Latest AI News
Databricks hits $188B valuation, extending its run as AI’s favorite second act
Databricks on Thursday announced a new round of funding that values the company at$188 billion. The round was led by Coatue. Databricks didn’t disclose exactly how much it raised; it said the money isn’t in its hands yet and that the round will close later in this summer. (Other outlets have since reported the raise is roughly$3 billion.) While it’s unusual for a company to announce before it gets the money, a VC tells TechCrunch that the deal is solid, with so many firms wanting in that the company had no reason to keep its shiny new valuation a secret. In fact, Databricks has been on a year-and-a-half fundraising tear as it successfully transitioned its image into an AI provider and not just a yesteryear SaaS sensation. Yesteryear being back in the BC times (Before ChatGPT). Only five months ago, in February, Databricks closed a $5 billion Series L raiseat a $134 billion valuation. Five months before that, in September 2025,it raised $1B at $100 billion valuation. And roughly nine months before that, in December 2024, it raised what was arecord-breaking round at the time of $10 billionat a $62 billion valuation. Databricks has raised so many rounds over the years that this latest one became the subject ofmemes about running out of lettersof the alphabet. “Turning on alerts for when we get a Series AA,” one person posted. But its image reconstruction has been legit. Founded in 2013, it initially grew to success back in the big data era, with software that enabled enterprises to store enormous amounts of data in the cloud, yet produce speedy analytics. Because it already sat on troves of enterprise data, Databricks was then well-positioned to respond as companies started wanting AI with the same security and governance they expect from traditional enterprise software. The company began rolling out one AI product after another, likeLakebase, its database built for AI agents, and Unity, its AI gateway, along with a “meta-harness” called Omnigent that manages multiple agents. Databricks also increasinglybecame knownas one of the big examples of enterprises adopting more affordable Chinese-based open-weight models (models whose underlying code is published for anyone to use and modify) for cost control,one of the big trends of 2026. It is a particular champion of Z.ai’s GLM 5.2 as a model for coding. Last week Databricks CEO Ali Ghodsishared the resultsof some internal benchmarking done to manage his own AI costs for his 3,000 software engineers. The company compared AI models on the actual tasks its programmers do. Not surprisingly,in the blog post revealing the results, Databricks shared that “open models, and GLM 5.2 in particular, are now able to handle even the highest level of task difficulty” in coding, and at a total lower cost than proprietary models from Anthropic and OpenAI. But it did surprise people by finding that the choice of harness — the agentic coding tool, like Codex or Claude Code, that wraps around a model and manages its context and instructions — equally impacted costs. It found that open-source harness Pi to be one of the best at managing context surrounding each prompt, and therefore one of the lowest costs choices without sacrificing quality. “The lesson here isn’t that one harness is always cheaper or that native harnesses are worse,” thepost declared. “Instead, model choice is only one piece of the puzzle.” All of this has added to Databricks image as an AI company, even if it wasn’t founded as an AI lab. This, in turn, has granted it the AI-halo for raising money and leaping its valuation. As we previously reported, the AI effect is so strong these days, that evensandwich shop Jersey Mike’s mentionedAI 22 times in its S-1 documents.
View
Vertu wants executives to pay $6,880 for an AI agent — here’s how it actually performs
AI has become the smartphone industry’s latest battleground, with manufacturers racing to add AI-powered features to attract mainstream consumers.Vertuis taking a different path: the UK-founded luxury phone maker, known for hand-finished devices often costing tens of thousands of dollars, has built its business selling status symbols to the ultra-wealthy rather than competing on specs. Its Alphafold targets affluent buyers, particularly chief executives, pairing luxury materials with an AI agent designed to automate parts of an executive’s working day. I decided to test that claim on its own terms. Rather than focusing on benchmark scores, camera comparisons, and media consumption — the staples of most smartphone reviews — I spent a few days using the foldable the way Vertu says its customers would: managing documents, analyzing spreadsheets and contracts, planning business trips, automating routine tasks, and relying on its AI agent as a digital companion throughout the working day. The question wasn’t whether it was a good smartphone, but whether it was a good executive smartphone. At the heart of the Alphafold is Hermes Agent, a pre-installed AI agent built on top of the open-sourceHermes project, which the company says can analyze files, automate tasks across apps, remember conversations, and hand off requests to a human concierge when needed. Unlike most smartphone AI assistants that largely just respond to prompts, Hermes is designed to execute multi-step workflows on users’ behalf, making it the centerpiece of Vertu’s pitch rather than the foldable hardware itself. Physically, the Alphafold, whichstarts at $6,880, looks and feels every bit like a luxury device. The review unit I received was wrapped in genuine calfskin leather with titanium accents, setting it apart from mainstream foldables that largely rely on glass or synthetic finishes. It’s clearly built for buyers who see their phone as both a tool and a status symbol. Compared with theSamsung Galaxy Z Fold 7, which I used as a reference device throughout this review, the 264-gram Alphafold feels noticeably heavier than Samsung’s 215-gram foldable. The extra weight is apparent during prolonged use, though it never feels unwieldy. The Alphafold’s curved frame also makes it easier to unfold than the Galaxy Z Fold 7’s flatter edges. Samsung’s design, however, feels sleeker and more comfortable to hold when folded, making it easier to use one-handed. The Alphafold also arrives in packaging that feels more akin to a jewelry presentation case than a smartphone box. The oversized box opens to reveal neatly arranged drawers containing bundled accessories, including a leather sleeve and charging cables, reinforcing the sense that Vertu is selling a luxury experience rather than just a handset. Beneath the premium materials, however, the Alphafold tells a different story. During the review, I noticed striking similarities between the device and the$1,100 ZTE Nubia Fold— from the hinge design and dimensions to the placement of the speakers, microphones, and the fingerprint reader. The most visible distinction is Vertu’s leather-clad rear panel, though. System information also revealed ZTE identifiers in parts of the software. When asked about these observations, Vertu confirmed to TechCrunch that the Alphafold was developed through a specialist supply-chain partnership involving ZTE/Nubia’s hardware platform, component integration, and production engineering. However, the company said it was responsible for the luxury materials, software experience, quality control, and after-sales service. ZTE did not respond to a request for comments. This isn’t new for Vertu. In a 2023 review of the MetaVertu, Wiredreportedthat the device appeared to be based on a ZTE Nubia handset, citing hardware similarities and comments from Counterpoint Research that Vertu had been adapting existing ZTE models with luxury materials and custom software. Still, focusing solely on the hardware misses the point of the Alphafold. Vertu’s real bet is not on building a better foldable but on whether executives will pay for an AI agent that helps them get through the working day more efficiently. Over several days, I used the Alphafold as my primary smartphone, replacing routine prompts with real executive-style workflows. Instead of asking Hermes to write emails or answer trivia questions, I tasked it with analyzing spreadsheets and contracts, planning business trips, managing my schedule, and automating actions across multiple apps. I then compared the experience with Samsung’s Galaxy Z Fold 7 runningGoogle’s Gemini. The testing evolved as I went. Early software builds struggled to upload files, analyze images, and connect to Vertu’s concierge service. After I reported these issues to Vertu, the company rolled out server-side fixes that restored the missing functionality, allowing the remaining tests to be completed. What emerged over days of testing was a more nuanced picture than the company’s claims might suggest. Hermes impressed when analyzing local files and spreadsheets, areas where Gemini on Samsung’s foldable still relied on manually uploaded documents during my testing. It was also more willing to automate actions across apps and complete multi-step workflows. But that greater autonomy came with trade-offs, raising questions about when an AI should act independently and when it should ask for clarification. One of the first tests simulated a common executive scenario before leaving for the airport. I asked Hermes Agent on the Alphafold to message a contact that I was running 20 minutes late, navigate to the airport, switch the phone to Do Not Disturb, and remind me to call the hotel in 15 minutes. The agent sent the message, enabled Do Not Disturb, and opened Google Maps with directions to the airport. It did not, however, automatically begin navigation and instead set the reminder for 9:08 p.m., despite the request being made at 2:32 a.m. for a reminder 15 minutes later. Running the same request on Samsung’s Galaxy Z Fold 7 produced a different experience. Rather than attempting every action immediately, Gemini asked follow-up questions, including which airport I wanted to travel to and whether the reminder should be created in Google Tasks or Samsung Reminder. Once I made those selections, it created the reminder for the correct time. Hermes was more willing to act autonomously, while Gemini preferred to confirm details before proceeding. As a result, Hermes completed more of the requested workflow, but Gemini ultimately produced the more accurate outcome. A second test focused on a more open-ended task. I asked Vertu’s Hermes Agent to organize a business trip from Mumbai to Pune, including a morning flight, a hotel recommendation, and adding the itinerary to my calendar. The agent responded that there were no direct morning flights available for the requested journey and offered a Contact Butler button to escalate the request to Vertu’s concierge service. It also created a calendar entry for the wrong dates, scheduling the trip for 7 July instead of 18–19 July, leaving the workflow incomplete. Gemini on Samsung’s Galaxy Z Fold 7 took a different approach. After determining that no suitable direct morning flight was available for the requested journey, it continued planning the trip by suggesting alternative travel options rather than handing the task off. Business documents revealed a mixed picture, too. I asked both Hermes Agent and Gemini to analyze a locally saved financial spreadsheet, summarize the quarterly results, and determine whether third-quarter sales figures were included. During my original testing, Hermes analyzed an uploaded sales spreadsheet and correctly summarised the Q2 figures. However, when I returned to the same conversation days later, it no longer recognized the previously shared document, instead responding: “I cannot access files stored directly on your local device. Please upload or attach the Sales spreadsheet here in the chat, and I will gladly analyze the Q2 data for you.” Gemini also required the spreadsheet to be uploaded initially, but retained the context of the conversation. Days later, it was still able to answer follow-up questions about the document, correctly identifying the North region as generating the highest sales without requiring the file to be uploaded again. Taken together, the testing suggested Hermes Agent is an ambitious AI assistant rather than a finished one. Its willingness to act autonomously often made it feel more like an agent than Gemini on Samsung’s phone, but that same approach occasionally produced incomplete workflows, incorrect outputs, and inconsistent behavior. The pace of updates during the review also suggested Vertu is actively refining the platform, meaning today’s experience may not be the same one buyers encounter a few months from now. Beyond general assistance, Vertu has built Hermes around a collection of specialist AI agents aimed at affluent professionals, including agents focused on legal advice and investment insights, along with the option to escalate certain requests to a human concierge. The idea is to position the Alphafold as more than a premium smartphone, instead presenting it as a digital assistant for executives. In practice, however, the specialist agents should be treated as starting points rather than authoritative advisers. They can provide useful summaries and recommendations, their responses remain AI-generated and should be independently verified before being relied upon for legal, financial, or other high-stakes decisions. The option to escalate certain requests to Vertu’s concierge service underscores the current limits of AI agents. Human expertise still matters. Vertu is also positioning the Alphafold as a business platform rather than just a smartphone. The company demonstrated an integrated enterprise resource planning (ERP) system designed to give executives access to business data and workflows from the device. My testing, however, was limited to a demonstration environment, making it difficult to assess how the feature performs in day-to-day use or how well it integrates with existing enterprise systems. For Alphafold’s target audience, security may matter as much as AI. Executives are unlikely to use an assistant that analyses contracts, financial reports, and business plans if they are uncertain where that data is processed or stored. Vertu says conversations with Hermes Agent are encrypted and are not used to train public AI models. According to the company, users can also choose where their data is processed, with enterprise deployments supporting private infrastructure for organizations that require greater control over sensitive information. Vertu backs those claims up with a dedicated “A5” security chip, which it says provides hardware-level protection for sensitive data, encrypted communications, and digital credentials. Those claims couldn’t be independently verified during testing, but they’re central to Vertu’s pitch to executives and enterprises. Away from AI, the Alphafold behaves much like any modern flagship foldable. The battery comfortably lasted more than a day during testing. However, the absence of wireless charging is a surprising omission at this price, particularly when Samsung’s Galaxy Z Fold 7 supports convenient Qi charging alongside wired USB-C charging. The camera app also includes a document scanning mode under a “Smart AI” setting that can recognize paperwork and save them with enhancements, making it useful for digitizing contracts, receipts, and other business documents. Samsung offers a comparable scanning experience through its own camera software, so this is feels more like a parity feature than a differentiator. The Alphafold is an ambitious attempt to build an AI-first luxury smartphone, but the execution falls short of its price tag. Despite its premium materials and exclusive services, the core hardware offers little that cannot be found in significantly less expensive foldables, while Hermes Agent remains an evolving platform rather than a compelling reason to spend thousands more. Ultimately, Vertu is asking buyers to pay a substantial premium for branding, craftsmanship, and an ecosystem of AI and concierge services built on top of an established smartphone platform. Based on my testing, that premium is difficult to justify, particularly when Samsung’s Galaxy Z Fold 7 offers a more mature foldable experience with comparable day-to-day functionality at a fraction of the price. With Samsung’s next-generationGalaxy Z Fold 8 expected very soon, the Alphafold’s value proposition becomes even harder to defend.
View
Apple’s lawsuit couldn’t come at a worse time for OpenAI
Applefiled a trade secrets lawsuit against OpenAIlast Friday, and it’s not messing around. The complaint alleges a pattern of misconduct reaching all the way up to OpenAI’s chief hardware officer and claims more than 400 former Apple employees now work at the company.OpenAI’s response so farhas been carefully hedged, and the timing couldn’t be worse with the company reportedly eyeing an IPO as early as later this year. On this episode of TechCrunch’sEquitypodcast, hosts Kirsten Korosec, Anthony Ha, and Sean O’Kane dig into what the lawsuit could mean for OpenAI’s own hardware ambitions and IPO timeline, plus a bigger theme running through the week’s news: how much should anyone trust AI companies with their data? Listen to the full episode to hear more about: Subscribe to Equity onYouTube,Apple Podcasts,Overcast,Spotifyand all the casts. You also can follow Equity onXandThreads, at @EquityPod.
View
Patreon stops asking AI bots not to scrape — and starts blocking them
Patreon, the membership platform for creators, is cracking down on AI scraping its content for training purposes. On Thursday, the company shared that it’s working with internet infrastructure provider Cloudflare to directly block access to AI bots designed to train their AI models on creators’ work without permission. The strengthened measures were necessary because AI scraping has become more sophisticated since it first put measures in place to deter AI crawlers in 2023, the company says. In addition, Patreon’s paywall has long locked much of creators’ content out of reach of crawlers. But more recently, the company introduced new discovery tools like a redesigned Home Feed and itstweet-like Quips, which could expose more content to crawlers. The changes come about as more online publishers and content creators are coming to grips with how AI is ingesting their work for the purpose of making their AI models smarter. To combat this, Cloudflare now offers tools that allow website publishersto restrict AI bots, including amarketplace that lets websites charge AI bots for scraping, dubbed Pay Per Crawl. Earlier this month,it changed its policiesso that “mixed-use” crawlers, meaning those that both index and train on a website’s content, are blocked by default on any pages that host ads. Patreon says that it’s extending its existing work with Cloudflare to use the company’s AI Crawl Control technology to update its AI policies and enforcement tools. The difference here is that instead of simply asking AI crawlers not to scrape content using the robots.txt files — a standard way to provide bots with instructions on how they can use its site — Patreon is now actively blocking AI training bots. “Consent shouldn’t depend on whether a scraper chooses to behave,” aPatreon blog postexplains, referencing the stricter measures. When testing the features, individual AI training crawlers’ weekly attempts to access Patreon went from “thousands of attempts to zero,” the post noted. That indicates that the AI scrapers were ignoring Patreon’s robots.txt file and scraping the site anyway, despite its requests. However, the company said that it will allow bots that index pages and organize information that can be used to send users back to Patreon. “As AI agents become increasingly powerful and popular, creators deserve a meaningful say in how their work is used by AI companies,” remarked Patreon’s product chief Drew Rowny in the announcement. “On most of the Internet, creators have to accept AI training on their work just to reach and grow an audience. Patreon has a different vision: creators should be able to grow their audience and control how their work is used.”
View
How Apple’s big lawsuit could disrupt OpenAI’s IPO plans
Loading the player… Applefiled a trade secrets lawsuit against OpenAIlast Friday, and it’s not messing around. The complaint alleges a pattern of misconduct reaching all the way up to OpenAI’s chief hardware officer and claims more than 400 former Apple employees now work at the company.OpenAI’s response so farhas been carefully hedged, and the timing couldn’t be worse with the company reportedly eyeing an IPO as early as later this year. On this episode of TechCrunch’sEquitypodcast, hosts Kirsten Korosec, Anthony Ha, and Sean O’Kane dig into what the lawsuit could mean for OpenAI’s own hardware ambitions and IPO timeline, plus a bigger theme running through the week’s news: how much should anyone trust AI companies with their data? Subscribe to Equity onYouTube,Apple Podcasts,Overcast,Spotifyand all the casts. You also can follow Equity onXandThreads, at @EquityPod.
View
Why the first GPU financiers are turning to inference chips in a $400 million deal
General Compute, an AI inference cloud startup, has landed a $400 million loan from Upper90, a tech investment firm. It might be the first deal to put up inference-specific chips as collateral — chips built to run already trained AI models quickly and efficiently, rather than the more expensive chips used to build the models in the first place. The financing is the latest signal that markets are responding to concerns over the price of AI tools and tokens by turning to infrastructure that runs open source models more cheaply than the newest LLMs from frontier labs. Founded by CEO Finn Puklowski and CTO Jason Goodison, General Computeraised a $15 millionseed round in May to build an inference neocloud around silicon from SambaNova, an Intel-backed chipmaker. (Neoclouds are purpose-built for AI workloads, unlike the general-purpose infrastructure offered by traditional hyperscalers like AWS or Azure.) The company’s SN50 chips are designed for inference. They’re power-efficient and don’t require expensive water-cooling systems, which means they can be deployed more quickly than GPUs across a larger variety of data centers. General Compute says the new chips will provide 16 times faster inference than GPU-based clouds. The challenge is getting a lot of these chips, especially when you’re a brand-new company. Upper90 co-founder and CEO Billy Libby, a former Goldman Sachs quantitative trader, had a playbook for this: In 2021, his firm financed GPU purchases by Crusoe, the energy-focused data center startup, which he believes was the first loan against the value of advanced chips. Traditional lenders eschewed such deals at the time because of the risks and uncertainties around GPU depreciation. But as CoreWeave made chips-backed loans into a business model and then the basis of a blockbuster IPO, this kind of financing has become common. “When we financed Nvidia GPUs as the first group to do that, the market was inefficient,” Libby told TechCrunch. “We could really put together something as an early participant, and kind of get compensated for the risk.” Now that GPUs are comparatively well understood andperhaps over-bought, Upper90 is turning to companies like General Compute to ride the next wave of the AI boom. “We think open source models are going to be important, and we went and looked for a player last year that was in inference,” Libby said. “Everyone doesn’t need a supercomputer, but they do need inference and AI.” That thesis has been growing stronger, with companies that provide access to open models, like OpenRouter and Fireworks, raising new rounds at huge valuations. New models likeKimi’s K3have proven to compete with the latest releases from Anthropic and OpenAI on coding benchmarks. And new chipmakers like Groq and Cerebras have drawn interest from acquirers and public markets alike. General Compute’s ability to access chips outside of Nvidia’s ecosystem matters for the same reason. TensorWave, another AI infrastructure company, is making a similar bet on a partnership with AMD. As more alternatives to Nvidia emerge, compute providers that aren’t locked into Nvidia deals may have an advantage in providing cost-efficient inference. “There are a bunch of chips that are starting to scale that have amazing [total cost of ownership], or that can operate much faster than Nvidia, but there’s not too many buyers for them,” Puklowski said. “By getting together with Upper90, this is not just, ‘a cool startup got some money to buy some compute.’ Like, this is the first signal of capital organizing itself and the fragmenting of Nvidia’s monopolistic dominance.”
View
Are GCCs Hitting Pause on Global Talent Moves After EY Tax Ruling?
Companies are reassessing overseas secondments as the industry seeks clarity from the government on tax treatment and compliance risks.
View
1Password Lets Claude AI Access Your Accounts Without Revealing Passwords
1Password has launched a new integration for Anthropic's Claude AI that allows the chatbot to sign in to websites using credentials stored in a user's 1Password vault without exposing passwords or one-time passcodes. The feature, called 1Password for Claude, is designed for browser-based AI tasks that require access to accounts. Alongside the integration, the company has also introduced Agentic Mode, a browser extension feature that limits AI agents to using only the credentials approved by users for a specific task.
View
The End of AI Subsidies is Forcing Enterprises to Rethink AI Agent Economics
Progress Software’s Chief AI Officer Ed Keisling says enterprises are reassessing AI deployments as discounted frontier models fade, exposing the true cost of running AI agents and accelerating a shift toward smaller, cheaper models.
View
Wipro Says AI is the Future, But Legacy Drag Keeps Growth and Margins in Check
CEO Srini Pallia said AI is creating new spending opportunities even as it compresses traditional IT budgets.
View
Govt to Match VC Funding for Chip Startups Under ₹1.27 Lakh Crore Semicon 2.0
The centre will exit its investments as companies mature, allowing founders to retain long-term control.
View
TCS, Google Cloud Open Gemini AI Centre in Kolkata
The new facility is TCS’ eighth Gemini Experience Center globally and third in India, as the IT major expands its AI co-innovation network with Google Cloud.
View
