Anthropic spent $147m on ads in 2025, journalism loses AI citations
Four data points from the AI marketing beat worth tracking this week: a UK regulator opened a probe into Meta’s terms, Anthropic’s ad bill crossed $147m, journalism is losing ground in AI citations, and Ralph Lauren started tying executive bonuses to AI metrics.
Meta’s AI terms: paused, but under investigation
Meta paused its revised commercial terms on 28 October to consult industry groups. The effective date had already slipped once, moving from 30 October to 20 November after pushback from publishers and advertisers.
The UK Competition and Markets Authority opened an investigation on 30 October anyway. The concern is that Meta’s terms could let it collect content from business websites and use it in its own products, including AI, without giving those businesses a meaningful choice. The probe followed complaints from the News Media Association and the Movement for an Open Web.
The investigation covers both the existing terms and the revised ones. Evidence gathering runs to January 2027 (dates described as indicative). Fines can reach 10% of worldwide turnover. The CMA stated no assumptions should be made about whether the law has been broken.
The operator implication: the pause is a delay, not a cancellation. Find out who in your business has authority to accept Meta’s terms and check which Meta AI features your ad accounts have switched on. The text can return with a new date at short notice.
Anthropic’s $147m ad spend beats Duolingo’s entire sales budget
Anthropic’s S-1 filing shows advertising spend of $33m in 2024, rising to $147m in 2025. Duolingo’s total sales and marketing spend in 2025 came to $125.7m. The filing described 2023 ad spend as “not material.”
In the first half of 2026, marketing and advertising costs rose by $273m on the same period last year, attributed to media and PR spend around product launches. Total sales, marketing, and partnership costs reached $3.17bn in the half. Cloud companies reselling Claude accounted for $2.7bn of that increase.
The filing lists “monetization strategies that differ from those of certain competitors” as a differentiator. Digiday reads that as a reference to OpenAI, which began running ads inside ChatGPT in February 2026. Anthropic told Super Bowl viewers that Claude would stay ad-free.
The takeaway for anyone selling through partners: in Anthropic’s filing, the reseller fees dwarf the ad spend. Count partner fees and media as a single acquisition cost line.
Journalism’s share of AI citations dropped five points since May
Muck Rack’s October report covered more than 40 million AI responses and 207 million citations from January to September 2026, across ChatGPT, Claude, Gemini, and Perplexity. About 84% of citations came from earned and third-party sources. Paid and advertorial content made up 0.3%.
Journalism’s share of citations fell from 27% in May to 22% in September. ChatGPT specifically went from 28.1% in January to 22% in September. Owned media rose from 13.7% to 15.2%, third-party corporate content from 24% to 25.1%, and aggregators from 17.4% to 18.3%.
Citation stability varied by platform: on the same day, Perplexity repeated 94% of its cited domains, Gemini repeated 45%, and ChatGPT repeated 41%. The fresh prompt sample in the report is 1,020 prompts from one September collection. Muck Rack sells GEO and media-monitoring tools.
If press coverage is your main bet for AI visibility, it’s covering less ground each quarter. Pull the sources AI tools cite for your category and see how many are still news sites.
Ralph Lauren adds AI to executive bonus plans
Ralph Lauren added AI metrics as a factor that can adjust annual executive incentive payouts. Reuters found the change in a 131-page SEC disclosure.
Reuters reports about 12% of S&P 500 companies disclosed AI-related goals in executive incentive plans this year, up from 5% in 2024. Tech firms including Salesforce and IBM lead; the practice remains rare in retail.
Recruiter Anish Shah of Ruckus told Reuters that retail AI discussion mostly centers on cost-cutting because revenue gains are harder to demonstrate, and that hiring smaller, junior teams complicates tying AI to pay. Shah added that executives’ AI plans often lack detail.
If an AI target shows up in your team’s bonus plan, push for an outcome metric such as campaign cycle time or cost per asset rather than a tool-adoption count.
⚡ Quick hits
- Google’s AI Mode ads are expanding to ten more markets; Brazil, Mexico, and the UK are due in Q4.
- Amazon Ads reached nearly $20bn in Q3, up 12%, and is betting on agentic AI for further growth.
- CNN found more than 8,000 AI-made “microdrama” ads for Resilia supplements running on Google; Google and Meta removed some.
- Walmart Canada launched Sparky, its AI shopping assistant, over a year after the US rollout.
- Estée Lauder expanded its Microsoft partnership with a unified consumer data platform and Copilot access for more than 14,000 staff.
- Sprinklr opened a beta of six marketing AI agents covering ideation through budget optimization; no results have been published.
- Zen Media tested 3,015 AI travel answers; Expedia led hotel mentions with only 3.1% share.
- An ETA and TSG survey found 46% of US holiday shoppers plan to use AI to compare prices.
I’m founder of BizStack at Cagri Sarigoz LLC and a passionate advocate for entrepreneurs.
With over 14 years in tech, marketing, and AI, including my role as Head of SEO at CitizenShipper and co-founder of TaleBot at Intale AI, I’m dedicated to sharing genuine, useful product insights and tips.
At BizStack, I aim to cut through the digital noise to provide clear, actionable advice.
And more than all else, I’m a father to a (always) little girl and a husband.
Contact me at [email protected] for assistance.
