AI Statics Are a Commodity

A commodity on a three to six month clock

Danny Yeung runs IM8's creative process on a deliberately temporary advantage. The brand currently leans on AI-assisted static ads because they are fast, cheap, and the most direct way to relieve pressure on creative volume, and Yeung reports the same discovery from his own broader marketing experience: overweighting video, then flipping toward roughly seventy five percent statics, improved performance. But he treats the static advantage as explicitly running out. "You have so many companies now that make the static on the AI, it's going to be so easy. It's going to be like a commodity. So if everyone's going to a commodity, how do you differentiate against everyone else? That's where video and authenticity is going to play a big factor later on," he said, putting a three to six month clock on the shift.¹

The underlying structure generalizes past advertising. A production capability that artificial intelligence drives toward zero marginal cost gets adopted universally, and universal adoption is symmetric by definition, which means it stops distinguishing anyone. The live question for any AI-enabled advantage is not whether it works today but how long until everyone has it, and what stays expensive once they do. Yeung's answer for advertising is real people and real stories: not because AI video is impossible, but because the scarce input is not the rendering, it is having an athlete's coaching team actually wear the product or a partner with final say on taste actually stand behind it.

The complicating case: their best ad was AI

The thesis sits awkwardly next to IM8's single biggest creative win. A twenty five thousand dollar AI-generated ad featuring tennis player Aryna Sabalenka drew two hundred million views, twenty million of them on the first day, ran on a Times Square billboard, and was described as one of the most-viewed social ads of 2025.¹ Two features explain the performance without undercutting the underlying argument. Yeung insisted the ad disclose that it was AI-generated from the start, and the novelty and the honesty together drove comment-section debate that fed distribution. And the ad ran in September 2025, timed to the US Open, which Sabalenka then won and shared herself, a launch he believes triggered the algorithmic amplification. The performance came from novelty, disclosure, timing, and a real athlete's endorsement, not from AI generation as such, each of which depletes or fails to repeat.

Open question

The clearest tension is that the evidence cuts both ways from the same mouth: the most successful creative IM8 has run was AI-generated, cost twenty five thousand dollars, and reached two hundred million people, which a skeptic could read as the opposite lesson. The three to six month timeline is also a guess rather than a measurement, and the deeper risk to the thesis is that AI video is improving fast enough that authenticity stops being visually detectable at all, in which case the real moat is not authenticity but verified provenance, a harder problem than the one Yeung is currently solving for.