AI Gamechangers Summit: Predictions, pushback and the new investment landscape

• Chris Zhou opens by asking the question: can AI actually change how games feel?
• The AI Paradox panel puts triple-A sceptics, F2P giants and investors in the same room to find out where the industry's real "iPhone moment" is coming from.
• Mattel163, BITKRAFT and Aream & Co bring real deployment data and live investment thinking to the table.

AI has become one of the defining conversations in games and China is increasingly central to where it goes next.

When Pocket Gamer Connects Summit Shanghai kicks off its four-track programme on July 29th, that debate will take centre stage as the four-track programme closes with the AI Gamechangers Summit.

The Summit frames the conversation in two directions: what AI could genuinely unlock for games and where the money is going to be made.

Here's what each session has on the agenda, including thoughts from the investors and studios taking part.

Session one: 'Four Wild Speculations On Gameplay Design In The Age Of AI'

The Summit opens with a solo talk from AI practitioner Chris Zhou and it's a deliberate departure from the AI conversation most of the industry has been having.

Zhou's premise is to set aside the commercial KPIs like the art-pipeline savings, the code-cost cuts and the productivity math everyone else is running and ask what AI could actually change about how games feel to play.

His talk is structured around four speculations, each built on one of the primal reasons people play games in the first place, long before monetisation enters the picture: Play, Create, Learn and Companionship.

It's less a talk for studio heads chasing efficiency gains and more a provocation aimed at anyone still asking whether AI can make games better, not just cheaper to make.

Session two: 'The AI Paradox - Creative Resistance vs. Industrial Revolution'

This panel puts triple-A scepticism and F2P appetite in direct conversation. Triple-A purists remain wary of AI's creative value, F2P giants are already rebuilding their LTV models around it, and investors are placing early bets on an "AI-native" future.

Rather than rehashing the hype cycle, the conversation turns to what's actually being built and deployed today and where the industry's own "iPhone moment" might come from.

The line-up: Wenfeng Yang (BBX Ventures), Devin Nambiar (Mattel163), Kartik Prabhakara (Aream & Co), Harri Manninen (Play Ventures) and Jonathan Huang (BITKRAFT Ventures) - bringing triple-A, F2P and investor perspectives to the same table.

Investors on the panel are already watching how that "AI-native" bet plays out differently by market. Kartik Prabhakara, co-founder and partner at Aream & Co, argues that state involvement is a distinct commercial advantage in China, at least in the short term.

"China's government-as-catalyst distribution delivers real advantages, including faster revenue growth, a bias toward open-source models that speeds enterprise adoption, and a cost advantage from state subsidies for power and infrastructure," says Prabhakara.

"Governments treat AI as strategic and direct resources to ensure national champions emerge. It also accelerates application-layer and vertically integrated offerings, embedding AI directly into industry-specific workflows rather than layering them on top."

He adds: "But policy-driven scaling carries structural risk. When the state accelerates adoption ahead of market demand, value tends to accrue to a country's industry position as a whole more than to company shareholders. There is no blanket valuation premium for Chinese AI companies based on adoption speed alone - policy dependency, overcapacity and margin compression are risks markets will eventually price in.

"Investors cross-reference multiples across markets, with Chinese investors benchmarking against US peers and vice versa, meaning pricing gaps rarely persist in isolation. Across both markets, given the size of the opportunity, perceived winners will continue to command premium valuations."

On the studio side, Devin Nambiar of Mattel163 sees the payoff in more concrete terms - not valuation dynamics, but day-to-day monetisation and content output.

"At Mattel163, AI has already moved the needle in several areas," he says.

"On monetisation it has helped us with both prediction and creating a more efficient content pipeline. Our pLTV modelling and ROAS prediction employ AI to help us dynamically adjust our UA plans to maximise paid and blended ROAS across a variety of channels, which is paramount in this highly competitive UA environment."

He continues: "On the content side, using AI to guide portions of our live content creation, work alongside our designers to create new, engaging mini-game content and aid with localisation efforts has helped improve the efficiency of our overall live content pipeline - including limited-time content, seasonal content and innovative mini-games.

"Despite our concerted efforts and AI-related initiatives at a company level, in many ways we still feel we are scratching the surface and we actively want to go deeper."

From the investor side of the table again, Jonathan Huang of Bitkraft Ventures sees that same shift changing what he expects from founders before he'll write a cheque.

"Pre-AI, seed investments could be underwritten without the existence of any product, but today I think the bar has shifted towards seeing a semblance of product-market fit, already at seed," he says.

"Today it is not only effortless to produce prototypes, we also see a ton of 'headfake' week-one retention data given the novelty and innate virality of AI-driven experiences, where the engagement is non-durable and craters shortly after.

"At seed, founders ought to come prepared with working prototypes and demonstrably durable retention metrics, because only then can one believe there is some form of a moat that a team can leverage to compound on."

He adds: "At Series B it's no longer about proving product-market fit but whether this company can scale and potentially be the category winner.

"During the ZIRP era we saw unprecedented 'growth at all costs' strategies which mostly collapsed eventually. Today the challenge is slightly different - specifically for AI-native entertainment experiences, can you scale user play time without inference costs eating all your margins away? Ability to scale economically is ultimately tied to whether you can beat the competition."

ChinaJoy and 'Level Up with AI'

PGC Shanghai serves as the gateway to ChinaJoy, which opens two days later under its own "Level Up with AI" banner.

The new Next Play zone turns AI NPCs and reactive storytelling into playable demos and the show floor is packed with AI infrastructure - rendering, localisation and monetisation tools all built directly into studio pipelines.

Be a part of the AI conversation - PGC Summit Shanghai tickets are available here.