The numbers tell a stark story

Just over a year ago, Similarweb reported that Google’s AI-generated search summaries — officially called AI Overviews — appeared in roughly 15% of all queries. Fast-forward to May 2026, and that figure has nearly tripled. According to the firm’s latest analysis, AI Overviews now show up in 43% of searches.

That’s not a gradual uptick. It’s a pivot. What began as a tentative AI layer on top of search results has become a core feature of the experience itself. Users are increasingly landing on pages where the first thing they see isn’t a list of blue links but a block of machine-written text that synthesizes information from across the web.

AI Mode is booming, too

Google’s more conversational AI Mode — a separate interface where users can ask follow-up questions and get multi-turn answers — is also growing fast. Similarweb data shows that AI Mode visits climbed from 126 million in June 2025 to 279 million by May 2026. That’s more than double in less than a year.

The average length of a Google search query has also crept upward. People are moving away from short, keyword-driven strings like “best coffee maker 2026” and toward longer, natural-language questions such as “What’s the best drip coffee maker for a household of four under $200?” The shift reflects a broader behavioral change: users are treating Google less as a directory and more as an oracle.

Publishers are losing out

For news sites, blogs, and content publishers, this trend is painful. AI Overviews often include citations — the report notes that the number of AI responses containing a citation has grown more than fivefold over the past year — but citations don’t always translate into clicks. Readers get their answer without ever leaving Google’s ecosystem.

Last year, Similarweb already flagged how devastating this was for news publishers. Now the picture is clearer: Google is becoming the destination, not just the gateway. Users spend more time on Google’s platform, and less time visiting the sites that produced the original information.

Some publishers have started fighting back. Cloudflare, the tech infrastructure company, introduced tools that let publishers block AI crawlers unless those AI companies pay for access through its marketplace. It’s a small but growing resistance movement.

ChatGPT is a mixed bag for referrals

Meanwhile, across the AI search landscape, ChatGPT shows a different pattern. As of May 2026, only 6.8% of U.S. desktop queries on ChatGPT included citations. That’s low, but some sectors fare better: travel, retail, and sports queries generate cited responses more frequently than, say, health or finance topics.

There is a glimmer of good news for publishers. After a May 7 search update, the proportion of ChatGPT desktop visits that landed on actual webpages more than doubled — from 25% in March 2026 to nearly 60% by May 30. That suggests that when blue links are made more prominent within AI-driven results, users do click through.

Still, the overall direction is clear. Google is transforming itself from a search engine into a self-contained answer engine. Whether users or publishers like it, AI search is becoming the default — and the era of the simple list of links is fading fast.

How a $10 billion dead end became Apple’s AI engine room

For more than a decade, Apple poured money and talent into a secret project that never reached a single customer. The Apple Car — officially code-named Project Titan — was scrapped in early 2024 after burning through an estimated $10 billion. But according to Bloomberg reporter Mark Gurman, that spectacular failure may have given Apple something far more valuable than a vehicle: the technological foundation for Apple Intelligence.

It is one of those ironic twists that corporate historians will chew on for years. Apple’s most expensive flop might also be its most consequential long-term bet. The car is dead. The AI brain it helped create is just getting started.

Level 5 autonomy forced Apple to think like an AI company

From the outset, Apple did not want to build just another electric car. The company set its sights on Level 5 autonomous driving — the kind where a vehicle needs zero human input. No steering wheel. No pedals. Just a machine that sees, thinks, and acts entirely on its own.

That target was absurdly ambitious. It forced Apple’s engineers to solve a problem most tech companies had barely touched: how to process enormous AI workloads locally, in real time, inside a moving vehicle. The cloud was too slow. The latency would kill you. Every decision had to happen on board.

So Apple invested heavily in machine learning research and custom silicon. Engineers designed a dedicated AI chip specifically for the car. That chip never made it into a finished product. But the work did not disappear. It evolved into the Neural Engine, Apple’s dedicated AI processor now embedded in virtually every modern Apple chip.

The Neural Engine’s quiet arrival

The first Neural Engine shipped inside the iPhone X in 2017. It powered Face ID and Animoji — neat features, but hardly world-changing. Since then, Apple has quietly expanded the technology across its entire lineup. Every Apple Silicon Mac launched since 2020 includes a Neural Engine, giving laptops and desktops dedicated hardware to run AI tasks locally instead of phoning home to a server.

That local processing matters more now than ever. Apple Intelligence, the company’s suite of generative AI features, depends on it. The Neural Engine handles on-device inference for image generation, text summarization, and Siri improvements — all without sending your data to the cloud.

From a canceled car to the data center

Bloomberg reports that the abandoned vehicle project’s influence extends well beyond consumer gadgets. The same research reportedly shaped Apple’s Ultra-class Mac chips, the ones that combine two M-series dies into a single powerhouse. It also influenced the custom processors currently running Apple Intelligence servers in data centers.

Here is the part that makes the story sting a little less for Apple fans. While the company has struggled to ship AI software features as fast as Google or Microsoft, it has spent more than a decade quietly building the hardware required to support them. Those early investments are now beginning to pay off as Apple continues expanding Apple Intelligence and rebuilding Siri around more capable AI models.

Apple has been playing the long game. It just looked like it was wasting time.

What the Apple Car actually achieved

The Apple Car is often remembered as a textbook failure. It never reached customers. It consumed billions. It went through multiple leadership changes and strategy pivots. By any conventional measure, Project Titan failed.

Yet internally, the project achieved something arguably more valuable: it accelerated Apple’s expertise in AI hardware years before generative AI became the industry’s biggest battleground. The engineers who worked on the car’s neural network and silicon designs didn’t vanish when the project was canceled. They moved to other teams. The knowledge stayed inside the company.

In hindsight, Apple’s abandoned vehicle may never transport people from one place to another. But the technology it inspired is already helping power Apple’s next generation of AI experiences. That might end up being the far more important destination.

For a deeper look at how Apple is handling AI on its devices, check out our guide on Apple Intelligence features and compatibility. And if you are curious about the hardware that makes it all run, read our breakdown of Apple Silicon chips compared.

AMD puts $5 billion on the table for Anthropic

Advanced Micro Devices (AMD) is making a bold move. The chipmaker has agreed to invest up to $5 billion in Anthropic, the startup behind the Claude AI assistant. This isn’t just a financial bet — it’s a sprawling infrastructure deal that could reshape how AI models are trained and deployed.

The agreement covers tens of billions of dollars’ worth of AI systems. Anthropic will deploy up to two gigawatts of computing capacity using AMD’s next-generation Instinct MI450-series accelerators. The first gigawatt is expected to come online in the first half of 2027.

Here’s the twist: AMD’s investment is tied to deployment milestones. If Anthropic hits certain targets, the money flows. Neither company has disclosed what those conditions are, or what happens if delays creep in.

How the deal works — and how it’s different

This isn’t a simple check-writing exercise. The transaction combines an equity investment from AMD with a separate, large hardware order from Anthropic. The companies haven’t said Anthropic must use AMD’s investment to pay for the systems.

It’s a structure AMD has used before — but with a twist. In October, AMD agreed to supply OpenAI with systems supporting up to six gigawatts of capacity. That deal included warrants allowing the ChatGPT developer to acquire roughly 10% of AMD’s stock, vesting in stages based on deployment, commercial, and share-price conditions.

AMD struck a similar deal with Meta in February, covering up to six gigawatts of GPU capacity, with performance-based warrants tied to shipment and purchase targets.

The Anthropic agreement is different. AMD is making a direct equity investment, not issuing warrants. That puts more skin in the game for the chipmaker — and gives Anthropic cash it can use however it sees fit, within the broader partnership.

What Anthropic is actually buying

Anthropic will deploy AMD Helios systems. These aren’t just GPUs — they’re integrated rack-scale platforms combining MI455X accelerators from the Instinct MI450 series, EPYC “Venice” processors, Pensando networking hardware, and ROCm software.

This matters because it’s a full-stack approach. Instead of buying chips and figuring out the rest themselves, Anthropic gets a pre-integrated system designed to work together. AMD chair and CEO Lisa Su called it a “deepened partnership” that brings together Anthropic’s AI leadership with AMD’s high-performance computing.

Anthropic has already been testing AMD’s earlier MI355X accelerators. The new deal scales that relationship dramatically — from evaluation to gigawatt-scale deployment.

Gigawatts: What that number really means

Two gigawatts of power capacity is enormous. For context, Anthropic has separately secured more than 300 megawatts through SpaceX’s Colossus 1 facility in Memphis, which houses over 220,000 Nvidia GPUs. The AMD deployment represents more than six times that power capacity.

But power capacity isn’t the same as computing performance. The two facilities use different hardware and deployment models. AMD executives have said that building one gigawatt of AI infrastructure can cost tens of billions of dollars, depending on the equipment and facilities involved.

Some of the AMD systems will go into Anthropic’s own data centres. Others will be hosted by cloud providers and specialist AI infrastructure companies. AMD and Anthropic are also hunting for operators capable of hosting these massive systems.

Lisa Su noted that gigawatt-scale capacity requires 12 to 24 months of planning before deployment. That’s why the first systems won’t arrive until 2027.

Anthropic’s multi-supplier strategy takes shape

Anthropic is building a diverse compute network. The AMD systems will run alongside infrastructure based on Nvidia GPUs, Amazon Trainium processors, and Google tensor processing units.

Amazon remains Anthropic’s primary cloud and training partner. The startup has secured up to five gigawatts of capacity from Amazon and uses more than one million Trainium2 processors. Trainium3 deployments are planned for 2026.

Anthropic has also locked in multiple gigawatts of next-generation TPU capacity from Google and Broadcom, with deployments expected in 2027. Claude is available through Amazon Web Services, Google Cloud, and Microsoft’s Azure-based Foundry platform.

Anthropic co-founder and chief compute officer Tom Brown explained the logic: “Running across a diversified range of hardware lets us map the right workloads to the right hardware.” He said the AMD partnership gives Anthropic additional capacity while allowing the companies to optimise systems for training and serving Claude.

The company has also agreed to use computing capacity at SpaceX’s Colossus facilities and has discussed leasing infrastructure from Meta. Reuters reported that potential Meta deal could be worth up to $10 billion over two years.

Engineering collaboration and software development

The AMD agreement includes a multi-year engineering programme. The companies will use Claude to optimise workloads for Instinct accelerators and support development of AMD’s ROCm software platform. AMD plans to deploy Claude across its engineering and product-development teams.

This gives Anthropic a dual role: customer of AMD’s infrastructure and participant in developing the software that runs workloads on that infrastructure. It’s a tight feedback loop that could help AMD close the software gap with Nvidia’s CUDA ecosystem.

Market reaction and what comes next

AMD shares rose 2.4% after the agreement was announced. The company’s stock has more than doubled since the start of the year, outperforming the broader Philadelphia Semiconductor Index.

The deal positions AMD as a serious player in the AI infrastructure race, alongside Nvidia’s reported talks about investing up to $30 billion in OpenAI. Chip suppliers are increasingly pairing investments or equity incentives with commercial agreements involving major AI developers.

For Anthropic, the AMD deal provides another pillar in its compute strategy — and signals that the startup is thinking long-term about hardware diversity. “Access to compute is central to keeping Claude at the frontier and meeting demand from our customers,” Brown said. “By partnering with AMD across the stack, we are securing the capacity we need and optimising it for training and serving Claude.”

The real test will come in 2027, when the first gigawatt of AMD-powered infrastructure is supposed to go live. Until then, the industry will be watching — and counting the milestones.