Ad Tech Briefing: AppLovin and Unity’s dual explained 

On Sept. 29, it emerged that AppLovin has taken Unity to court while alleging that Unity’s Ad Quality SDK improperly collected information generated when AppLovin won and served ads. Unity refutes the allegations and is characterizing the litigation as an attempt by a dominant incumbent to impede a growing competitor.  

The underlying filings make the technical dispute particularly interesting, with AppLovin contesting that the disputed data leakage was derived from ad auctions Unity did not win — and, in some cases, potentially auctions in which it did not even participate. 

The outcome of the legal dispute will likely have a ripple effect. A bidder ordinarily receives the information needed to evaluate an impression and then learns whether its bid won or lost. AppLovin argues that doesn’t give a competitor the right to assemble a record of the creative another bidder served, the user who received it, the revenue associated with the impression and the subsequent engagement. 

AppLovin further claims Unity’s Ad Quality software uses AppLovin-specific connector scripts, interacts with callbacks inside AppLovin’s SDK, and subscribes to an internal channel used for impression and revenue events. The filing also alleges that Unity can configure that collection remotely on an app-by-app basis. 

In the filing, AppLovin further alleges, “on information and belief,” that the resulting information was useful for training Unity’s advertising models and modeling AppLovin’s own ad decisions — an allegation Unity disputes and which has not been adjudicated.  

For its part, Unity opposes all the allegations, and in its subsequent filing, it argued Ad Quality obtains information from publishers’ apps or users’ devices with publisher permission, and not “from AppLovin.” 

Unity also argued that AppLovin operates its own ad-review technology and characterized AppLovin’s Ad Review product as “far more intrusive.” AppLovin disputes that comparison, saying its product is restricted to MAX-mediated impressions and isn’t used to train its advertising models. 

Furthermore, according to the filings, Unity said it could remotely stop Ad Quality from collecting data relating to MAX-mediated auctions within five business days. But AppLovin says Unity made that offer contingent on resolving the dispute. AppLovin also argues that such a change wouldn’t address previously collected information

Modern mobile monetization requires multiple rival SDKs, mediation platforms and demand sources to operate inside the same app. The AppLovin–Unity fight is effectively asking how much information one participant can legitimately extract from that shared environment — particularly when the same companies increasingly use transaction data to improve machine-learning systems that subsequently compete against one another.

The eventual answer could matter well beyond these two companies. AppLovin’s filing itself says Unity’s Ad Quality technology supports data collection involving numerous other ad networks, such as Google, Liftoff, Digital Turbine and InMobi. That remains AppLovin’s characterization of the technology, not a judicial finding.

For an industry rapidly attaching AI to auction optimization, the dispute therefore raises an increasingly consequential question: which signals are legitimately observable when rivals participate in the same transaction?

What we heard

“We are figuring out how to give them more transparency on the audience side.”

— Amit Bhattacharyya, vp of full-funnel products, AI & modeling at Amazon Ads, speaking about the controls available within Full-Funnel Campaigns. 

Numbers to know

3,300+: Community members AgenticAdvertising.org says now participate in its ecosystem as it attempts to establish common infrastructure for agentic advertising.

67%: Higher return on long-term ad spend Amazon says advertisers in the Full-Funnel Campaigns beta generated compared with manually configured full-funnel campaigns.

29%: Lower cost per new-to-brand customer Amazon says the same beta campaigns produced.

20%: Approximate share of eventual sales value Amazon says can appear beyond the immediate attribution window, according to its work developing its Long-Term Sales metric.

What we’ve covered

At its flagship unBoxed conference, Amazon Ads collapses buying silos as AI takes a bigger role in media decisions

The company has renamed its unified platform Amazon Ads Agent, with further product launches.  

OpenAI moves to make ChatGPT Ads more measurable — and more visual

Since OpenAI started selling ads in ChatGPT in February, advertisers have kept coming back to the same complaints: they can’t measure the ads properly, they don’t have enough formats to work with and there isn’t enough inventory to absorb what they want to spend. OpenAI’s latest round of updates takes aim at two of them.

What we’re reading

⁠Amazon locks out Meta’s Muse in agentic shopping standoff

Amazon blocked Meta’s Muse agent from shopping on its website, another emerging example of platforms deciding which outside agents can interact with their infrastructure.

⁠Citi is jumping into the ad business and entering a fast-growing and competitive market

Citi is joining the financial-media-network rush, using purchasing signals from more than 70 million customers and billions of annual transactions.

⁠Adtech vet Brian O’Kelley’s startup is pivoting to AI agents and rebranding with a new name

Scope3 becomes Apostra as O’Kelley makes another attempt to rewire how digital advertising is traded.

⁠Apple’s latest operating system blocks The Trade Desk from serving ads on Safari

The technical dispute is one to watch closely for signs of whether Apple changes course.