The week of May 9 through 16, 2026, delivered some of the most consequential AI developments of the year so far, a funding round that could value Anthropic near a trillion dollars, Google redefining Android as an intelligence platform, OpenAI opening its advertising engine to businesses of every size, and a wave of AI-driven layoffs reshaping the corporate workforce from crypto exchanges to global e-commerce.
Key stories this week:
- Anthropic targets a $900 billion valuation in a planned $50 billion funding round
- Google unveils Gemini Intelligence at the Android Show, positioning Android as an agentic OS
- OpenAI’s self-serve Ads Manager goes live for U.S. advertisers with CPC bidding
- GPT-5.5 Instant becomes ChatGPT’s new default model, cutting hallucinations by 52.5%
- Claude Mythos Preview autonomously finds thousands of zero-day vulnerabilities across every major OS
- Cohere acquires Germany’s Aleph Alpha in a sovereign AI deal valued at approximately $20 billion
- Coinbase cuts 700 jobs (14% of workforce) in an explicit AI restructuring
- Amazon continues rolling layoffs as AI reshapes its Selling Partner Services division
- Novo Nordisk partners with OpenAI to accelerate drug discovery end-to-end
- Google I/O 2026 set for May 19, with a new Gemini model expected as centerpiece
- Five Eyes intelligence agencies publish joint guidance on agentic AI security
- Big Tech’s combined 2026 AI capex approaches $725 billion as workforce cuts accelerate
Anthropic Eyes $900 Billion Valuation in Planned $50 Billion Fundraise
Anthropic’s planned funding round is taking shape, aiming to raise up to $50 billion at a valuation of roughly $900 billion, a figure that would make it one of the most valuable private companies in history and signal that investor appetite for frontier AI has not cooled despite broader market uncertainty. The round reflects surging enterprise demand for Claude across regulated industries, with Anthropic’s annualized revenue approaching $19 billion as of early May.
The financial composition of Anthropic’s customer base is striking. About 40% of its top 50 customers are financial institutions, which points to how deeply Claude has embedded itself in compliance-sensitive, high-stakes environments where accuracy and auditability matter more than raw conversational capability. CEO Dario Amodei has been direct about the competitive stakes: at a May financial services event, he warned that SaaS firms failing to adopt AI risk obsolescence, while simultaneously calling for regulatory frameworks around the most powerful model releases.
The $900 billion target arrives at an inflection point in the enterprise AI market. Anthropic now wins about 70% of head-to-head matchups against OpenAI among businesses purchasing AI services for the first time, and Anthropic’s fight with the Trump administration may actually be benefiting the company as enterprise buyers seek providers with a credible safety narrative. The gap between Anthropic and OpenAI in revenue is narrowing rapidly, but the gap in enterprise trust appears to be widening in Anthropic’s favor, a dynamic that investors are clearly pricing into the round’s terms.
Source: AIToolsRecap | https://aitoolsrecap.com/Blog/ai-news-may-9-2026
Claude Mythos Preview Finds Thousands of Zero-Day Vulnerabilities Autonomously
Anthropic’s restricted-access model Claude Mythos Preview has reset the conversation about what AI can do in cybersecurity, and how much caution the industry should exercise before models with these capabilities reach general availability. Anthropic used Claude Mythos Preview to identify thousands of zero-day vulnerabilities, many of them critical, in every major operating system and every major web browser, along with a range of other important pieces of software.
The technical demonstration is striking in its specificity. Mythos Preview fully autonomously identified and then exploited a 17-year-old remote code execution vulnerability in FreeBSD that allows anyone to gain root on a machine running NFS, triaged as CVE-2026-4747, and no human was involved in either the discovery or exploitation of this vulnerability after the initial request to find the bug. Anthropic has also noted that Mythos Preview saturates most existing cybersecurity benchmarks, forcing the company to pivot toward real-world zero-day discovery as a more meaningful measure of the model’s capabilities.
Anthropic’s own system card confirms the model can run “end-to-end cyber-attacks on at least small-scale enterprise networks with weak security posture,” and it is the first model to complete a 32-step corporate network attack simulation, an operation that would take humans with some technical expertise 20 hours to complete. The company has released Mythos Preview only to a limited group of critical infrastructure partners and open-source developers through Project Glasswing, precisely to give defenders a head start before similar capabilities become broadly available. The Five Eyes intelligence agencies, whose joint guidance is covered below, released their agentic AI security document in the same week, a timing that was not coincidental.
Source: Anthropic | https://www.anthropic.com/glasswing
Google Launches Gemini Intelligence at Android Show, Turns Android Into an Agentic Platform
At its Android Show: I/O Edition event on May 12, Google made its most ambitious AI platform statement yet. Google announced a number of new Gemini Intelligence-branded AI features, including the ability for AI to complete tasks across apps, browse the web, fill out forms, dictate speech, and allow users to vibe-code their own Android widgets. The announcement reframes Android not as an operating system but as an intelligence layer. Sameer Samat, president of the Android Ecosystem, put it plainly during the keynote by saying Google is transforming Android from an operating system into an intelligence system.
The feature set is meaningfully agentic rather than cosmetically so. Gemini will be able to handle a multistep process, like copying a grocery list from your notes app, then adding items to the cart in your shopping app, activated by pressing the phone’s power button and describing the task. The Rambler feature in Gboard lets users speak naturally, then cleans up filler words and reformats speech in real time. A new Create My Widget tool lets users generate custom home screen widgets through natural language descriptions. Gemini Autofill can pull data from connected Google services to complete forms dynamically, and the company says it is opt-in. These features are launching this summer on Pixel 10 and Galaxy S26 before rolling out to watches, cars, glasses, and laptops later in 2026.
The strategic significance extends well beyond phone features. Google also unveiled Googlebooks, a new line of AI-first laptops built with partners including Acer, Asus, Dell, HP, and Lenovo, shipping this fall. Googlebooks are the first laptops designed from the ground up for Gemini Intelligence, featuring Magic Pointer (a cursor with Gemini built in), compatibility with Android phones, and custom widget creation. Android Auto is also getting its biggest Maps update in a decade, alongside Gemini integration across more than 250 million vehicles. Google I/O on May 19 is expected to go further still, with a new flagship Gemini model anticipated as the centerpiece announcement, likely landing in the class of a Gemini 3.5 or 4.0 based on pre-event reporting.
Source: Google Blog | https://blog.google/products-and-platforms/platforms/android/gemini-intelligence/
OpenAI’s Self-Serve ChatGPT Ads Manager Goes Live, Targets $2.5 Billion Ad Revenue in 2026
OpenAI’s push into advertising moved from pilot to platform during this week’s window. OpenAI launched a self-serve advertising platform, marking a significant step in its goal of generating $2.5 billion in ad revenue this year and $100 billion by 2030. A beta version of the new Ads Manager tool began rolling out to advertisers in the U.S. The move eliminates the minimum spend barrier that previously locked out smaller buyers, making ChatGPT’s 900 million weekly active users accessible to businesses of any size.
Advertisers can now create ChatGPT ads through partners or a new beta self-serve Ads Manager, and OpenAI is also introducing cost-per-click (CPC) bidding and expanded measurement tools, giving businesses more flexible ways to buy, manage, and understand campaign performance without sharing conversations or personal details with advertisers. Agency partners include Dentsu, Omnicom, Publicis, and WPP, while technology integrations span Adobe, Criteo, Kargo, Pacvue, and StackAdapt. A Reuters report in March suggested OpenAI banked over $100 million in revenue for its U.S. ad pilot after just six weeks, a signal that demand from advertisers was robust even before the self-serve layer opened.
The structural question this development raises is whether ChatGPT’s core value proposition, a utility people trust for accurate, unbiased answers, survives aggressive monetization. OpenAI has been emphatic that ads do not influence the organic model, with ads appearing as clearly labeled, subtly tinted boxes at the bottom of responses rather than embedded within them. Pro, Business, Enterprise, and Edu accounts see no ads. The expansion is also going international: the ad pilot is set to extend into the UK, Japan, Brazil, Mexico, and South Korea. For Google and Meta, the maturation of OpenAI’s ad platform represents a direct challenge to their combined search and social advertising dominance, particularly as ChatGPT increasingly captures high-intent queries that previously flowed through search.
Source: OpenAI | https://openai.com/index/new-ways-to-buy-chatgpt-ads/
GPT-5.5 Instantly Becomes ChatGPT’s Default Model, Slashes Hallucinations Sharply
OpenAI updated ChatGPT’s default model, available to everyone, replacing GPT-5.3 Instant with GPT-5.5 Instant, designed to be smarter and more accurate, with clearer, more concise answers that feel better tailored to the user. The rollout on May 5 marked a practical upgrade for the hundreds of millions of people who rely on ChatGPT daily, not just for power users running complex reasoning tasks.
The hallucination reduction numbers are meaningful. In internal evaluations, GPT-5.5 Instant produced 52.5% fewer hallucinated claims than GPT-5.3 Instant on high-stakes prompts covering areas like medicine, law, and finance. The model also reduced inaccurate claims by 37.3% on especially challenging conversations users had flagged for factual errors, a direct response to criticism that AI assistants remain unreliable for consequential decisions. The new model achieved a score of 81.2 on the AIME 2025 math test compared to 65.4 for its predecessor, and outperformed on the MMMU-Pro multimodal reasoning benchmark with 76 versus 69.2.
Personalization is another pillar of this release. Enhanced personalization from past chats, files, and connected Gmail is rolling out to Plus and Pro users on the web, with plans to extend to Free, Go, Business, and Enterprise users in the coming weeks. Memory sources, a new transparency layer letting users see and correct what context the model is drawing on, are rolling out across all consumer plans. For developers, GPT-5.5 Instant is available via API as “chat-latest,” with GPT-5.3 Instant remaining available for paid users for three months before retirement.
Source: OpenAI | https://openai.com/index/gpt-5-5-instant/
Cohere Acquires Aleph Alpha in $20 Billion Sovereign AI Deal
Canadian AI startup Cohere announced it would merge with Aleph Alpha, a German company building AI systems for businesses and governments, with the explicit goal of giving enterprises and governments an alternative to dominant U.S. AI providers, one that offers greater independence and control over their data. The deal, which values the combined entity at approximately $20 billion, is being backed by Schwarz Group (the German retail conglomerate behind Lidl) and received formal support from both the Canadian and German governments, with Digital Ministers from each country attending the announcement in Berlin.
As part of the deal, Schwarz Group plans to invest $600 million in Cohere’s upcoming Series E round. Cohere, last valued at $7 billion in 2025, will lead the new entity, with CEO Aidan Gomez staying in place. The merged entity will target heavily regulated industries, including defense, finance, and healthcare, as well as European public sector organizations seeking alternatives to U.S.-controlled AI infrastructure. Aleph Alpha brings a team of 250 people and deep relationships with the German public sector, including contracts with the German Ministry for Digital Affairs and state governments in Baden-Württemberg.
The deal is analytically significant for what it reveals about the state of the enterprise AI market. Training runs for state-of-the-art models now exceed $100 million, creating existential pressure on undercapitalized AI companies that cannot keep pace with the hyperscalers. While Cohere reported $240 million in annual recurring revenue in 2025, Aleph Alpha had generated little revenue and significant losses, meaning this is a consolidation born of necessity as much as strategy. For U.S. AI companies, the deal may accelerate similar mergers among European and mid-tier global players who cannot independently sustain frontier competition.
Source: TechCrunch | https://techcrunch.com/2026/04/25/why-cohere-is-merging-with-aleph-alpha/
Coinbase Cuts 700 Jobs Citing AI Shift, Becomes a Canary in the Workforce Coal Mine
On May 5, Coinbase announced a restructuring plan involving a reduction of its workforce by approximately 700 employees, representing approximately 14% of the company’s global workforce as of May 1, 2026, to manage operating expenses and optimize for the AI era. The restructuring is expected to cost $50 million to $60 million in severance and termination benefits, substantially all cash-based, recognized in Q2 2026.
CEO Brian Armstrong’s framing of the cuts went further than most corporate restructuring announcements. Rather than citing market conditions alone, Armstrong described Coinbase as “fundamentally changing how we operate, rebuilding the company as an intelligence, with humans around the edge aligning it.” AI mentions more than doubled in Coinbase-adjacent job listings in 12 months, rising from 23% in early 2025 to 53.1% by March 2026, and more than half of Web3 job listings now ask for AI proficiency outright. The company is explicitly hiring for what it calls “Agent Managers”, people who oversee AI agents rather than execute tasks themselves.
Coinbase’s situation is a concentrated version of a pattern playing out across the technology sector. Google, Amazon, Meta, and Microsoft collectively committed roughly $725 billion in AI capital expenditure in 2026, up approximately 77% year over year. Nikkei Asia attributed 47.9% of Q1 tech layoffs to AI and automation, while Bloomberg data suggests roughly half of AI-attributed layoffs will result in the same roles being rehired offshore or at lower salaries, framing this as a labor repricing story rather than a pure reduction. The workers being laid off are largely not the workers being hired for the 275,000 AI-related roles sitting open in the U.S. market, a skills mismatch that is widening rather than narrowing.
Source: SEC Filing (Coinbase 8-K) | https://www.sec.gov/Archives/edgar/data/0001679788/000167978826000049/coin-20260505.htm
Amazon’s AI-Driven Layoff Cycle Continues with Selling Partner Services Cuts
Amazon’s workforce reductions in 2026 represent a sustained restructuring rather than a single event, with its Selling Partner Services division becoming the latest unit affected. Amazon confirmed that a “small number” of employees in the Selling Partner Services team were affected in its latest round, adding to the company’s series of workforce reductions across divisions, including gaming, cloud services, devices, advertising, and HR. The cumulative total of Amazon job cuts in the past six months has reached approximately 30,000 positions.
In earnings calls, Meta and Amazon executives collectively referenced “efficiency” 15 times, while Microsoft’s top finance executive said headcount would decline in 2026 as the company stresses “pace and agility.” The efficiency narrative at Amazon is backed by concrete investment numbers: Amazon committed $200 billion in technology capital expenditure for 2026, making human payroll the most flexible cost to cut as infrastructure spending balloons. CEO Andy Jassy has repeatedly signaled that AI will help Amazon operate more efficiently over time and could eventually reduce portions of the workforce in operations, coordination, and support.
The implications for enterprise AI adoption are double-edged. On one side, large companies are demonstrating real productivity returns from AI integration, genuine enough to restructure around. On the other hand, the human cost of that transition is concentrated in roles that are hardest to retrain rapidly, and the skills premium for AI fluency has created a 56% wage differential between workers who have it and those who do not. Amazon’s ongoing restructuring is less a story about one company and more a stress test of whether the AI-efficiency thesis translates from investor decks to actual operating leverage.
Source: Business Insider / Outlook Business | https://www.outlookbusiness.com/amp/story/corporate/amazon-cuts-more-jobs-after-30000-layoffs-ai-expansion
Novo Nordisk and OpenAI Partner to Apply AI Across the Entire Drug Development Pipeline
Novo Nordisk announced a strategic partnership with OpenAI to apply artificial intelligence across its business, from drug discovery to manufacturing and commercial operations, with full company-wide integration targeted by the end of 2026. The deal is designed to accelerate the Danish drugmaker’s identification of new obesity and diabetes treatments as it works to regain market ground against U.S. rival Eli Lilly, whose oral weight-loss tablet Foundayo received FDA clearance in the same period.
Advanced AI will be applied end-to-end, extending beyond discovery into manufacturing, supply chain, and commercial execution, with a central component being workforce upskilling, where OpenAI will support Novo Nordisk in building AI fluency across the company’s global organization. CEO Mike Doustdar framed the partnership as positioning Novo Nordisk to lead in the next era of healthcare, explicitly noting that the goal is to accelerate scientists rather than replace them, though the company acknowledged AI would reduce future hiring growth in certain functions.
The partnership matters beyond its immediate commercial rationale. Big Pharma has poured billions into AI drug discovery across multiple years, yet no market-ready drug developed entirely through AI has reached patients. The industry consensus emerging from deals like this one is that AI delivers most immediate value in patient recruitment, regulatory submission, and supply chain optimization rather than in de novo molecule design. Novo Nordisk’s full-pipeline integration with OpenAI, covering both the science and the operations, is one of the most comprehensive pharma-AI deals disclosed in 2026 and will be watched closely as a template for whether AI can meaningfully compress the 10-year average drug development timeline.
Source: Novo Nordisk | https://www.ddw-online.com/novo-nordisk-partners-with-openai-across-operations-41337-202604/
Five Eyes Agencies Release Joint Agentic AI Security Guidance for Critical Infrastructure
Cybersecurity and intelligence agencies from the United States, Australia, Canada, New Zealand, and the United Kingdom jointly released a document titled “Careful Adoption of Agentic AI Services,” addressing security risks in agentic AI systems deployed in critical infrastructure and defense. The joint release is the first coordinated Five Eyes guidance document specifically focused on agentic systems, a recognition that AI agents capable of multi-step autonomous action represent a qualitatively different risk surface than traditional software deployments.
The document identifies five risk categories: privilege escalation, design and configuration vulnerabilities, behavioral unpredictability, structural risks in multi-agent pipelines, and accountability gaps when AI agents make consequential decisions without clear human authorization chains. The agencies stress that organizations should deploy agentic AI incrementally, maintain strong governance over what permissions agents are granted, and ensure rigorous human oversight is preserved, particularly for actions with irreversible consequences such as financial transactions, access control changes, or infrastructure modifications.
The timing relative to Anthropic’s Mythos Preview disclosure was not coincidental. The same week that Mythos demonstrated autonomous exploitation of zero-day vulnerabilities, the Five Eyes collectively flagged that AI agents introduce credential theft, prompt injection, and token exploitation risks that traditional cybersecurity tooling was not designed to catch. For enterprises deploying AI coding assistants, customer service agents, or autonomous workflow tools, the guidance provides a governance framework at a moment when the attack surface is expanding faster than security posture can adjust.
Source: AIToolsRecap | https://aitoolsrecap.com/Blog/ai-news-may-9-2026
Google I/O 2026 Preview: New Gemini Model, Android XR Glasses, and AI Search Evolution
Google I/O 2026 opens on May 19 in Mountain View, and the week’s buildup from the Android Show has set expectations high for one of the company’s most consequential developer conferences in recent years. Sources indicate that Google plans to announce a new Gemini model at I/O, with the release landing roughly in the class of OpenAI’s GPT-5.5, a significant step but, by some accounts, still short of the bar set by Anthropic’s Mythos.
Major Gemini model updates are widely expected to be the centerpiece announcement, with a newer, more capable version of Google’s flagship AI almost certainly coming. Android XR smart glasses built with Samsung are anticipated to receive more details, along with updates to Veo video generation, Lyria music generation, and Google’s AI Mode for search. The conference is also expected to address agentic AI capabilities for Vertex AI and Firebase, giving developers the tools to build the next generation of autonomous applications on Google’s cloud infrastructure.
As of Q1 2026, paid subscribers for Google services reached 350 million, with growth primarily driven by YouTube Premium and Google One, whose premium plan already includes access to Gemini Advanced. Google Cloud’s backlog is nearing $462 billion, a figure that reflects how deeply enterprise AI workloads are consolidating around the major cloud providers. The market is watching I/O not just for product announcements but for a clear articulation of how Gemini translates into Google’s commercial growth story, particularly as AI Mode evolves toward becoming the default search experience.
Source: Yahoo Tech | https://tech.yahoo.com/general/article/google-io-2026-what-to-expect-next-week-including-android-17-ai-announcements-and-more-131200995.html
Key AI Funding and Valuation Snapshot: May 9–16, 2026
| Company | Event | Valuation / Amount | Key Detail |
|---|---|---|---|
| Anthropic | Planned funding round | $900B valuation / $50B raise | ~40% of top customers are financial institutions |
| Cohere + Aleph Alpha | Acquisition/merger | ~$20B combined entity | Backed by Schwarz Group’s $600M Series E commitment |
| Meta (AI capex) | 2026 infrastructure spend | $115–135B committed | Nearly double 2025 capex; includes Muse Spark model launch |
| OpenAI | ARR milestone | $25B annualized revenue | Exploring IPO path; Anthropic at ~$19B ARR |
| Coinbase | Restructuring cost | $50–60M severance | 700 employees cut (14% of workforce) |
| Amazon | Cumulative 2026 cuts | ~30,000 positions | Latest round: Selling Partner Services division |
Closing
The week of May 9 through 16, 2026, will be remembered as the period when several of the AI industry’s most consequential long-term questions moved from theoretical to operational: Can Anthropic sustain its valuation trajectory while competing with OpenAI’s revenue lead? Can Google turn Android’s 3-billion-device installed base into an agentic computing platform before Apple responds at WWDC? Can OpenAI build a durable advertising business without eroding the trust that made ChatGPT indispensable?
And, cutting across all of it, can the economic disruption from AI-driven workforce restructuring be absorbed at the pace these companies are accelerating? Google I/O on May 19 arrives at exactly this inflection point; the answers that come from Mountain View will shape the competitive landscape for the rest of 2026.
Frequently Asked Questions
What is Anthropic’s current valuation in 2026, and why is its latest funding round significant?
Anthropic is targeting a valuation of approximately $900 billion in a planned $50 billion funding round, which would make it one of the most valuable private companies ever formed. The round reflects surging enterprise adoption of Claude, particularly in financial services, and Anthropic’s annualized revenue approaching $19 billion, a rapid climb from $1 billion at the end of 2025.
What is Claude Mythos Preview, and why is Anthropic not releasing it publicly?
Claude Mythos Preview is a restricted-access Anthropic model with advanced autonomous cybersecurity capabilities, including the ability to find and exploit zero-day software vulnerabilities with no human intervention. Anthropic has limited its rollout to critical infrastructure partners through Project Glasswing specifically because broad access would give attackers the same capability currently being offered to defenders.
What is Gemini Intelligence, and how does it change Android?
Gemini Intelligence is Google’s new agentic AI layer for Android, announced on May 12 at the Android Show. It allows Android to complete multi-step tasks autonomously across apps, such as copying a grocery list and adding items to a shopping cart, and introduces features including Rambler speech cleanup, Create My Widget for natural-language widget creation, and Gemini-powered Autofill, all launching on Pixel 10 and Samsung Galaxy S26 devices this summer.
How does OpenAI’s self-serve Ads Manager work for businesses?
The beta Ads Manager, launched May 5, allows U.S. advertisers of any size to create and manage campaigns directly inside ChatGPT. Advertisers can use CPM or CPC bidding, access measurement tools including a Conversions API and pixel-based tracking, and work through agency partners such as Dentsu, Omnicom, Publicis, and WPP. Ads appear as clearly labeled sponsored units at the bottom of responses and do not influence the model’s answers.
Why did Coinbase lay off 14% of its workforce in May 2026?
Coinbase cut approximately 700 employees on May 5, citing a deliberate restructuring toward an AI-first operating model. CEO Brian Armstrong framed the cuts as the company moving from a traditional headcount structure to one where smaller teams manage AI agents, reducing the need for individual contributor roles in areas increasingly handled by automation.
What is the significance of the Cohere and Aleph Alpha merger?
The deal creates a transatlantic enterprise AI company with backing from both the Canadian and German governments, targeting regulated industries that require sovereign AI, systems where organizations retain full control over their data without routing it through U.S. hyperscalers. The combined entity is valued at approximately $20 billion, with Schwarz Group (owner of Lidl) committing $600 million to Cohere’s upcoming Series E.
What did the Five Eyes agentic AI security guidance say?
The joint guidance from cybersecurity agencies in the U.S., UK, Australia, Canada, and New Zealand identified five agentic AI risk categories: privilege escalation, design flaws, behavioral unpredictability, structural pipeline risks, and accountability gaps, and recommended incremental deployment, strong governance over agent permissions, and rigorous human oversight for irreversible decisions.
What is GPT-5.5 Instant, and who does it affect?
GPT-5.5 Instant became ChatGPT’s default model for all users on May 5, replacing GPT-5.3 Instant. It produces 52.5% fewer hallucinations on high-stakes topics such as medicine, law, and finance compared to its predecessor, uses fewer words per response, and introduces memory sources, a transparency feature showing users what past context the model is referencing to shape its answers.
How much is Big Tech spending on AI infrastructure in 2026?
Google, Amazon, Meta, and Microsoft are collectively committing approximately $725 billion in AI capital expenditure in 2026, up roughly 77% year over year. Amazon committed $200 billion, Microsoft $190 billion, and Meta raised its full-year guidance to $125–145 billion. These figures represent a historically unprecedented rate of infrastructure investment, partially offset by workforce reductions across all four companies.
What can we expect from Google I/O 2026 on May 19?
Google I/O is expected to feature a new flagship Gemini model, potentially Gemini 3.5 or 4.0, as its centerpiece, alongside expanded agentic AI capabilities for developers, updates to AI Mode in Google Search, further details on Android XR smart glasses built with Samsung, and new tools for Vertex AI and Firebase. The conference follows the Android Show’s Gemini Intelligence announcement and sets the direction for Google’s entire AI roadmap through the end of 2026.
