> ## Content Index
> Fetch the complete content index at: https://www.regenraygun.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# Meta's AI gaming push
- URL: https://www.regenraygun.com/meta-ai-gaming-push-horizon-create-studio/
- Published: 2026-09-28T19:16:29.000Z
- Updated: 2026-10-21T19:17:18.000Z
- Description: Meta has an AI game creation strategy and distribution at scale.
- Author: Jon Jordan
- Tags: Meta

Meta opened a waitlist on 24th September for two game creation tools announced at Connect. 

- **Horizon Create** is a mobile app: describe a game in plain language on your phone and it builds one.
- **Horizon Studio** is the browser version, adding real visual editing — placing scenes by hand, swapping assets from a library, adjusting script parameters.

Both run on Meta's own Horizon Engine, share a backend so projects can move between them with iteration history intact, and are claimed to produce complete 2D or 3D mobile games with progression systems, balanced difficulty, art direction and multiplayer.

Unity fell on the news. Meta rose as much as 4.5%, to its highest level in a year.

The coverage has framed this as AI coming for game development. That reading misses what Meta actually announced, because prompt-to-game is not new and Meta is not first to it. Roblox has creation AI. Rosebud has it. Combos was demonstrating precisely this on the Gamescom floor five weeks ago, with games built by people who had never made one.

What none of them have is a feed.

Games built in Horizon Create and Horizon Studio become eligible for native distribution across Facebook, Instagram and Horizon. That is the sentence the market reacted to. It is also the one thing Unity, Unreal and Godot cannot answer, because none of them own their audience.

Saxs Persson, Meta's VP of Horizon, described the goal as letting creators get an idea in front of players without spending months on tools, hardware and marketing. Note the third item on that list. Meta is not primarily selling a cheaper way to build a game. It is selling a shorter path to players — and it owns the path.

Count what that involves. Meta supplies the creation tool, the engine, the runtime, the distribution surface, the discovery ranking that decides who sees what, the content safety review that gates publication, the automated age rating, and the monetization terms. Every layer between a creator's idea and a player's attention belongs to one company.

The discovery language in Meta's own announcement is worth reading closely. Games that are engaging, stable, culturally relevant and good at retention will earn more reach. It's not a marketplace. It is an algorithm, and it is the same algorithm that has governed Facebook Pages, Instagram publishers and Reels creators through a decade of reach changes that nobody outside Meta could predict or appeal.

Monetization, meanwhile, is unspecified. Creators may earn through in-game purchases and something called the Meta Horizon Creator Fund, with details promised later.

As for the competition, Unity has now been knocked back twice in 2026 by announcements it does not directly compete with.

In January, Google demonstrated Project Genie, which generated explorable 3D environments from text prompts. Unity shares fell nearly 26% in a morning; the wider selloff briefly erased around $47 billion from video game stocks. CEO Matt Bromberg argued such tools expand creative possibilities without replacing engines. Jefferies took the more durable line: Unity was likelier to integrate generative models into its own tools than be displaced by them.

Unity then did exactly that. On 9th September it shipped [an official first-party plugin for Claude Code](https://www.regenraygun.com/unitys-latest-customer-is-claude/) — 29 Unity-authored skills, the Unity CLI, an MCP server giving the agent live Editor control. 

Its director of AI products, Rachel Zhao, explained the reasoning with unusual candor: developers increasingly begin by describing a game to a coding agent rather than downloading an engine, and the agent picks the engine for them, tending to choose whatever it can install and run inside that same session. The plugin exists to make sure the agent picks Unity and then uses it properly.

Nine days later, Meta announced a product in which there is no engine to pick.

That is the significance, and it is not the same as "AI replaces game engines." Unity's strategy is a bet on winning a selection event. Meta's architecture deletes the selection event. A creator who opens Horizon Create never chooses an engine, never installs a toolchain, never publishes to a storefront, and never encounters a decision Unity could have won.

Reported figures for Unity's fall on 24th September range from 5% to 6.9% intraday to 10.59% for the session, depending on source. And Meta was one of four cited drivers, alongside profit-taking on the wind-down of Unity's legacy advertising business, rising long-term Treasury yields pressuring growth software, and a stock that was technically overbought after a strong run. Meta supplied the catalyst, not the entire explanation.

Larger caveat: Meta hasn't shipped anything. Their is a waitlist with gradual early access and no general availability date. Every claim about output quality — balanced difficulty, coherent art direction, working multiplayer — is Meta's own, with no independent verification. And the track record here is *Horizon Worlds*, a platform whose struggles are well documented and which this effectively relaunches.

But the direction is legible regardless of whether these particular tools work. The interesting contest in game creation is no longer about which engine, or even which model. It is about who stands between a creator and an audience.