Fengxing Online CEO Yi Zhengchao: Full Coding First, Then All-in Co-creation | AIGC2026

<img id="wx_img" src="https://www.qbitai.com/wp-content/uploads/imgs/qbitai-logo-1.png" width="400" height="400">

2026-05-21 17:27:25 Source: QbitAI

AI amplifies self-indulgence; the cure is delivering results

Edited by the editorial department, sourced from AIGC2026

QbitAI | Official Account: QbitAI

A video company spent three years teaching all employees to code – including non-technical roles like operations, marketing, and content teams.

This is the counterintuitive stance from Fengxing Online CEO Yi Zhengchao:

What your core business is doesn't matter. In the AI era, the true public infrastructure is coding.

At the 2026 China AIGC Industry Summit, he labeled his company as the "second place player" – neither a trailblazer burning cash nor a laggard waiting观望, but a pragmatic practitioner that quickly follows proven directions cost-effectively.

After three years of AI transformation, Fengxing's profits grew over 10x. But Yi attributes this less to AI video (which only penetrated business in the last year) than to a overlooked factor: pushing AI programming capabilities from engineers to frontline non-technical employees closest to business operations.

He also shared sharp insights:

• Why most companies shouldn't build their own models
• Why individual efficiency gains can harm organizations
• Why even powerful Agents aren't the core lever

And a repeated warning: AI's biggest side effect is amplifying self-indulgence.

The following transcript preserves Yi's core ideas, edited for clarity. The 2026 AIGC Summit hosted 20+ industry leaders, drawing over 1,000 attendees and 4 million live viewers, with extensive media coverage.

Core Ideas

• "Believe in AI, avoid model development" – profitability exists in the middle layer.
• Regardless of your "AI+X" focus, the foundational infrastructure is always coding.
• Shift from role-based to task-based organization – overpowered individuals with misaligned systems create disasters.
• Companies shouldn't hoard "super employees", but become co-creation levers amplifying collective intelligence.
• Boundaries between creation and consumption dissolve – guard against AI-amplified self-indulgence. The cure? Deliver results.

Full transcript of Yi Zhengchao's speech:

Mainstream Culture & AI's "Practical Imagination"

Good morning everyone! Earlier, Mr. Fang categorized companies into first, second, or third place.

I'm the second speaker today, and Fengxing positions itself as the second place player – not a pioneer or laggard, but a cost-effective pragmatic practitioner.

My topic is "AI Programming and AI Video", with a core keyword: co-creation.

As a video company, why start with programming? We've invested 3+ years in AI programming transformation, far longer than our AI video efforts (only penetrating business in the last year). This contrast reflects key thinking for many enterprises.

Before discussing coding, let me summarize the Mainstream Culture & AI relationship with my term "practical imagination".

First, it's imaginative with huge growth potential. Second, it delivers tangible business growth.

I asked AI to generate four images: 1) rapid AI video evolution with demand and tech leaps; 2) time-axis focused opportunities; 3) combined concepts; 4) global visibility for Chinese audiovisual content.

These are consensus views, but today's focus is my personal insights from AI video practice.

AI Video's Paradigm Shift & Co-Creation Imperative

AI video transforms the industry in five ways:

1. Lowering video barriers

Short drama costs dropped from RMB 1 million (live-action) to RMB 100,000 (premium AI dramas), with some even under RMB 10,000. This opens opportunities for many.

1. Massive supply challenges

When platforms list over 40,000 new works/month, standing out requires higher creator standards.

1. Extended value chains

Short videos evolve from novels/IPs to interactive content/games, creating long chains.

1. Operational innovation

Businesses must use AI for operations – not just creation – to realize full potential.

1. Imagination filtering

Co-creation is inevitable. While creation/consumption boundaries blur (creation itself becomes consumption), over-immersion risks arise.

5 Strategic Choices for Enterprise AI Transformation

As a pragmatic AI application company, we've made five key choices over three years:

First, profit growth exceeding 10x through products like ChengXing Dream Factory, ChengXing Push, Xiao JianDao, and Fun Series. While these aren't mass-market hits, they form our AI foundation.

We partnered with MaaS providers, built a 500,000+ creator ecosystem, and saw Zhaohui Group's valuation hit RMB 60 billion. These results came from five strategic choices:

1. Believe in AI, avoid model development

Model development is too resource-intensive for most companies. We position ourselves as MaaS service connectors between providers and cultural industries.

1. Beyond AI animation

Though popular, AI animation accounts for <10% of our business. AI-video applications span short dramas, commercial/entertainment videos, and more.

1. Success through AI programming

As CEO, I consume over 1 billion tokens daily writing demos. Our non-engineer training focuses on business-facing roles, delivering higher efficiency through tools like Codex.

1. Organizational capability over individual talent

Shift from role-based to task-based organization. We categorize employees as:

• "Old pros"
• AI-native youth
• Super connectors

Individuals need to improve intent expression, AI utilization, and cross-domain innovation. Organizations must align systems – misalignment creates inefficiency even with strong individuals.

1. Co-creation over Agents

Our platform provides 80% token subsidies and >50% revenue sharing, combining incentives/tools/Agents to amplify collective intelligence.

We combine employees + digital workers + partners into a social ecosystem. But beware: AI amplifies execution and self-indulgence. The antidote? Deliver results.

From "ChengXing Series" to "FunX Series" – Building an Agent Ecosystem

Our product evolution:

• Fengxing Series: Legacy apps/accounts
• ChengXing Series: 100,000+ creators producing 3,000 episodes/day
• FunX Series (2026):
• FunToken
• FunOPC
• FunStudio (natural language AI app development)
• FunClaw (enterprise-grade content ops solution)

FunStudio enables AI audiovisual app creation through Agents, while FunClaw handles infrastructure for creators. These embody AI's opportunities – turning hot trends into operational reality.

One crucial aspect here is task-oriented thinking, as the implementation of AI in enterprises requires focusing on task execution. We need to ensure that human-machine collaboration is goal-oriented, followed by empowerment. On this basis, feedback and acceptance processes become critical infrastructure for successful AI adoption.

The ultimate goal is to safeguard OPC's agility, enabling OPC and its co-creation partners to focus on productive work and avoid ineffective efforts or aimless activities.

In the future, we will continue to amplify our practices as a video company in AI transformation, fulfilling our corporate mission: Enable the free flow of creativity and make AI accessible to all.

Finally, thank you all! Special thanks to Quantum Bit!

_Copyright reserved. Unauthorized reproduction or use in any form is prohibited. Legal action will be taken against violations._