618 2026: GEO Goes Mainstream — Brands Are Now Paying for AI Discoverability

On June 13, 2026, 36Kr published an article titled "This 618, Brands Are Lighting Incense for GEO" — a reference to the traditional Chinese practice of burning incense for good fortune. The message was clear: as 618 enters its final stretch, brands are no longer relying on luck. They are investing in a structured, measurable discipline called Generative Engine Optimization, and they are doing it at scale.

The shift from concept to commerce has happened faster than most analysts predicted. In early 2025, GEO was a theoretical discussion at AI conferences. By 618 2026, it had become a line item in marketing budgets — with publicly listed companies forming dedicated GEO departments, cross-border e-commerce agencies adding GEO service lines, and beauty brands already reporting conversion from GEO-driven AI traffic.

💰 The Numbers That Define the Shift

The 36Kr article provided the first concrete pricing data for China's GEO services market:

At RMB 30,000-60,000 per brand per year, GEO is positioned between SEO (cheaper, less effective in the AI era) and SEM (more expensive, transactional). The pricing sweet spot suggests the market believes GEO delivers measurable ROI — and the early data supports it.

Service providers interviewed by 36Kr reported that brands who began GEO work in 2025 are seeing tangible sales conversion from AI-generated recommendations during this 618 cycle. The lead time matters: GEO is described by practitioners as a "long-cycle" investment, not a "quick win."

🛠️ How GEO Works in Practice

The article details a six-step GEO service process that has become the industry standard:

1. Brand needs analysis — mapping the brand's category, competitive position, and target AI platforms
2. Keyword and question scenario mapping — building a matrix of natural-language queries that users ask AI assistants
3. Original content production — creating fact-based, structured content that addresses each query scenario
4. Authoritative channel distribution — publishing content on high-authority platforms that AI models trust
5. AI monitoring — tracking whether brand content is being referenced by major AI models
6. Delivery and verification — confirming that pre-defined queries return positive brand references

A concrete example: for a robot vacuum brand, the GEO process begins by mapping queries like "best robot vacuum under RMB 2,000," "robot vacuum for pet-owning households," and "2026 robot vacuum value rankings." Each query receives a dedicated piece of content with factual specifications, user reviews, and authoritative citations.

⚖️ The Question of Legitimacy

The 36Kr article draws a critical distinction that will shape the GEO industry's development:

Legitimate GEO is built on real brand assets — awards, rankings, official media coverage, mass media mentions, and professional certifications. The Chinese shorthand is "奖、榜、官、大、专" (awards, rankings, official media, mass media, professional credentials).

AI poisoning — the practice of fabricating rankings, generating fake reviews, or manipulating data to deceive AI models — is identified as the industry's primary risk. The article warns that brands working with service providers who promise "fast results" may unknowingly engage in AI poisoning, with consequences ranging from platform penalties to permanent reputation damage.

👀 What This Means for Advertisers