AI content co-creation platform "FunloomAI" secures tens of millions in Pre-Series A funding, returning creativity to its core | 36Kr First Release

This is a funding article detailing a startup’s capital infusion and strategic positioning within the AI and gaming intersection. The analysis will focus on the company’s claimed paradigm shift, its technical differentiation, and its early market validation.

A Strategic Pivot from Pure Code Generation

Funloom’s founding team identified a critical flaw in the initial “vibe coding” approach: most users lack the prompt engineering skills or domain knowledge to translate a vague idea into a polished game. The founder, Wu Tong, observed that people can articulate a feeling but current tools cannot interpret it. This led to a pivot away from pursuing a mythical “one-step-to-AAA” generative AI. Instead, the strategy became iterative, stepping-stone development aimed at stripping away all non-creative overhead for existing creators, rather than creating creators from scratch. This pragmatic focus on serving those with ideas but technical or financial barriers distinguishes their approach.

Deconstructing Game Creation into AI-Powered “Atomic Capabilities”

Funloom’s product, Funloom AI, decomposes a game’s core into four atomic capabilities, each tackled sequentially:

1. Storyline: Already deployed. The team uses “cognitive engineering,” collaborating with professional writers to abstract narrative structures (conflict, foreshadowing) into training patterns, producing plots that avoid clichés.
2. Visuals: The next frontier. While AI image generation is mature, their focus is on automating the integration of assets—replacing rough prototypes with rendered art by identifying and swapping UI elements and backgrounds.
3. Numbers & Interaction: Under development. Notably, Funloom rejects treating numerical systems (like character stats) as a pure optimization problem. Instead, they view numerics as an extension of narrative, tied directly to managing player “flow state” and emotional pacing—creating bottlenecks or power spikes in sync with the story.

This modular, problem-focused framework prioritizes elevating creative expression over technical completion.

Validating a Niche Market and a Sustainable Business Model

The article highlights early traction that challenges common assumptions in AI gaming:

• Demand Validation: Within a month of testing, the platform attracted over 5,000 creators and tens of thousands of players. A single user-generated title, Chongzhen Simulator, drew nearly 20,000 players, spawning derivative simulations set in other dynasties. This proves latent demand for niche historical and thematic content that traditional studios de-risk away from.
• Business Model Validation: Unlike platforms relying on free tiers to build scale, Funloom implemented a token-based monetization system from the outset. Users spend tokens to create and play, achieving a daily average customer price of 25 RMB. This early paid model successfully filters for serious users—those with customization needs who are both creators and consumers—creating a self-sustaining commercial loop. The COO argues that this niche of serving small creators with imaginative projects is structurally defended against large studios, whose risk-averse, high-budget models favor proven formulas.