For months, Google has stood firm on its message that the web is healthy, AI tools aren’t draining clicks, and its search engine is funneling people to a wider variety of websites than ever before.

However, a recent court filing revealed a strikingly different narrative.

In a filing submitted last week, ahead of a trial over its dominance in the advertising technology market, Google acknowledged that “the open web is already in rapid decline.” The admission, spotted by analyst Jason Kint and reported by Search Engine Roundtable, marks a sharp departure from the company’s upbeat public narrative.

The disclosure is part of Google’s strategy to defend itself against the U.S. Department of Justice (DOJ), which has recommended breaking up the company’s ad tech business. Google argues that such a divestiture would not revive competition but instead “accelerate” the collapse of the open web, harming publishers reliant on open-web display advertising.

Google’s Filing

“Finally, while Plaintiffs continue to advance essentially the same divestiture remedies they noticed in their complaint filed in January 2023, the world has continued to turn,” Google wrote in the filing.

The company added that the DOJ is acting as if the “incredibly dynamic ad tech ecosystem had stood still” while judicial proceedings dragged on.

According to Google, ad tech is undergoing seismic changes: AI is reshaping the industry, and advertisers are shifting rapidly toward formats outside the open web, such as Connected TV and retail media.

“The fact is that today, the open web is already in rapid decline and Plaintiffs’ divestiture proposal would only accelerate that decline, harming publishers who currently rely on open-web display advertising revenue,” the filing read.

Public Narrative vs. Courtroom Reality

The admission stands in stark contrast to Google’s recent public defense of its search product. As AI-powered tools such as ChatGPT and Perplexity gain popularity and Google itself rolls out AI-powered search overviews, many publishers and independent site owners have reported declining traffic. Yet, Google executives have consistently countered that the web is “thriving.”

In May, CEO Sundar Pichai told Decoder that Google was “definitely sending traffic to a wider range of sources and publishers” with its new AI search features. Nick Fox, Google’s senior vice president of knowledge, echoed that defense in June on the AI Inside podcast, insisting “the web is thriving.” Search chief Liz Reid cited Pew Research data to argue that while user behavior is changing, click volume remains “relatively stable” year-on-year.

After the courtroom filing surfaced, Google sought to downplay the remarks. Company spokesperson Jackie Berté told The Verge the “decline” line was being misrepresented, emphasizing that the context referred to “open-web display advertising,” not the open web as a whole.

“We are pointing out the obvious: that investments in non-open web display advertising like connected TV and retail media are growing at the expense of those in open web display advertising,” Berté said.

Still, the dissonance between Google’s courtroom arguments and its public messaging highlights the shifting reality of the digital economy. Publishers—already struggling with reduced referral traffic and tighter advertising margins—see the rise of AI-driven answers as an existential threat, one Google seems reluctant to acknowledge outside a courtroom.

Some analysts note that Google’s acknowledgement, even if framed around advertising rather than search, reinforces the mounting challenges faced by newsrooms and independent sites. With advertising dollars flowing toward walled gardens such as Amazon’s retail media network, YouTube, and streaming platforms, the open web’s share of digital ad spending continues to shrink.

As the DOJ trial continues, Google’s lucrative ad tech business faces an existential challenge – that could redefine how digital advertising operates. For publishers caught in the middle, the fight is not just about competition—it’s about survival.

Artificial intelligence startup Cognition has cemented its place among the industry’s elite, announcing on Monday that it closed a $400 million funding round at a $10.2 billion valuation — a milestone that includes the newly raised capital.

The funding, led by Peter Thiel’s Founders Fund with participation from Lux Capital, 8VC, and Bain Capital Ventures, underscores the appetite among investors for high-growth AI companies despite growing concerns over saturation in the sector.

Cognition is best known as the developer of Devin, an AI software engineer designed to automate and accelerate coding tasks. The company vaulted into the spotlight in July when it acquired Windsurf, a rival startup that had recently been rocked by leadership departures. That bold move has since redefined Cognition’s trajectory.

Windsurf’s turbulent path

Windsurf had been in advanced talks with OpenAI for a multibillion-dollar acquisition earlier this year, but the negotiations collapsed. Soon after, its founders and top researchers defected to Google, in a deal that included a $2.4 billion payout from the search giant in licensing fees and compensation.

In the aftermath, Cognition swooped in to acquire Windsurf, a move that surprised many in Silicon Valley given the sudden shifts in its leadership. The acquisition has already delivered measurable gains. According to CEO Scott Wu, Cognition’s annual recurring revenue (ARR) has more than doubled since the deal closed, with the combined enterprise ARR growing more than 30 percent in just two months.

“We’ll continue to invest significantly in both Devin and Windsurf, and our customers are already seeing how powerful the combination is together,” Wu wrote in a blog post announcing the funding.

The integration has been aided by Jeff Wang, who stepped in as CEO of Windsurf following the founders’ exit. In a LinkedIn post on Monday, Wang reassured customers that the tighter alignment between Windsurf and Cognition products would unlock new synergies: “It’s been quite an eventful last few months, and now it’s time to show what we’re made of.”

The $10 billion club

Cognition’s fresh valuation puts it in rare company, joining the likes of OpenAI and Anthropic, which have already established themselves as industry leaders. It also follows last week’s news that Sierra, the AI startup founded by former Salesforce co-CEO Bret Taylor, raised $350 million at a $10 billion valuation.

The flurry of billion-dollar valuations illustrates how quickly capital is flowing into the AI sector, but also highlights the mounting competition among startups trying to carve out space in a market dominated by Big Tech players such as Microsoft, Google, and Amazon.

For investors like Thiel’s Founders Fund, Cognition’s trajectory offers a chance to back an AI player that is both product-focused (through Devin) and strategically opportunistic (through the Windsurf acquisition). By comparison, OpenAI has leaned heavily on its partnership with Microsoft, while Anthropic has built its war chest through backing from Amazon and Google. Cognition’s independent path — combining product innovation with aggressive M&A — could set it apart in the long run.

The big question is whether Cognition can sustain its growth at scale. While ARR expansion has been promising, the company will need to prove that its combined suite of AI engineering and development tools can achieve widespread enterprise adoption. If successful, Cognition could emerge as one of the few independent challengers capable of standing alongside the most well-funded players in the industry.

President Donald Trump will be accompanied by some of America’s most powerful business executives on his state visit to the United Kingdom next week, with Nvidia CEO Jensen Huang among the delegation, a source familiar with the plans told CNBC’s Kristina Partsinevelos on Monday.

According to Sky News, the group will also include Sam Altman of OpenAI, Stephen Schwarzman of Blackstone, and Larry Fink of BlackRock. Apple CEO Tim Cook received an invitation as well, though it remains unclear if he will attend. The trip will feature a state banquet hosted by King Charles, where Trump and the business luminaries are expected to be present.

CEOs seek soft landing amid Trump’s tariff war

The growing presence of leading technology and financial executives at Trump’s side highlights how corporate America is recalibrating its strategy amid the president’s escalating tariff war. With new tariffs targeting Chinese imports and the broader tech sector under increasing regulatory scrutiny, many CEOs see proximity to Trump as a soft-landing strategy—a way of shielding their companies from harsher restrictions or positioning themselves to secure exemptions.

Nvidia is a clear example. Earlier this year, the Trump administration cut off Nvidia’s access to the Chinese market by restricting sales of advanced AI chips. Huang immediately moved to build a closer relationship with Trump, arguing that U.S. leadership in AI required some level of continued access to Chinese buyers. His lobbying paid off: after a series of White House meetings over the summer, Trump approved export waivers for Nvidia’s H20 chip, a move that helped the company recoup potential losses.

Following those talks, Trump also opened consideration for the export of other Nvidia chips to China, a step that Huang recently told investors presented a “real possibility” for approval. Such decisions could prove pivotal to Nvidia’s growth, given that the company had earlier been forced to scrap H20 chips that might have accounted for $8 billion in quarterly sales.

Depending on the geopolitical environment, Nvidia now estimates it could sell as much as $5 billion worth of H20 chips in the current quarter.

Trump has publicly praised Nvidia for being a technological leader, celebrating the moment the company’s valuation soared past $4 trillion. Huang, in return, has positioned Nvidia as a national asset, telling the administration that allowing exports strengthens U.S. national security by ensuring that the U.S. stays ahead in the global AI race.

Trump even claimed that he negotiated a 15% cut of Nvidia’s chip sales in China, though Nvidia clarified last month that details of the government’s share had not been finalized.

A pattern of engagement

However, Huang’s attendance on the U.K. trip is consistent with his broader engagement with Trump. In May, he joined the president on a visit to Saudi Arabia for an investment forum, another sign of Nvidia’s strategy of staying closely aligned with the administration’s global economic initiatives.

Meanwhile, other CEOs are navigating similar pressures. While Altman’s OpenAI depends on U.S. government support for research and regulatory clarity, and Cook’s Apple has been vulnerable to tariffs on Chinese-made devices, each executive has reasons to keep ties with the White House warm.

Diplomacy through business

The inclusion of executives like Huang, Altman, Schwarzman, and Fink on the U.K. state visit underscores how Trump is fusing business leadership with foreign policy, using corporate figures as both economic ambassadors and political allies. Their presence at the banquet with King Charles signals a deliberate effort to project American corporate strength abroad, even as tensions over trade, AI, and semiconductor dominance with China continue to loom large.

But for U.S. firms, these appearances are more than ceremonial—they are part of a broader survival strategy in an era of tariffs, export controls, and geopolitical realignments.

Google is pushing forward with a new wave of Gemini-powered updates, adding more versatility to its AI products and broadening access to its Search and research tools.

The company on Monday rolled out three key improvements: audio file compatibility for the Gemini app, five new languages for Search’s AI Mode, and expanded report formats for NotebookLM.

Josh Woodward, vice president of Google Labs and Gemini, announced the changes in a post on X, highlighting that audio uploads were the “#1 request” from users of the Gemini app.

Audio comes to Gemini

The addition marks a significant step for the app. Free users will be able to upload audio up to 10 minutes in length and are capped at five prompts each day. Those subscribed to AI Pro or AI Ultra plans can upload audio recordings up to three hours. Gemini now accepts up to 10 files per prompt across multiple formats, including compressed ZIP folders.

NotebookLM, Google’s research-oriented tool, already supported audio, but the Gemini app’s adoption widens its reach to users who want transcription, analysis, or conversation with audio content on the go.

Search expands with new languages

Google Search’s AI Mode is also extending its reach globally. With the integration of Gemini 2.5, the service now supports Hindi, Indonesian, Japanese, Korean, and Brazilian Portuguese.

A company blog post described the move as a major step toward inclusivity: “With this expansion, more people can now use AI Mode to ask complex questions in their preferred language, while exploring the web more deeply.”

The update signals Google’s intent to make Search’s AI features more accessible outside English-speaking markets, a push likely aimed at strengthening its footprint in Asia and Latin America, where rival platforms have been growing.

NotebookLM becomes a content generator

Meanwhile, NotebookLM is moving from a research companion into a content generator. The tool can now create structured reports in a variety of styles—study guides, briefing documents, blog posts, flashcards, and quizzes—drawing directly from a user’s uploaded files.

Reports are available in more than 80 languages and can be customized in tone, structure, and format. According to Google, the new features “should be 100%” rolled out by the end of this week.

This makes NotebookLM more competitive with general-purpose AI writing tools, while maintaining its edge as a research platform that can detect themes, patterns, and connections across diverse file types.

A month of rapid releases

The updates cap a whirlwind stretch for Google’s AI division. In August, Gemini began automatically recalling user details and preferences from prior conversations, while free users gained access to Workspace’s video generator, Vids. In September, Photos upgraded to the latest video model, Veo 3, and introduced the ability for free users to generate short, four-second silent videos from their still images.

The steady drumbeat of updates underscores Google’s urgency in the AI race, as it tries to match and outpace rivals like OpenAI and Anthropic. By layering audio, languages, and flexible content creation into its products, Google is signaling that Gemini is not just a conversational AI—but an ecosystem designed to permeate everyday work, study, and online exploration.

How Google’s updates compare in the AI race

Google’s recent spree comes as competitors are consolidating their own AI bets. OpenAI, backed by Microsoft, has been integrating its models more deeply into Office, Windows, and Azure while experimenting with memory features for ChatGPT. Meta, meanwhile, has doubled down on Llama and struck a high-profile partnership with Midjourney to strengthen its image-generation capabilities. Anthropic has been refining its Claude assistant to attract enterprise clients.

Against this backdrop, Google’s Gemini upgrades stand out for their consumer focus—rolling advanced features like audio uploads, multi-language Search, and structured reports directly into apps used by students, workers, and casual users. While OpenAI and Meta emphasize ecosystems tethered to enterprise and social platforms, Google appears intent on embedding Gemini in everyday digital habits, from search to note-taking to personal research.