly rebelled. A year later the sequel still feels unfinished, and the publisher has handed the project to an internal studio while letting the original creators bow out gracefully.
Does this mean CS2 is doomed? Not necessarily. Iceflake inherits a game with a solid core and a passionate community. The studio’s history with Surviving the Aftermath shows it can shepherd a complex management game from rough early access to a polished release. Paradox’s decision to move development in-house suggests a willingness to allocate resources and accept delays. Significant performance fixes – better LODs, occlusion culling, asset optimisation – are engineering tasks that can be accomplished over time. Simulation adjustments to traffic and economy are harder but not impossible. What CS2 will never become is CS1 with all the modding freedom; the Paradox Mods platform and console parity goals make that clear. For players willing to accept that constraint, there is still hope that Iceflake can turn CS2 into a stable, satisfying city builder. The road will be long, but at least the car is now being driven by a team that isn’t running on fumes.
> The engine lacked occlusion culling and relied on high-resolution shadow maps, causing “an innumerable number of draw calls”.
The engine does not lack or cause these things. The fact that the developers chose the HDRP pipeline for this game should be the most obvious dead bird in the entire coal mine. These games should be running on URP without question. We don't need advanced lighting systems in a top down city builder.
If we want an art workflow that allows artists to shit arbitrary content into the editor without thinking, we should probably reach for Unreal and flip on TAA like everyone else is doing.
> We don't need advanced lighting systems in a top down city builder.
You don't need a lot of things. The developers wanted the game to look better than the first, so went with HDRP because it claims to be a production ready pipeline that helps them achieve that.
But it was not and the developers did not have the time, or perhaps skill, to work around it's issues.
> we should probably reach for Unreal and flip on TAA like everyone else is doing.
First of all: "Just ask the entire studio to throw out all existing work and retrain staff to change their engine."
Second of all: Do you mean DLSS? TAA is an AA technique, it does not improve performance.
Third of all: Unreal? The engine notoriously ragged on as dragging down the performance of countless games in the last 2 years because it too has features that are easy to turn on and look good, but require skill and knowledge to fine-tune to be reasonably performant? That Unreal?
But you are rarely looking at it from street level, you spend most of your time in a birds-eye view
I guess you could argue that top down should be defined narrower than that, but the steam tag Top Down is full of games like this [1]
Detailers spend a lot of time in 3D perspectives and zoomed in close so I don't think that this is generally true.
Honestly the gameplay is so shallow I'd spend my time zoomed in too. Talk about a series where every release went down hill. The original Cities in Motion was the best of the 4.
The story is familiar: small team nails a niche, publisher scales expectations, sequel inherits AAA scope without AAA staff. Ten years later we call it "mismanagement", but really it's the same incentive loop that breaks most creative partnerships once success hits Excel.
Yep. It's the curse of too much money.
Just because you hit on something and gamers threw their money at you because you deserved it, it doesn't mean the next iteration has to have MORE OF EVERYTHING.
Even some series that have maintained quality have got a bit too big for their own good if you ask me. Did Horizon Forbidden West need to be that big? Zero Dawn was the perfect length if you ask me.
Even Witcher 3 has a faint whiff of 'it could have been a bit shorter and still brilliant'.
I'm not sure it's always the publisher's fault though. Success and the worldwide obsession for cancerous business growth can go to your head even without outside pressure.
It's also strange to do with with a city builder.
Now Skylines certainly uped the game game in graphics, but honestly I would pay good money for an updated Sim City 4 or ... Sim City 3000.
A city builder doesn't have to LOOK amazing to be great.
I still don't know why we needed a sequel... Couldn't they just keep working on the original game, which already worked really well and lots of people loved? I had similar feelings about kerbal space program, but at least there it's somewhat understandable, given the jank that crept in over time
KSP2 made some degree of sense: the game had outgrown its engine and architecture, so you start fresh with a bigger dream.
But before that had a chance to fail from second system syndrome it was doomed to fail by insane demands from Take 2. News of work on KSP2 could harm sales of KSP1, so when hiring people to work on KSP2 they couldn't mention what they were hiring for. So you had a team who didn't know KSP1, and due to budget constraints were mostly juniors. Then to "save time" they were not allowed to only pick the good parts of the old source code or to even switch engine, they were supposed to just expand the janky KSP1 code base. Obviously without being allowed to talk to the developers of that code base, because secrecy. And no talking to fans about what they would want from a KSP2 either, because, you guessed it, secrecy.
So an inexperienced team disconnected from the fan base was supposed to fix a code base they were not familiar with, without speaking to the people who wrote it, add some cool features to it that the original team never tackled due to engine limitations, and release it to massive fanfare. Surprisingly this did not work. As the project was failing went back on many of those decisions, but it's hard to fix a project that starts off so wrong
Compare to Kitten Space Agency: hire KSP1 devs and KSP1 modders so you have people who can judge what worked and what didn't, start with a home-grown engine that fits the unique demands of a KSP-like game, talk with the community during development. Obviously they aren't far enough along yet to call it a success, but I give them much better chances
Ironically, the KSP 2 studio in all their secrecy couldn't help but post multiple videos as early access approached, praising themselves for how innovative and smart they were - how they're so inspired and it's gonna be amazing.
I'm really excited for KSA, hoping this is finally the sequel KSP1 deserves!
You can only sell a game once. Once you have your customers' money, you've achieved your goal. What else is there to do? DLC has a hard cap on your possible sales...
You could work on a totally new game, but, I think companies are looking to cut costs by reusing content.
Factorio 2.0 seemed to pull it off. I think that as long as users don’t feel misled by a DLC that only adds a few skins, they generally appreciate larger updates to a game.
> DLC has a hard cap on your possible sales...
DLCs have a hard cap on possible sales in the same way a new game has a hard cap on possible sales. DLC releases can, and do, bring in new players to the base game.
or you go "online" and milk customers for decade? yeah that is done by Rockstar.
Development teams are expensive to fund, and people who have bought a game will pay full price for a sequel, but won’t pay full price for updates/DLC.
And releasing a sequel gets you hype and press coverage - potentially expanding your customer base - in a way that releasing updates won’t.
There are some exceptions (No Man’s Sky?) but they are very few and far between.
Well ok, I know "value to shareholders" is a good enough reason for some people... I guess I'm not thinking capitalistically enough about stuff
The point above wasn’t about value to shareholders but rather about being able to pay the people doing the actual work.
Even before you get to "value to shareholders" you have to actually pay your developer salaries to update the game. Where does that money come from when you're updating a game that's been on the market for 10 years and sales of new copies have tapered off?
Free major updates make your existing customers happy but don't pay salaries. This is why so many games have moved to some kind of ongoing revenue model with Battle Passes, cosmetics, item marketplaces, etc.
This is one reason why we should either cut back copyright to more like 10-15 years and require source escrow (so public domain materials can come with source) to obtain copyright, or just require all computer programs come with source code as part of consumer protection laws. Then people can fix the engine themselves, or find a way to fund someone to do it.
Or just eliminate copyright entirely and focus on economic models that are based on funding creation. You raise money to build the thing, and once it's built, it's there for all.
In most cases I would agree with you but ultimately games get older and can’t be sustained forever without people being compensated. People don’t pay for DLC like they pay for sequels, as the other person said.
It’s not about shareholders necessarily. It’s also about sustainability and people paying bills - they live in a capitalist society and can’t choose not to participate at the end of the day. You can’t ask a dozen or more developers to keep working on a game for free for a decade or more. They have to eat too.
The only other option is keep playing the exact same game with little to no changes. Which you can! The original is still available. But if you want it to improve and change over time or receive substantial DLC’s, somebody has to get paid at some point.
I don’t want to harp on as you had a couple answers on this already but if you need to pay your devs, what is your suggested alternative to “having money in the bank”? The latter only happens with more sales, and that only happens if you have something to sell.
People will pay for DLC what the DLC is worth, which should in theory be directly proportional to how much effort it makes to produce the DLC. 4 small $20 expansions could be much more lucrative than an $80 new game which needs to not only include those changes but also make the rest of a functional (and presumably higher quality than the original) game.
> Development teams are expensive to fund, and people who have bought a game will pay full price for a sequel, but won’t pay full price for updates/DLC.
So, you never fall in the trap of Paradox Games and the eternal launch of DLCs for Stellaris/Victoria/Hearths Of Iron/etc?
> will pay full price for a sequel, but won’t pay full price for updates/DLC.
I'm not sure this is true, see Factorio as an example. They released Space Age as a "DLC" but for full price and with clear messaging that it's version 2.0 of the game.
To claim Factorio Space Age as counter example would require to show that it was as successful as DLC as it would have been as a new game. Probably not easy to show...
I think that's impossible to prove and I don't see how it's relevant. The OP claimed that people aren't willing to pay full price unless it's a clear, separate sequel with hype around it. The devs of Factorio proved otherwise with 400k copies being sold in the first week of Space Age's release[0].
Rust (not the language) is another good exception that is mostly powered by DLCs and skins today. Continuous updates with balance changes keep the game fresh, ensuring you maintain your playerbase that will in turn buy DLCs.
Both could definitely have used a completely updated engine at the very least (not just graphics, but scaling/capabilities around the core gameplay had grown quite a bit beyond what made sense originally), which would enable a lot of things which weren't as feasible in the original games, but it's hard to do that kind of reset and match 10 years of building and tweaking on the original. Hopefully KSA (Kitten Space Agency) can have better luck.
As far as I remember, KSP did keep pretty on top of engine updates. And I never thought the graphics were that bad, TBH. Sure, its not raytraced to hell and back, but I thought it looked just fine. However, there was a lot of physics jankery that never really got fixed (the kraken likes to eat complicated ships), and it did have performance issues in some areas. I think the community kinda wanted multiplayer and colonisation too, and the codebase was probably getting quite mangled and convoluted, making that hard. It would have been nice to see Squad get some time to be able to rework their systems over the course of a few updates, not really adding many features but focusing on performance and de-spaghettification, but I'm guessing Take-Two wouldn't see it the same way and just wanted MORE CONTENT, MORE SALES, MORE PROFIT
> As far as I remember, KSP did keep pretty on top of engine updates. And I never thought the graphics were that bad, TBH. Sure, its not raytraced to hell and back, but I thought it looked just fine. However, there was a lot of physics jankery that never really got fixed
This is what I mean by engine updates that aren't just about graphics but the scaling/capabilities around the core gameplay. Updating Unity again or adding raytracing graphics wouldn't have fixed the actual problems with the rest of the engine.
There isn't much of a business model in paying developers for a year for updates that won't generate more revenue.
If I told my boss that I wanted to spend 12 months refactoring our entire system in ways that would benefit our existing customers (who have paid once and won't ever pay again) but likely result in no additional revenue being generated, I doubt that project would be green-lit.
There definitely were improvements to be made in both KSP (physics of large vessels) and CS (FPS in large cities).
Instead we get this... 0/2
Kitten Space Agency is looking like it will be the KSP2 we deserved but didn't get
> Couldn't they just keep working on the original game, which already worked really well and lots of people loved?
Of course. If history (and "IME") is any guide, this was all marketing and product manager driven. Creators of all stripes now create the thing THEY want, not the thing paying CUSTOMERS want.
That's perhaps overly cynical and sometimes this is not the case if some new set of features can't be easily done in the architecture of the original, of course.
He mentions Prison Architect 2, which like Kerbal switched studios for the sequel and ended up an unfinished mess that's objectively worse than the original. Meanwhile, Rimworld is raking in the cash (presumably) by continuing to make popular DLCs for a 12-year-old game! But it sounds like they wanted to go big, and I kind of get it since graphics matter a lot more for a city builder than a lot of other simulation games.
The interesting part about Cities 2 is that the simulation is much more in-depth: pops have a real job where they commute to (versus taking any available one in the first game), they don't just teleport around, companies have to import&export resources and make profit based on that, etc.
Also the graphics/lighting seems much improved with a more realistic art style.
Both things which you cannot really retrofit into Cities 1.
> I still don't know why we needed a sequel...
$$$. I think they need to design a long-term monetization strategy that does not require new major releases, but rather just more DLC, seasons, etc.