Amazon Game Studios Exits MMORPGs: A Decade of Burning Player Trust

The prehistory: Breakaway and Crucible

Before Amazon ever touched an MMORPG, it had already established a pattern. In 2016, Amazon Game Studios announced a trio of ambitious PC titles built around Twitch integration: New World, Crucible, and Breakaway. The message was clear: Amazon was not dipping a toe into gaming. It was building a platform.

Breakaway was a 4v4 “esport brawler” that Amazon pitched with a live tournament on a 355-foot naval patrol ship. It was cancelled in 2018 before it ever reached a meaningful public release. The team claimed they were “setting a very high bar.” The reality was simpler: the game was not fun enough to justify continued investment.

Crucible was worse. Amazon’s first major original release, a free-to-play third-person shooter, launched in May 2020 to near-total silence. Player counts cratered within weeks. In June, Amazon “un-released” the game, pulling it back into closed beta in a move so bizarre it became industry shorthand for corporate panic. By October, Crucible was dead. It had lasted five months from launch to cancellation, four of them in a state of public beta purgatory.

Between Crucible’s death and New World’s launch, Amazon also quietly killed its Western publishing deal for Blue Protocol, the anime MMORPG from Bandai Namco. It was a mercy killing, perhaps, but it was also another data point in a growing trend. Amazon Game Studios was very good at announcing games and very bad at keeping them alive.

Game One: New World and the 913,000-player mirage

New World launched on September 28, 2021, and for one glorious week, it looked like Amazon had finally cracked the code. 913,027 concurrent players on Steam. Over a million at peak across all platforms. Queue times measured in hours. Streamers fighting to get into Aeternum. The game was beautiful, the combat was satisfying, and the territory control system felt genuinely novel. For a moment, it seemed like Amazon had built the first true next-generation MMORPG.

The problems emerged almost immediately. At launch, New World allowed players to duplicate gold through a simple trading post exploit. Endgame content was thin: a handful of expeditions, repetitive elite zones, and a crafting system that funneled everyone into the same material bottlenecks. The PvP territory war system, which was supposed to be the game’s unique selling point, devolved into off-peak timezone captures by the largest guilds. Wars were scheduled at times convenient for the defending company, meaning most players never got to participate in the game’s central competitive feature.

What Amazon promised vs. what players got

The player exodus was brutal and well-documented. Within three months, New World had lost over 75% of its peak concurrent players. By six months, it was below 20,000. The Steam review score settled into “Mixed” territory and never recovered. Amazon tried everything: server merges, new weapons, new zones, a raised level cap, a seasonal model. Nothing moved the needle meaningfully.

The Aeternum relaunch: a second chance squandered

In October 2024, Amazon attempted what is rarely attempted in MMORPGs: a full relaunch. New World: Aeternum was not just an expansion. It was a rebranding, a console port, a mechanical overhaul, and a narrative reset. New voice acting. New storylines. New solo scaling. PlayStation 5 and Xbox Series X/S versions. This was Amazon’s Final Fantasy XIV moment, or so they hoped.

It was not Final Fantasy XIV. Reviews remained stubbornly “Mixed” at 67% positive. Veterans felt betrayed by systems changes that invalidated their years of progress. Console players found a game that still felt like a PC port from 2021. And the fundamental problem, that there was simply not enough compelling content to justify an ongoing subscription to your time, remained unaddressed.

By August 2026, New World: Aeternum sits at roughly 423 concurrent players on Steam with a 24-hour peak of 545. From 913,000 to 545. That is not a decline. That is erasure. The game has lost 99.95% of its peak Steam audience. And unlike Lost Ark or Throne and Liberty, there is no Korean developer waiting to take New World back. It is Amazon’s corpse to bury.

Game Two: Lost Ark and the bot apocalypse

Lost Ark should have been Amazon’s redemption story. Unlike New World, which Amazon developed internally, Lost Ark was a proven product. Smilegate’s action MMORPG had been running successfully in Korea since 2019, refining its systems, building its endgame, and establishing itself as one of the most polished ARPGs ever made. Amazon’s job was not to build the game. It was to publish it in the West, run the servers, manage the community, and keep the bots out.

On February 11, 2022, Lost Ark launched on Steam and hit 1,325,305 concurrent players. It was the second-biggest launch in Steam history at the time, behind only PUBG. The action combat was sensational. The class variety was deep. The horizontal progression systems, collectible hunting, and island exploration gave players hundreds of hours of content before they ever touched the controversial gear honing system. For the first month, it felt like Amazon had finally done something right.

The bot problem

Then the bots arrived. Not dozens. Not hundreds. Millions. Lost Ark’s free-to-play model, its gold-based economy, and its lack of meaningful account verification at launch created the perfect environment for automated farming. Bots filled every starting zone. They clogged dungeon queues. They flooded the auction house with illegally obtained gold, crashing prices for legitimate players and inflating the cost of everything that mattered.

Amazon’s response was slow, then inadequate, then farcical. In March 2022, the company banned “over a million” accounts. The bot population barely dipped. In 2023, another wave of bans hit inactive accounts, some belonging to legitimate players who had simply taken a break. The community was outraged. By March 2025, Amazon was issuing statements about “recent ban waves” that had removed hundreds of thousands more accounts. In March 2026, Amazon acknowledged the “frustration within the community” and announced it had permanently banned “several million” bot accounts over the game’s lifetime.

Think about that timeline. Four years. Several million bans. And bots were still a problem significant enough that Amazon had to publicly acknowledge it in a farewell-adjacent blog post. This is not a technical failure. It is a operational failure so comprehensive it becomes a moral one. When your players cannot log into a starting city without wading through armies of identical berserkers auto-running the same path for four years, you are telling those players that their experience does not matter as much as your margin.

Monetization and the honing wall

The bot problem was compounded by Lost Ark’s monetization, which Amazon inherited from Smilegate but was responsible for communicating and contextualizing to a Western audience. The gear honing system, which required players to spend gold and materials to upgrade equipment with a chance to fail, was controversial in Korea and became radioactive in the West. Players who had enjoyed hundreds of hours of horizontal content suddenly hit a vertical wall where progression slowed to a crawl without either extreme grinding or real-money purchases.

Amazon’s communication around this was consistently poor. When the Western version launched without certain quality-of-life features present in the Korean client, players accused Amazon of deliberately creating a more grind-heavy experience to drive microtransaction sales. Whether this was true or not, the perception took hold. And in free-to-play games, perception is the economy.

By August 2026, Lost Ark’s Steam concurrent count averages around 5,300 with a 24-hour peak of roughly 6,700. From 1.3 million. That is a 99.5% collapse. And the final indignity: Amazon announced it is handing Western operations back to Smilegate in early 2027 and merging the remaining servers down to just five. Five servers for a game that once had 1.3 million concurrent players. The symbolism is almost too perfect.

Game Three: Throne and Liberty and the pay-to-win trap

Throne and Liberty launched on October 1, 2024, and once again, the numbers were extraordinary. 333,393 concurrent players on Steam at peak. NC’s Unreal Engine-powered MMORPG was gorgeous, ambitious, and positioned as the spiritual successor to Lineage. Amazon had learned from Lost Ark, surely. This time, the publisher would get it right.

It did not. Within four months, Throne and Liberty had lost roughly 90% of its players. By July 2025, it was down 95%. By August 2026, the game hovers around 5,000 concurrent players with a 24-hour peak of roughly 7,000. The Steam reviews sit at 68% “Mixed,” with recent reviews dipping into truly negative territory.

What went wrong?

The problems were different from Lost Ark’s but no less fatal. Where Lost Ark was killed by bots and a progression wall, Throne and Liberty was strangled by its own monetization and systems design. The game was aggressively free-to-play on the surface and punishingly pay-to-advance at its core. The Rune system, which allowed players to enhance gear through randomized upgrades, became a focal point of community rage. Daily contracts felt like a checklist rather than an adventure. Guild schedules became mandatory rather than optional. The game asked for your time, your money, and your social availability, and offered diminishing returns on all three.

Amazon’s handling of the Western version compounded these issues. The publisher cut servers in early 2026 in a cost-saving measure that disproportionately affected PlayStation 5 players. The community felt, with justification, that Amazon was managing decline rather than investing in recovery. When a publisher starts removing infrastructure from a one-year-old MMORPG, players understand exactly what message is being sent.

On August 12, 2026, Amazon announced that Throne and Liberty would transition to NC by the end of Q4 2026. The announcement was polite, corporate, and utterly damning. Amazon had given up on its newest MMORPG before it had even reached its second anniversary.

The August 12, 2026 announcement: Amazon’s quiet surrender

The dual announcements on August 12 were masterpieces of corporate euphemism. Lost Ark’s blog post spoke of “transitioning live service operations” after “nearly five years of shared adventures.” Throne and Liberty’s announcement described a “transition from Amazon Game Studios to NC by end of Q4 2026.” Neither post used the words “failure,” “exit,” or “giving up.” But the message was unmistakable.

Amazon Game Studios, the division that Jeff Bezos reportedly told to “win at games,” had lost every MMORPG it ever touched. New World: dead. Lost Ark: handed back. Throne and Liberty: handed back. Blue Protocol: cancelled before launch. The only remaining question was whether Amazon would try a different genre or finally admit that live-service game publishing might not be something you can solve with AWS infrastructure and A/B testing.

The timing was particularly brutal for Lost Ark players. Amazon’s announcement came alongside news of a final server merge down to five total servers. Players who had stayed loyal through years of bot plagues, ban waves, and monetization controversies were now being told that their publisher was leaving before their game was. Smilegate taking over may ultimately be better for the Western version, but the manner of the handoff, announced in the same breath as server consolidation, felt less like a transition and more like a retreat.

What is the pattern? Three games, one failure mode

Amazon’s MMORPG collapse is not a story about three unlucky games. It is a story about one company that never learned how to operate live services at the human level. The pattern is consistent across New World, Lost Ark, and Throne and Liberty:

1. Amazon treats launch as the finish line, not the starting line

All three games had extraordinary launches. Amazon is exceptional at marketing, platform placement, and generating day-one buzz. What it cannot do is sustain. New World’s endgame was empty within weeks. Lost Ark’s bot problem was allowed to fester for years. Throne and Liberty’s monetization was not adjusted despite months of community feedback. In each case, the post-launch operation was an afterthought.

2. Amazon does not solve problems, it manages their optics

Consider the bot issue. Amazon did not ignore it. The company issued ban waves, posted updates, and acknowledged “frustration.” But it never solved it. The same bots, the same gold farmers, the same auction house manipulation continued for four years because Amazon’s approach was reactive and public-facing rather than structural. When your anti-bot strategy is measured in press releases rather than player experience, you are not running a live service. You are running a PR campaign with a game attached.

3. Amazon’s monetization philosophy alienates its core audience

New World was buy-to-play with a cash shop that felt increasingly necessary as the economy broke. Lost Ark was free-to-play with a honing system that pushed players toward real-money solutions at endgame. Throne and Liberty was free-to-play with a Rune enhancement system widely described as pay-to-win. In all three cases, the monetization model created friction between what players wanted to do and what the game allowed them to do without paying. The richest company in the world kept making players feel like they were being squeezed.

4. Amazon’s corporate culture is incompatible with MMORPG community management

MMORPGs are not products. They are relationships. They require publishers to listen, adapt, apologize, and occasionally sacrifice short-term revenue for long-term trust. Amazon’s corporate culture, optimized for efficiency metrics and quarterly targets, struggles with the messy, emotional, unpredictable reality of community management. When Lost Ark players screamed about bots for four years, Amazon responded with the pace of a company that had other priorities. Because it did. Amazon Game Studios was never important enough to Amazon to get the resources and attention that live services demand.

5. Amazon overestimates its own competence

This is the deepest pattern. Amazon believed that its technical infrastructure, its data analytics, and its platform power would translate into gaming success. They did not. Running a global e-commerce platform and running a live-service MMORPG require different skills, different patience, and different values. Amazon’s assumption that it could apply its general corporate excellence to a specialized creative industry led directly to a decade of failures.

The full Amazon Game Studios graveyard

That is six major gaming initiatives across eight years. Zero enduring successes. One internal development that collapsed. Two Korean partnerships that ended in retreat. One cancelled Western localization. And one shooter so mishandled it became a verb in game development circles: “to Crucible” something is to panic-un-release it.

What happens now?

For the games themselves, the future is mixed. Smilegate taking direct control of Lost Ark’s Western operations may actually improve things. The Korean developer understands its own game better than Amazon ever did, and without a publisher middleman, decisions about content pacing, monetization, and bot prevention may happen faster and more effectively. NC taking over Throne and Liberty similarly removes a layer of corporate abstraction. Whether either game can recover from years of damage is uncertain, but the handoff is not inherently bad for players.

For Amazon, the question is whether it learns or repeats. The company still operates games. It still has resources that dwarf almost any other publisher. But its track record suggests a fundamental mismatch between Amazon’s corporate DNA and the requirements of live-service gaming. You cannot A/B test community trust. You cannot machine-learn your way out of a bot apocalypse. You cannot optimize a relationship.

The tragedy of Amazon’s MMORPG decade is that the company had everything it needed to succeed except the one thing that mattered: the willingness to treat players as humans rather than metrics. New World, Lost Ark, and Throne and Liberty did not die because they were bad games. They died because their publisher kept making the same mistake, over and over, across three different titles, two different business models, and nearly a decade of opportunities to learn.

For players looking for MMORPGs that are actually being cared for by their publishers, the landscape is not as bleak as Amazon’s exit suggests. Games like Final Fantasy XIV, Guild Wars 2, and Old School RuneScape continue to demonstrate that the genre works when the relationship between developer and community is treated as sacred. If you want to understand how other live-service games succeed and fail in the current market, our breakdown of Marathon’s collapse traces a similar pattern in the extraction shooter space, while our best games to play this month highlights titles that respect your time. For a broader look at how gaming businesses actually sustain themselves, see our analysis of how esports teams make money and what separates sustainable models from burning piles of player goodwill.

Frequently Asked Questions