Watching something fail—not because it failed on its own terms, but because something completely outside of its control decided it no longer mattered—causes a particular kind of dread. That’s essentially what happened to Stormgate, a real-time strategy game created by some of the same Blizzard employees who once shaped StarCraft 2, when a quietly devastating announcement appeared in its Discord server towards the end of March 2026.
The message was delivered in a succinct, almost professional manner. Stormgate’s developer, Frost Giant Studios, notified the community that Fireworks AI, an AI company, had acquired Hathora, the company that operated its online game servers. By the end of April, Hathora was closing its gaming services. The primary motivation for most players, online multiplayer, would cease to exist. The studio stated that it hoped to eventually bring things back. However, the wording was cautious and hesitant, the kind of language people use when they are genuinely unsure of how things will turn out.
| Field | Details |
|---|---|
| Game Title | Stormgate |
| Developer | Frost Giant Studios |
| Genre | Real-Time Strategy (RTS) |
| Early Access Launch | August 2024 |
| Full Release (v1.0) | August 2025 |
| Business Model | Free-to-play |
| Kickstarter Funding (2023) | Over $2.3 million raised |
| Server Partner (former) | Hathora |
| Acquiring Company | Fireworks AI (founded by ex-Meta AI division employees) |
| Fireworks AI Customer Base | 10,000+ customers (AI code assistance, agentic systems, conversational bots) |
| Multiplayer Shutdown Date | End of April 2026 |
| Other Affected Games | Splitgate 2, Predecessor |
| Developer Response | Patching offline mode; searching for new server partner |
| Online Mode Status | Uncertain — dependent on finding new hosting provider |
By then, Stormgate’s life had already been complicated. With a 2023 Kickstarter campaign that raised over $2.3 million and an aesthetic reminiscent of Blizzard’s golden age, it entered early access in August 2024 with high expectations. Veterans leaned in, recalling Warcraft III and StarCraft. Reviews of the early access build on Steam were mixed; it wasn’t a huge success as the studio had hoped, but it wasn’t terrible either. Instead of stepping outside of StarCraft’s shadow, some players thought the game relied too much on it. The developers updated, listened, and made adjustments. Even though the larger gaming community had not yet fully embraced it, a genuine community was emerging.
Then, subtly, the AI industry’s insatiable demand for computing infrastructure pulled the rug out from underneath all of that, not because of poor sales or poor design. Hathora’s servers, processing power, and network capacity were necessary for Fireworks AI, a company that provides tools ranging from conversational agents to code assistance to about 10,000 clients. Simply put, the requirements of an AI platform were greater than those of a mid-sized role-playing game with a devoted but small player base. The new owners of Hathora might not have given Stormgate a second thought. Perhaps the most disturbing aspect is that.
A curtain that many in the gaming industry probably wished had remained closed was pulled back by the circumstances. For multiplayer hosting, smaller studios have always depended on outside infrastructure firms. It’s a sensible choice because creating and managing your own server network is costly and difficult, particularly for a studio without the financial resources of a publisher. That issue was manageable thanks to services like Hathora. However, this specific arrangement was predicated on the unstated assumption that the organization managing your servers would still be interested in managing your servers. That assumption was never guaranteed, as demonstrated by the Fireworks AI acquisition.
The damage wasn’t even limited to Stormgate. According to reports, Splitgate 2 and Predecessor were also supported by Hathora’s infrastructure, indicating that several games and their communities were simultaneously impacted by a single corporate acquisition. As this develops, there’s a sense that the industry has a sneak peek at something it should probably start taking seriously. When the demand for GPU clusters and cloud capacity in the AI industry begins to directly compete with large-scale gaming infrastructure, what will happen? Whether this is an isolated incident or the first sign of something bigger is still unknown.
The AI boom has already had a noticeable impact on industries, increasing the cost of chips, changing the way developers approach creative work, and causing anxiety in a variety of creative fields. However, this specific collision was unique. It wasn’t abstract. Because the underlying architecture of the internet is subtly being reorganized around different priorities, real players who booted up Stormgate one day expecting a match would soon find that option simply gone, not because of anything they or the developers did wrong. For video games, that is a truly novel kind of issue.

Frost Giant Studios is still going strong. The team is actively searching for a new server hosting partner while patching in an offline mode to keep the game playable in some capacity. However, the search is truly unpredictable. Their statement stated, “This work will be dependent on Frost Giant finding a partner to support ongoing operations,” which says a lot. Stormgate’s online future now rests on whether another infrastructure company finds it profitable.
The larger lesson is difficult to ignore, regardless of Stormgate’s final destiny. When developing multiplayer games based on third-party hosting agreements, game studios might want to pay closer attention to the stability and independence of those partners. There is only one direction for the infrastructure requirements of the AI industry, and it is not slowing down. The risk is significant for smaller developers who lack the power to include survivability clauses in hosting contracts. Stormgate was the first to arrive.


