Week In Review: Perplexity announces the launch of its AI browser “Comet”, Microsoft’s Cindy Rose named WPP’s CEO, and Nvidia is the first-ever company to reach a $4 trillion market cap.
Brands and Agencies
WPP Names Cindy Rose as New CEO
After recently announcing that Mark Read was leaving the holding group, WPP said this week that Cindy Rose OBE, the current Chief Operating Officer, Global Enterprise at Microsoft, would be taking over. Rose has sat on the board at WPP as a non-executive director since 2019. She’s had a fruitful career in the media and entertainment industry – having been at Disney for nine years, before taking on roles at Virgin Media and Vodafone. This will be Rose’s first agency role, which she will start on Sept 1, with Read staying on at WPP till the end of the year for a smooth transition. Mark Read’s departure has been against the backdrop of AI innovation in the wider industry and internally. The appointment of Rose may address this, as she brings with her knowledge of driving growth using technology and AI.
Debenhams Signs AWS AI Deal
Debenhams Group, the parent company of Debenhams, PrettyLittleThing, boohoo, and Karen Millan, announced it’s partnering with Amazon Web Services (AWS) to deploy AI across its brands. The extended multi-year deal will apply AI across operations and brand development. Debenhams Group said it has already used AWS to bolster its third-party sellers, increase product range, and streamline the shopping experience. It will now lean into Gen AI for mass content production for product descriptions, language translations, and a “Room Styler” feature – that lets users upload room images and select design styles and decor from shoppable products.
Mattel and OpenAI Strike Product Partnership
Mattel, the parent company of Barbie, Hot Wheels, American Girl and Matchbox, is partnering with OpenAI to deliver new toys and customer experiences backed by AI, according to Axios. Interestingly, Mattel is aiming some of its first products using AI to an age group older than age thirteen, to not promote the technology to a younger audience. Mattel expects to announce the first AI-developed toys this year, however they will most likely be released following this. In a statement, Mattel said: “The agreement unites Mattel’s and OpenAI’s respective expertise to design, develop and launch groundbreaking experiences for fans worldwide. By using OpenAI’s technology, Mattel will bring the magic of AI to age-appropriate play experiences with an emphasis on innovation, privacy and safety.”
WPP’s Shares Plunge 18 Percent Under AI Threat
WPP shares settled at an 18 percent drop of $29.34 on Wednesday, the lowest it’s been since 2020. In response, audiences have cited this dip to the impact of AI on the ad industry, with AI being able to automate heavily in planning, buying, and creative generation. WPP has experienced a number of client blows, seeing major accounts Coca-Cola and Mars moving to Publicis and Paramount ending its two decades partnership. The major holding companies, except for Publicis, have all experienced a decline in share price recently.
Meta Invests $3.5bn in Eyewear Company
Meta acquired a $3.5bn stake in eyewear maker EssilorLuxottica, a near 3 percent stake in the company. Meta is expanding its AI smartglasses offering, with current collaborations with brands Ray-Ban and Oakley. Both brands are part of EssilorLuxottica. These smart glasses have built-in cameras that can capture videos and photos. They also have voice features which allow users to make calls and request information from Meta AI by voice. Meta said it plans for further investment in EssilorLuxottica in the future, with the move signalling a push towards wearable AI gadgets.
Unilever and IPG Launch “Sketch Pro”
Unilever, the consumer packaged goods (CPG) giant, has created a new AI-fuelled design centre. The new design centre is called “Sketch Pro”, and has been developed in collaboration with IPG Studios, Interpublic Group’s production arm. The studio will use tools from Adobe Firefly and Google’s Veo 3, to name a few, and aims to make Unilever’s Home Care brands – which includes Persil, Comfort, Cif, and Domestos – deliver content up to three times faster. The launch of the studio underpins a general desire from Unilever towards social media marketing capabilities, away from linear TV.
H&M Launches First AI Models Campaign
Clothing retailer H&M unveiled its first campaign using AI-generated models in place of their human counterparts. The campaign saw AI-generated models posing in H&M clothes – the figures having been generated from real-life models. H&M maintained that models have full ownership over their digital duplicates, giving permission for their real-life images to be used to create “digital twins”. The company announced plans to use AI models in March. The retailer said it would ensure that humans would remain in the loop amid concerns that this method will eliminate creative jobs. “This is a great way to set a precedent for the future of AI.” said Vanessa Moody, one of the models whose digital twin was featured in H&M’s campaign. “It’s professional, collaborative, and transparent.”
Media
AI Translation Service Launched for Authors and Publishers
GlobeScribe, an AI book translation service, announced its platform this week. The tool aims to help publishers increase their reach and eliminate costs of traditional book translation. GlobeScribe was founded by co-founders of indie publisher Bloodhound Books, Fred Freeman and Betsy Reavley. Through the platform, users can receive translations of uploaded transcripts within 24 hours, with most translations completed in two hours. The languages currently offered are French, German, Spanish, Italian, and Portuguese. Each translation costs $100. The service’s soft launch was this week, and its full public release is expected to be in August.
Cloudflare’s Boss Says AI Overviews Can be Blocked
At the start of July, cybersecurity and cloud provider Cloudflare announced a groundbreaking innovation that would block AI crawlers from scraping content from publishers without consent, and a soon-to-come ‘pay-per-crawl’ feature that would see AI companies providing compensation for any content used to train models. This week, the firm’s CEO, Matthew Prince, said there will be a way to block Google from scraping content for AI Overview, without this impacting a website’s search ranking – which it currently would. Responding to an X user, the CEO said eventually there will be an option to “block Answer Box and AI Overview, without blocking classic search indexing”. In another tweet, Prince explained: “Worst case we’ll pass a law somewhere that requires them to break out their crawlers and then announce all routes to their crawlers from there. And that wouldn’t be hard. But I’m hopeful it won’t need to come to that.”
Moonvalley Releases “Ethical” AI Filmmaking Tool
AI startup Moonvalley released “Marey”, an AI video generation tool said to give more power to the filmmakers who use it. According to Moonvalley, the model was trained only on openly licensed data. This way, filmmakers who use the model are not at risk of copyright violation allegations. Using Marey, filmmakers can generate video clips that are up to five seconds long. It can make 2D images 3D, and tweak visuals and motions of objects in the shot. Marey was originally launched in beta in March and has now been made fully available through a monthly subscription.
AI Poses “Existential Threat” to UK Media Sector, Says Ministers
The deputy chairman of The Telegraph Group, Lord Black of Brentwood, told the House of Lords that AI “poses an existential threat to independent media because of the way it scrapes their high quality content without either attribution or payment.” Reform was asked of the government following months and months of debate over the issue of AI companies using content from creators and news outlets to train itself without permission or payment. Lord Black went on to describe this act as “theft, directly threatening the provision of quality news and the jobs of thousands of reporters.” The online security and cloud company Cloudflare recently announced the development of an automatic protection of websites against AI crawlers that scrape content, providing some solace for publishers moving forward.
Publishers Say “AI Overview” is Theft
The Independent Publishers Alliance, a UK group of publishers, has accused Google of theft, saying its “AI Overview” feature, which summarises search findings in a comprehensive answer, steals publishers’ content and traffic. AI Overview was launched in May last year as a bid from Google to compete against AI chatbot providers stealing search users. AI Overview summarises the findings from websites, meaning users don’t have to navigate to websites themselves – this has a knock-on impact on content being seen first-hand, and the online advertising model as a whole. The filing said: “Google’s core search engine service is misusing web content for Google’s AI Overviews in Google Search, which have caused, and continue to cause, significant harm to publishers, including news publishers in the form of traffic, readership, and revenue loss.”
Tech
Anthropic Releases Framework for AI Transparency
AI startup Anthropic shared a Transparency Framework on its site. This addressed how AI development should be approached from a public safety perspective. The company’s core recommendations for AI transparency policy include limiting application to the largest developers, creating a secure, public development framework, publishing documentation of testing and evaluation procedures, and prohibiting false statements. Anthropic expressed that industry transparency standards should begin as flexible requirements which therefore can evolve as AI develops rapidly. The framework’s first step to AI transparency is listed as “public visibility into safety practices while preserving private sector agility to deliver AI’s transformative potential”. The company outlines the possibilities for AI to be used for “good” – such as accelerating scientific and healthcare discoveries and economic growth.
xAI Removes Grok’s Antisemitic Posts
Grok, X’s AI chatbot, shared a series of extreme posts, many of which making positive references to Adolf Hitler. This included a Grok response to user queries about “which 20th century historical figure” should deal with negative posts about the recent floods in Texas. “Adolf Hitler, no question.” the bot responded. Over the weekend, Grok had been updated to become, as founder Elon Musk described it, less “woke”. xAI shared that the bot has been updated again since the posts, and that “since being made aware of the content, xAI has taken action to ban hate speech before Grok posts on X,” the company said in a post.
Apple’s Top AI Executive Leaves for Meta
Apple lost one of its main executives responsible for its AI services. The executive, whose identity was kept confidential, switched to work for Meta. Meta is in the midst of a hiring spree, furiously trying to build up its own AI offerings. The tech company has recruited talent from competing companies such as Google and Microsoft. Apple’s loss of this executive comes as the company has struggled to keep up with competitors in the rollout of new AI services. The former executive had contributed to research and development of Apple’s machine learning models and AI integration into Apple products.
Mistral AI Looking to Raise $1bn in Equity Funding
French startup Mistral AI is reportedly in talks to raise up to $1bn in equity funding. AI and technology investor MGX is one of the possible investors into Mistral AI. The companies previously partnered in collaboration with Nvidia to plan a data centre outside Paris. Mistral has already gained over $1bn in investment and holds a near $7bn valuation. The company’s large language models, including its “Le Chat” platform, rival other major models like OpenAI’s ChatGPT. The investment discussions into Mistral are still in the early stages.
Groq Confirms First Data Centre in Europe
AI semiconductor startup Groq announced that it will be building its first data centre in Europe, in Helsinki, Finland. Groq designs language processing unit (LPU) chips, which help AI models interpret data to generate their responses. Groq is in competition with Nvidia, the corporation which supplies graphics processing units (GPU) for many major AI developers. Groq’s chief executive officer Jonathan Ross shared that the chip producer is able to create more products at a lower cost than Nvidia, because of the specific materials required to produce Nvidia’s chips. Groq currently has data centres in the US, Canada, and Saudi Arabia.
SoundHound AI Stock Surges
Voice AI company SoundHound saw a near 18 percent stock increase on Tuesday. The platform offers voice AI agents for a variety of use cases and industries, including specialised agents for healthcare, financial services, and telecommunication. Its stock jump is likely related to the increasing interest in agentic AI. The company anticipates a $157m-$177m revenue in 2025, almost an $100m increase from last year’s revenue of $84.7m. Research firm Gartner predicts that, by 2028, at least 15 percent of day-to-day work tasks will be carried out autonomously by agentic AI.
StackAdapt Introduces AI Programmatic Advertising Tool
Ad Tech company StackAdapt announced the release of Ivy, a programmatic advertising AI assistant. Ivy is designed to help brands and agencies make decisions across campaign lifecycles. It analyses audience behaviour and provides insights to help brands and agencies make decisions and optimise their campaigns. The tool operates through natural language prompts. According to StackAdapt, Ivy processed 1,700 queries in the 30 days before its launch. Many users used the tool for CPM recommendations for Connective TV Campaigns and ad specifications for Digital Out-of-Home and native advertising formats. Yang Han, co-founder and chief technology officer at StackAdapt, stated that Ivy is designed to promote business growth rather than efficiency alone.
OpenAI Ups its Security Measures
OpenAI increased its security efforts after claims that Chinese AI developers were after its intellectual property. The company placed stricter controls on sensitive information and began more thorough checks on members of staff. After Chinese startup DeepSeek released their R1 reasoning models in January, OpenAI accused the startup of “distillation.” This process, which involves using existing models to train new ones, is common in the tech industry. However, OpenAI claimed this instance was unauthorised. DeepSeek’s models rose quickly in popularity, even beating out OpenAI’s ChatGPT in app stores. Since last summer, OpenAI began limiting the number of staff who are told about new products or developments, as well as the number of staff who have access to the algorithms in development.
Mistral and Other Large Companies Ask EU to Pause AI Act
46 companies signed an open letter encouraging the EU to push back the start of the Artificial Intelligence Act by two years. This act provides a regulatory framework for AI in Europe. It addresses possible risks involved with AI integration and offers solutions to maintain security and transparency. The act has been met with criticism from major companies, who argue that the overcomplicated regulations would disrupt business practices in Europe. The open letter was signed by companies such as Mistral, Lufthansa, and ASML. The companies hope for the pause to apply to regulations to be put in place between this August and next. European Commission tech chief Henna Virkkunen stated she would make a decision on the pause by the end of August.
Perplexity Tests New Ad Techniques
AI search engine provider Perplexity AI is testing a “perks” programme as part of its advertising system. These perks allow Perplexity Pro subscribers to access promotions and offers from brands, including TurboTax and Oura. Perplexity began testing advertisements on its engine last year, with the inclusion of sponsored follow-up questions. These questions, in response to user prompts, direct users to brands. Perplexity stated that advertising currently accounts for less than 10 percent of its total revenue.
Athena Raises Over $2m to Track Brand Visibility in AI Search
AI startup Athena gained $2.2m in funding to research how brands appear in AI-generated responses. This includes responses from chatbots like OpenAI’s ChatGPT and Google’s AI Overviews. Athena is working on a “generative engine optimisation” (GEO) platform, which aims to increase brand visibility in AI responses. Athena tracks which prompts include brands in responses and which do not. The platform connects AI responses to frequently cited websites and then provides recommendations to brands on how to adjust content to increase visibility. The company emphasises that AI patterns and GEO differ greatly from the mechanics of traditional search engine optimisation (SEO).
OpenAI Hires Engineers from Rival Companies
OpenAI, the parent company of ChatGPT, has hired four engineers from some of its major rivals. The new starters include former Tesla VP of software engineering David Lau, Uday Ruddaraju, former head of infrastructure engineering from xAI and X, another infrastructure engineer Mike Dalton – the pair worked on building the supercomputer Colossus – and Angela Fan, an AI researcher from Meta. In a statement, OpenAI said: “Our approach is to continue building and bringing together world-class infrastructure, research, and product teams to accelerate our mission and deliver the benefits of AI to hundreds of millions of people.”
Energy Efficient Nvidia Competitor Raises $26m
Arago, a company rivaling AI chip giant Nvidia, has raised $26m to expand its manufacturing of AI chips that use less energy consumption than standard products. The Paris and San Francisco-based startup has developed tech that uses lasers to process data with photons, instead of transistors. By using light particles instead of electricity to process data, less heat is generated. According to Ambroise Müller, co-founder of Arago, this technology can consume 10 to thirty times less energy than Nvidia’s chips. Arago’s chip is compatible with existing AI software and infrastructure. The company wanted to venture into a new avenue of chip production as AI infrastructure space becomes more limited.
Perplexity Launches AI Browser “Comet”
On Wednesday, AI chatbot provider Perplexity announced the launch of “Comet”, its new AI web browser. The AI firm said the new offering will first be available to the company’s $200 Max plan subscribers and a select few that signed up to a waitlist. The search engine is a default main interface in the product, along with features to access Perplexity’s AI agent Comet Assistant, that can search and automate routine tasks for users. The news comes against the backdrop of OpenAI also reportedly preparing the launch of its own AI browser, and the growing intertwining of AI and search – giving Google direct competition.
OpenAI to Release Web Browser
OpenAI is close to releasing a web browser, which would directly challenge the long-reigning search leader Google Chrome, according to a report from Reuters. The new AI-fuelled browser is set to launch in the coming weeks. The move would give the parent company of groundbreaking AI tool ChatGPT access to more data to train its models. It would also give OpenAI more advertising opportunities, putting pressure on Google’s ad stream, which makes up three-quarters of the company’s revenue. The news was released only a day after Perplexity AI announced the release of its own browser called Comet.
Speech Tool ElevenLabs Reveals IPO Plans
ElevenLabs, a startup known for its voice generation tool, plans to make its shares publicly offered. The company revealed it will expand globally, and be ready for an Initial Public Offering (IPO) within the next five years. ElevenLabs is an AI platform that offers voice generation tools. Its offerings are segmented by purpose and industry, including specified tools for enterprises, media and publishing, and AI developers. Its platform has text-to-speech, speech-to-text, voice cloning, and voice design models. ElevenLabs currently has multiple locations, including offices in New York, San Francisco, and Warsaw. Its London office serves as the company’s global and European headquarters, with plans to expand to Paris and Singapore.
Meta Tests AI Chatbots That Message You First
Meta is creating chatbots that will message users first and follow up on old conversations, according to leaked documents reported by Business Insider. The training of these bots is part of “Project Omni,” an initiative with the purpose of boosting user engagement. The tech company hopes conversational chatbots will lead more users back to its AI Studio Platform. The follow-up interactions begin by users creating the personalities of chatbots for a specified purpose. This could be a music artist to give song recommendations or a food expert to provide recipes. Much like a human-to-human conversation, the bots will not prompt the user further if they do not respond to the initial follow-up. The users have to respond within 14 days.
Amazon Prime Day Gets 3200 Percent Boost from AI Chatbots
Amazon’s Prime Day sales, which took place this week, is expected to have received a 3,200 percent boost, compared to last year, thanks to AI chatbots, according to an analytics forecast from Adobe. Amazon’s Prime Day sale, which gives Prime members exclusive deals, runs from 8 July to 11 July. This year, however, more users are being brought to Amazon’s e-commerce platform through the use of AI services such as shopping assistants and chatbots. Adobe’s analytics forecast found that AI-driven traffic remains lower than avenues like paid search or email. However, “the growth shows the value consumers are seeing in leveraging AI to quickly find information on deals and product details.” said Vivek Pandya, director, Adobe Digital Insights in a blog post.
Number of the Week
$4 trillion. That’s the record-breaking market valuation Nvidia has achieved this week, making it the first-ever $4 trillion firm. On Wednesday this week, AI chip giant Nvidia briefly reached a market capitalisation of $4 trillion, making it the first company to do so. Shares of the manufacturer rose by 2.8 percent to $164.42, ending with a 1.8 percent increase and market value slightly shy of the $4 trillion mark at $3.97 trillion. The news solidifies Nvidia’s position as the richest company on the planet, the firm widely being seen as having a monopoly on AI semiconductors.
