'100% a defensive move': How Google is pushing AI innovation while keeping its core businesses humming; OpenAI’s biggest-ever acquisition and more

Welcome back to LinkedIn News Tech Stack, which brings you news, insights and trends involving the founders, investors and companies on the cutting edge of technology, by Tech Editor Tanya Dua. You can check out our previous editions here and here. Catch up on the latest edition of VC Wednesdays with Neotribe Ventures' Neeraj Hablani.

Google unveiled a slew of new AI updates at its annual developer conference on Tuesday, including an AI filmmaking tool, a reasoning mode for Gemini’s Pro version and AI Mode in Google search in the U.S. — a move CEO Sundar Pichai described as a "total reimagining" of Search.

But the push didn’t end there. On Wednesday, the company said it was expanding ads within AI Overviews atop search results to desktop in the U.S. It also said it was testing ads within AI Mode, the new AI-powered search feature where users can dive deeper into topics using the Gemini AI chatbot, which was just rolled out across the U.S.

Reading between the lines, the tech giant is essentially declaring that while it’s betting on an AI future, it isn’t doing so at the cost of its core advertising and search businesses — and ads are coming right along. The bulk of Google's revenue, after all, is derived from traditional search ads, with the company raking in about $265 billion in ad revenue last year.

“AI is helping us give better answers [...] and that also means better and more helpful commercial answers when it's relevant,” Dan Taylor, VP of Global Ads at Google, told LinkedIn News. “Already, Google and YouTube show up in about two-thirds of purchase journeys, and AI Overviews and AI Mode create even more opportunities for us to introduce brands earlier in the process."

But while Google is positioning the new offerings as a net positive both in terms of innovation and for its partners, experts see it as old wine in a new bottle — saying Google may be trying to reimagine search but the way that search is being monetized isn’t changing.

“This is 100% a defensive monetization move,” said A.J. Ghergich, VP at AI-powered search visibility company Botify. “They’re retrofitting every AI product with advertising, because they can’t afford to lose their cash cow.”

For Garrett Sussman, director of marketing at SEO firm iPullRank, Google’s current position is similar to the innovator’s dilemma that both Blockbuster and BlackBerry faced with the rise of Netflix and the iPhone. But while Google was initially caught flat-footed in the AI race, Sussman gives the company credit for trying to innovate and accelerate its product pipeline unlike BlackBerry or Blockbuster, which were like “an ostrich with its head in the sand."

To his point, the search ecosystem is unraveling at the hands of AI, which is changing how people access information online and gradually shifting behavior away from traditional search engines. Google’s global market share in April fell to 89.65% — the lowest since 2013, when it dropped to 89.41%, according to Statcounter data.

And Google leaders seem to be well aware that their position is threatened, and have pushed toward AI monetization.

“We have 3 options: (1) Search doesn’t erode, (2) we lose Search traffic to Gemini, (3) we lose Search traffic to ChatGPT,” Nick Fox, Google’s SVP of knowledge and information, is revealed to have said in an internal exchange. “(1) is preferred but the worst case is (3) so we should support (2).”

At the same time, monetizing AI products isn’t just Google’s headache, given how expensive it is to actually run LLMs. While OpenAI seems to be going for an affiliate model with ChatGPT recently starting to link out to merchants, Perplexity has dabbled in both affiliate links and advertising, pointed out Sussman.

Venture capitalist Vanessa Larco sees AI-driven search as an opportunity, as it opens the door for new players to chip away at incumbents like Google. But she doesn’t think advertising is the right way forward.

“Advertising in search works because ultimately I’m still in control, but not in AI products because I’m using it to validate decisions,” she said. “How can I trust that Google is not just feeding me the brand that’s paying them the most dollars?”

Google, for its part, says it’s driven by consumer behavior and the desire to make advertising easier by using AI to automate ad creation and purchasing. It’s rolling out a new feature called Smart Bidding Exploration, for example, which will automatically detect new types of conversational queries that advertisers can bid on based on more complex AI searches that go beyond typical keywords. It’s also offering an AI agent within Google Ads and Google Analytics, which can make recommendations on ad campaigns. 

“Instead of someone saying 'expandable suitcase,’ it could be, 'I'm traveling for business and looking for a business travel essentials packing list,'" said Taylor. “We’re delivering higher-quality ads and better results using the conversational experience. Smart Bidding Exploration beta participants saw an 18% increase in search query categories with conversions."

Ultimately advertisers, not Google’s claims, will determine whether the new features deliver incremental value, said Gartner analyst Eric Schmitt. He also questioned how exactly ROI would be measured in an AI-driven conversational world, with Google’s pay-per-click model, where businesses pay Google only when someone clicks on their ad, no longer being the best metric of success.

“The handoff from a blue link isn’t the point of monetization anymore — time spent now becomes the key metric, more like TV,” said Schmitt. “Users paying for either a better-quality answer engine or less ads seems like one possibility.”

In the end, experts see Google perpetuating what has worked in its favor rather than abandoning ads. While newer challengers may experiment with affiliate and direct models, Google has unmatched resources — and still calls the shots.

“We’re in this moment where we can get rid of a dirty system [...] and we’re going to use the same old advertising model?” Ghergich said. “They’re repeating the same model with tomorrow’s technology. That’s what got us here in the first place.”

• OpenAI acquires Jony Ive's startup. The ChatGPT maker is preparing to venture into hardware and has acquired io, the AI device firm led by legendary Apple designer Jony Ive, for $6.4 billion. The all-stock deal will also see Ive step into a creative leadership role at OpenAI, where he'll oversee a family of new consumer devices, including one slated to debut in 2026. The acquisition is OpenAI’s largest to date. After announcing the acquisition, chief executive Sam Altman offered staffers a sneak preview of the hardware he is developing with designer Jony Ive. In a recording reviewed by The Wall Street Journal, Altman said the deal could add $1 trillion in value to OpenAI, noting the company expects to ship 100 million so-called AI companions that promise to be pocket-sized, screen-free and equipped for full awareness of a user's surroundings .Read more about the company's product pipeline and hardware products in the works, here.

• Anthropic rolls out new AI models. Anthropic on Thursday debuted what it's calling "the world’s best coding model." Claude Opus 4 can work autonomously for seven hours, the company says, utilizing multiple software tools at once. That's a crucial step closer to AI agents that can handle complex tasks without human supervision, per Engadget. Still, Amazon-backed Anthropic faces stiff competition from other AI firms, including OpenAI, which just announced a blockbuster $6.4 billion deal to acquire Apple legend Jony Ive's startup.

• The AI agent space is seeing advances. Earlier this week, OpenAI also launched a research preview of its AI coding tool, marking the ChatGPT maker’s push into software engineering. The company says the AI agent, called Codex, can perform multiple tasks simultaneously — including writing code, fixing bugs and answering questions. But OpenAI isn’t the only one. LinkedIn parent Microsoft too is envisioning an AI-agent future and leaning into an "open agentic web," one where AI agents can make decisions and perform tasks for individuals and organizations. Among the announcements at its annual Build developer conference: Windows is "embracing" Model Context Protocol as it lays the groundwork for AI agents; Azure AI Foundry models now include xAI's Grok 3 and Grok 3 mini; and GitHub too unveiled a new AI coding agent that can fix bugs. See the list of announcements by Microsoft CEO Satya Nadella here. 

• Apple opens AI models to outsiders. Apple is planning to open up its AI models to third-party developers, according to Bloomberg, citing anonymous sources. The new software development kit would allow outsiders to build software using Apple Intelligence. The move aims to drive innovation and give Apple a competitive edge in AI, an area in which it has been lagging. A rise in apps may also help boost its profitability amid increased regulatory pressure over its in-app purchase commission fees. More details are expected June 9.

• Meanwhile, Meta pushes back the debut of its AI model. Meta is delaying the release of its most powerful AI model after internal concerns that its capabilities don't match its public claims, The Wall Street Journal reports, citing anonymous sources. Touted as an OpenAI challenger, the advanced Llama model's release has been pushed back to fall or later. Meta is reportedly weighing "management changes" as a result. Some consider the move a sign that AI advancements will continue to progress but may come at a "slower pace," with "tremendous cost."

• Nvidia, Qualcomm branch out. NVIDIA has debuted a new program that will allow its customers to create "semi-custom AI infrastructure," per CEO Jensen Huang. NVLink Fusion, announced at the Computex 2025 conference in Taiwan on Monday, will allow Nvidia customers to use other companies' central and graphics processing units with Nvidia products. Also on Monday, Qualcomm said it will roll out CPUs for the data centers that power artificial intelligence, linking to Nvidia chips. With that move, Qualcomm steps into a market currently dominated by Intel Corporation and AMD. In related news, Nvidia CEO Jensen Huang on Wednesday called the U.S. restrictions on the sale of advanced chips to China a "failure,” arguing the export controls have only encouraged Chinese companies to advance their own AI technologies — and Nvidia has lost market share to those rivals as a result.

Here’s a list of AI deals, funding announcements and other notable developments from this week:

• SAP announced a slew of new partnerships at its SAP Sapphire conference this week, including a partnership with Perplexity as well as Candian LLM maker Cohere. 

• Coding startup Replit has rolled out a bunch of new safety features catered to the enterprise. See this LinkedIn post by its developer relations exec Matt Palmer for more.

• AI notetaking startup Granola has raised $43 million in Series B funding and unveiled a bunch of new product updates. See this LinkedIn post by CEO Christopher Pedregal for more details. 

• AI roleplays startup Yoodli AI Roleplays that uses AI to help people improve their communication has raised $13.7 million in Series A funding. See this LinkedIn post by co-founder and CEO Varun Puri for more details.

• AI startup Greenlite AI , which is building AI agents for the financial services industry, has raised $15 million in Series A funding. See this LinkedIn post by co-founder and CEO Will Lawrence for more.

• A new AI startup called Coworker.ai that claims to have a better way to handle enterprise AI tasks has emerged from stealth and announced $13 million in seed funding. See this LinkedIn post by co-founder and CEO Alex Calder and this post by investor Hadley Harris for more.

• A new AI startup called Upscale AI that creates and optimizes for streaming ads has emerged from stealth with $5.5 million in funding. See this LinkedIn post by co-founder Kevin Weatherman for more details.

• IT consulting firm Movate has acquired data and AI company Prescience Decision Solutions, CEO Sunil Mittal shared on LinkedIn. 

• Enterprise data intelligence platform Alation has acquired the startup Numbers Station AI, which builds AI-native data applications in a bid to help its customers access AI agents that run on top of their structured data. See this LinkedIn post by Numbers Station investor Scott Beechuk for more information. 

• GitHub has hired former Microsoft and Meta executive Vladimir Fedorov as its new chief technology officer, he shared on LinkedIn. 

• ServiceNow has hired former GE and Salesforce executive Sydney Williams as its vice president of global brand marketing, she shared on LinkedIn.

• AI robotics startup Figure has promoted its director of systems integrations Kyle Edelberg to chief engineer, he shared on LinkedIn. 

• Google DeepMind research director Aleksandra Faust is departing the research lab after a decade, she shared on LinkedIn. 

• Mayo Clinic has hired Micky Tripathi as its chief AI implementation officer, president and CEO Gianrico Farrugia, M.D. shared on LinkedIn.  

As always, pitch me the interesting investors, founders, ideas and companies powering AI, and reach the inboxes of nearly 1 million subscribers plus thousands more on LinkedIn. Follow me for other tech updates, and click 'Subscribe' to be notified of future editions.