NEW AI VIDEO MODEL
OpenAI’s new video app –and video ads
AI video ‘slop’ has a new competitor: compelling AI video content.
Yesterday, OpenAI announced the next iteration of its AI video model, Sora 2, and introduced a video app that also includes audio generation and sound effects as well as multiple languages. The “Cameo” feature is what makes the app particularly compelling according to the OpenAI developers who presented on the livestream. The creator can have their image ingested by Sora via a phone’s camera and put permissions on it.
On the Sora 2 product launch page, the Sora ‘team’ concluded:
“Video models are getting very good, very quickly. General-purpose world simulators and robotic agents will fundamentally reshape society and accelerate the arc of human progress. Sora 2 represents significant progress towards that goal. In keeping with OpenAI’s mission, it is important that humanity benefits from these models as they are developed. We think Sora is going to bring a lot of joy, creativity and connection to the world.”
- Read: Sora 2 is here – OpenAI (September 30)
- Download: Sora 2 iOS app (iOS-only for now. Availability is on the Sora 2 launch page.)
- Livestream recording: YouTube
- Hype video: via OpenAI’s Fidji Simo on X
- Coverage: OpenAI debuts Sora 2 AI video generator app with sound and self-insertion cameos, API coming soon (September 30) – VentureBeat
The OpenAI video ads example
What will not be lost on ad industry participants is the Sora 2 video ad example beginning at 11:31 of the recorded livestream.
OpenAI engineer Rohan Sahai explained:
“One of my favorite features of this app, and this model, and I think something that is uniquely made possible with this technology, is the ability to immediately participate in a trend, a storyline, in some lore that a creator is working on (…) via this remix feature. (…) I can click the remix button and say, ‘Make this an ad for a top hat with giant feathers…’ And boom I fire off that generation…”
From tipsheet #1: Imagine the data that will pour out of the video ads made using OpenAI’s Sora model… the OpenAI LLM will continued to ingest that data along with data from the Instant Checkout purchase path and build an AI-powered, ad targeting capability to supercharge its eventual ad product.
From tipsheet #2: Sora 2 isn’t just about video, it’s about creating a virtual reality. Virtual reality is here. Meta’s pre-emptive launch of Vibes last week makes sense (They knew Sora 2 was coming.). Who’s gonna own VR? “We are,” says Meta. “We are,” says OpenAI.
Content creators will likely run wild with OpenAI’s Sora app and Meta’s Vibes (via Instagram?). AI slop is going to be transformed into compelling content generated by AI. TikTok is gonna need some Vibes or Sora ‘sauce.’
Across the ad tech industry, do new models such as these change the game for genAI creative tools? OpenAI’s video API will be popular.
Open AI is going in all AI directions.. it’s not just text and a chatbot. It’s re-thinking our media lives.
ECOMMERCE & CHATGPT
Reaction: ChatGPT’s Instant Checkout & ads
Reviewing the launch of Instant Checkout for OpenAI’s ChatGPT, Stratechery’s Ben Thompson concluded by looking at answer engine versus Meta mechanics:
“…ChatGPT is basically saying that if you want to be the link that is purchased, you need to use Instant Checkout, and need to give them a cut. That this is potentially possible is a function of ChatGPT helping customers find products, as opposed to Meta’s goal of helping merchants find customers.
This has always been how I’ve thought a ChatGPT ad or affiliate model would work: if there are multiple merchants for a recommendation, then the merchant who pays ChatGPT the most would be the one who is featured — that’s exactly what is happening here. What is interesting is that this approach does mean that ChatGPT actually isn’t directly monetizing its most valuable function, which is uncovering the unique product you can’t find anywhere else; that, though, is par for the course for consumer tech companies: the idea is to generate so much consumer surplus that taking a skim off of only a fraction of the attention you generate pays the bills (and a whole lot more!).
Read: “OpenAI Instant Checkout, AI and Long Tail E-Commerce, Is AI Different?” on Mr. Thompson’s Stratechery. (September 30 – subscription)
Analyst Eric Seufert cut right to the chase on X:
“ChatGPT’s Instant Checkout is an advertising data feed.
Instant Checkout, applied to ChatGPT’s 700 million weekly active users, provides OpenAI with the sine qua non of a scaled advertising platform: user-level conversion data. This data, wielded by the recently acquired team from Statsig, can be used to target ads to ChatGPT users on the basis of confirmed transactions routed to Etsy, Shopify, and whoever else onboards to the platform.”
Read: “ChatGPT’s Instant Checkout is an advertising data feed” on Mr. Seufert’s Mobile Dev Memo. (September 30 – subscription)
More: “The OpenAI Take Rate” by TBPN’s John Coogan, Jordi Hays and Brandon Gorrell
-
“If OpenAI captures 20% of the e-comm market in the next 5 to 10 years, that’s around $300B of activity — and a 20% take rate would mean $60B in revenue. This is all back-of-the-envelope, but I think it’s a reasonable framework for thinking about the market.” – TBPN on Substack (September 30)
THE TRADE DESK AND AI
The Trade Desk CEO begins war on narratives
The Trade Desk CEO Jeff Green took to LinkedIn yesterday and said, in so many words, his company was taking back the open web —and its advertising opportunity— courtesy of AI and innovations.
Mr. Green teased “more to come” on LinkedIn (September 30):
“…We’re at a fork in the road. AI is making it easier than ever to shine the light of transparency on everything that stands between the advertiser and the publisher. As an industry we need to grasp that opportunity. In doing so, we will showcase the value of the open internet in contrast to the limitations of walled gardens.
Later this week, I will announce innovations that we’re launching in October to improve the digital ad ecosystem. All in service of our buy-side clients. But indirectly, as always, we want to help raise all boats. Because if we do that, the growth opportunity for everyone is so much greater.
But first, let’s set the stage.”
Then, Mr. Green pointed to his own op-ed on The Trade Desk’s ad trade, The Current: “The open internet 2025, part 1: Framing a year of change.” Read it here.
He concluded “part 1” of his op-ed by setting the stage for the imminent innovation announcements:
“At The Trade Desk, improving the supply chain of the open internet for the most sophisticated buyers in the world is a major focus. Innovation can accelerate the inevitable long-term movement toward a transparent and efficient marketplace for digital ads.
I will elaborate on our latest innovations — always in service of the buy side — in the second part of this op-ed later this week.”
From tipsheet: With the promised “innovation” announcements, CEO Green will be aiming at changing two narratives: the open web is dying and The Trade Desk is struggling. His comments suggest AI innovations will be key. So, too, will be the quarterly earnings reports from the $25 billion public company – the next one is likely in early November. Stay tuned.
Related: AppLovin Stock: Is the AI-Advertising Stock a Buy, Sell, or Hold? (September 30) – Motley Fool
TOOLBAR ADS FOR CHATBOTS
‘Toolbar’ ads for answer engines
PayPal SVP & General Manager Dr. Mark Grether threw his payment company into the agentic commerce ring yesterday with the launch of a new answer engine product from the company’s Honey platform.
Essentially, the PayPal Honey browser extension displays products mentioned in an answer engine’s response. Honey derives affiliate revenue from the sale of those products.
Dr. Grether wrote on LinkedIn (September 30):
“If you’re paying attention to the headlines, agentic commerce is arriving at scale. The platform that shapes new user habits in AI commerce will win—everyone else will be left following. The challenge is clear: It’s not just about building shopability. It’s about delivering trust and confidence with shoppers’ money.
Today, we’re introducing new capabilities in PayPal Honey that turn answers into sales. When someone asks AI for product recommendations, they should instantly discover what merchants have the item, see what the prices are and know what kind of value they’ll receive…”
More: PayPal Continues to Shape Agentic Commerce with Instant Product Discovery (September 30) – press release
The release said that the new agentic “features will be available at no cost to new and existing PayPal Honey users in the U.S. soon, in time for Black Friday and the holiday shopping season.”
From tipsheet #1: “Download my toolbar!” redux. This is potentially a clever way for anyone with a browser extension to overlay an ad model on top of answer engines. It would take a huge install base to have ad impact/reach. Honey is believed to have had an install base of around 14 million Chrome users in July – good for an affiliate business, but not compelling for an ad business.
From tipsheet #2: Speaking of ads and chatbots, Taboola is implementing ads across answer engines on publisher sites with its DeeperDive product. On an aggregated basis, it remains to be seen if that will scale to a compelling ad opportunity or more appropriate for affiliate marketing/publishing revenue.
Related: Extending E-Commerce Revenue With Browser Extensions (September 30) – Forbes
MORE
- Amazon unveils new generation of AI-powered Kindle and other devices (September 30) – AP
- AppLovin Stock: Is the AI-Advertising Stock a Buy, Sell, or Hold? (September 30) – Motley Fool
- Report: The Modern Marketing Data Stack 2026 – Snowflake (PII required)
- ‘The red line keeps moving’: Advertising creatives have lost the AI ick (September 30) – Digiday
- Google’s Liz Reid: AI isn’t replacing search – it’s augmenting it (September 30) – Search Engine Land
- Zeta Global to Acquire Marigold’s Enterprise Business (September 30) – press release
- Venture Firm Touring Capital Decided to Use AI to Scout Deals. It Wasn’t All Smooth Sailing (September 29) – The Wall Street Journal