Latest AI News
WhatsApp Introduces Incognito Chat With Meta AI for Private Conversations
WhatsApp has announced Incognito Chat with Meta AI, a new feature that lets users have private, temporary conversations with the company's AI assistant. The feature is aimed at people who want to ask sensitive questions involving personal matters, finances, health, or work without keeping a lasting record of the interaction. Incognito Chat is built on Meta's Private Processing technology and will begin rolling out to WhatsApp and the Meta AI app over the next few months.
View
Poppy debuts a proactive AI assistant to help organize your digital life
Smartphones can be distracting with their dizzying array of apps and constant stream of notifications. A new app calledPoppyaims to organize the chaos by combining your calendar, email, messages and other sources into a single dashboard. The idea, per the company’s website, is that “Poppy pays attention so you don’t have to.” Users can connect various services to Poppy’s app, like their email, calendar, and, at a minimum, their location. Poppy then uses that data along with AI to guess what’s important to you right now based on what’s going on in your life. At a high level, this means you can open Poppy’s app or glance at its widgets to see the meetings or tasks you have on your plate. But Poppy’s most powerful feature is likely its proactive suggestions. For instance, if Poppy has access to your calendar and sees that you have a 30-minute gap while you’re near a park, it could suggest you take a break and go for a walk before your next appointment. And if you’re planning a brunch with a friend who mentioned their food preferences in a previous communication, it could factor in that information when suggesting restaurants. You can also message Poppy with questions or requests, almost as if you had a personal assistant working on your behalf. Poppy can track your flights and alert you to changes, or nudge you when it’s time to take your medication. Poppy’s maker,Sai Kambampati, says he’s always been fascinated by human-computer interaction, having earned his Master’s degree in Computer Science with a specialization in this area. Previously a software engineer at the AI hardware startup Humane, he said he has seen first-hand how people are trying to rethink how we engage with technology. "I've always been interested in challenging what computers are able to do, especially the idea of ambient computing and computers that can proactively sense what you need and anticipate your needs," Kambampati told TechCrunch. "That's something that I found very, very exciting. And I felt like with all the AI technology that we're seeing around us, it has never been more possible to embark on something like this." At launch, Poppy works with everyday apps like Apple Calendar, Google Calendar, Gmail, Outlook, iCloud Mail, Apple Health, Reminders, Contacts, iMessage, WhatsApp, and others. (It uses a Mac app to access iMessage, which could later be a problem as Apple generally doesn't allow third-party apps to access its messaging service.) It also works with apps like Uber and Instacart, and Kambampati plans to extend support to others over time. The company says users' data is encrypted when stored in its database, and it has a zero-retention policy enabled when it uses cloud-based LLMs for its suggestions. In time, however, Kambampati would like make the switch to using local, on-device AI models when technology advances. "My hope, my dream is — within two to three years from now, when our devices have much more powerful compute, and the models get much smaller, cheaper and more high quality — eventually we can have all of this running on our own devices, and there won't even be a need to hit the servers," he says. Poppy's San Francisco-based team of four is backed by $1.25 million in pre-seed funding led by Kindred Ventures, with various angels also participating, including DeepMind's Logan Kilpatrick.
View
Anthropic now has more business customers than OpenAI, according to Ramp data
For the first time, Anthropic has more verified business customers than OpenAI, according tothis month’s AI Indexfrom the fintech firm Ramp. The survey, compiled from Ramp’s clients’ expense data, shows 34.4% of participating businesses are paying for Anthropic services, more than any other AI lab, while only 32.3% pay for OpenAI. It is the first time Anthropic has held the top position. “Anthropic has already been in the lead amongst the high adoption groups like finance, tech, professional services,” Ramp economist Ara Kharazian told TechCrunch. “It’s across the other firms where OpenAI still has a lead, but that has been shrinking over the past couple of months.” Because the index only represents companies that use Ramp, it’s not a perfect proxy for the marketplace at large. Still, the sample includes more than 50,000 companies, making it both broad and diverse enough to carry weight. More importantly, the general trend can be seen across the industry. OnOpenRouter’s leaderboard, which samples a different portion of users, OpenAI last ranked above Anthropic in December 2025. According to Ramp’s figures, the past 12 months have been particularly transformative for Anthropic. In May 2025, a mere 9% of businesses were paying for Anthropic products, a figure that climbed 26% in the following 12 months. Over the same period, OpenAI’s share declined by 1%, and the overall share of businesses using some kind of AI product increased by 9%. Kharazian is skeptical about whether this advantage will last, for reasons he explained in a blog post, but said the success of the past year was proof that Anthropic had chosen a good strategy. “What Anthropic did worked really well,” Kharazian told TechCrunch, “which was — start with a very technical customer base, focus on their needs, really succeed in execution and then start broadening out through tools like Cowork.”
View
WhatsApp adds an incognito mode in Meta AI chats
Meta on Wednesday said it is adding the ability to start “incognito” conversations with its Meta AI chatbot within WhatsApp. These conversations, the company said, will be processed in a secure environment and can’t be seen by anyone. Users can start an incognito session by tapping on a new icon in one-on-one chats with Meta AI. The company said the feature will also be available on the standalone Meta AI app as well. Incognito chats will roll out to WhatsApp and the Meta AI app over the next few months. Loading the player… Meta said these incognito conversations are not saved, and messages will disappear by default once you close the chat. The session will also end if you close the app or lock your phone, and Meta AI will lose the context of that particular conversation, the company said. “People are starting to use AI for everything, including some of their most private thoughts, whether that’s tackling financial or health questions, or for advice on how to respond to a tricky message from a friend or a colleague. We think it’s really important to give people the ability to ask these questions as privately as possible,” Alice Newton-Rex, VP of Product at WhatsApp, told TechCrunch over a call. The company has been laying down the groundwork for secure AI chats on WhatsApp for a while now. Last year, it detailed its private processing infrastructure that would let it build AI features withoutbreaking end-to-end encryption. Since then, WhatsApp has added features likeAI-powered summaries of messagesthat use this architecture. Newton-Rex said Meta used smaller models to power its previous features, but the new incognito chat uses itslatest Muse Spark model, which was released last month. The company is already working on its next feature that taps its private processing infra. Called Side Chat, it will let users invoke Meta AI within chats to ask questions and get answers privately without notifying or showing it to other people in the chat. Currently, you need to tag a message and ask a question to the AI assistant to get an answer that other participants in the chat can see. If you privately need to ask a question, you have to paste the text in a separate chat window. ChatGPT and Claude offer incognito modes, too, and companies likeDuckDuckGoandProtonhave launched their own privacy-first chatbots. Meta's move towards private AI chats comes at a key time. Last month, Reuters cited lawyers who opined that users'conversations with an AI chatbot could be used against themin litigation.
View
Introducing the 6 stages at TechCrunch Disrupt 2026 — built for today’s tougher startup market
The biggest risk for founders and investors right now isn’t moving too slowly. It’s reacting too late to where the market already shifted. The new stages atTechCrunch Disrupt 2026help founders and investors make faster, more informed decisions in today’s highly complex, volatile markets. FromOctober 13–15 at Moscone West in San Francisco, Disrupt brings together 10,000+ founders, investors, and operators for 250+ sessionsacross six stagesfocused on the operational pressures reshaping startup innovation, from AI-native competition and infrastructure bottlenecks to changing venture dynamics and enterprise adoption. Explore the six stages planned for 2026 andsecure your pass before prices increase. Right now, save up to $410 with Early Bird rates, plus 50% on a second ticket. Without further ado, get to know thesix stages at Disrupt, designed for a hands-on approach to launching, building, and selling in today’s tech industry. TheDisrupt Stageremains the center of gravity forTechCrunch Disrupt,bringing together headline founders, major technology leaders, and top-tier investors to discuss the broader shifts reshaping the market. Startup Battlefield 200also takes place here, giving attendees direct visibility into which startups investors and media believe have breakout potential before the rest of the market catches up. Think your startup has what it takes to compete?Nominate and apply by May 29. Conversations focus on: For founders, investors, and operators alike, it brings to light the signals shaping opportunity: where attention is concentrating, which categories are accelerating, and how successful companies are positioning themselves in a much tougher market. For a limited time, bring your community to Disrupt and save up to 30% on tickets.Register here. TheBuilders StageatDisruptfocuses on the operational realities of building a company right now: fundraising, hiring, product-market fit, go-to-market execution, and scaling in a more demanding environment. Unlike traditional founder content, these sessions are built around current pressure points having a real impact. Sessions like “How to Win When You’re Not Building AI” tackle one of the defining challenges in today’s market: how non-AI startups compete for attention and capital while investors chase AI-first companies. Other programming explores: Speakers includeNina Achadjian, partner atIndex Ventures;Rajeev Dham, managing partner atSapphire Ventures;Josh Reeves, CEO and co-founder ofGusto;Grant Lee,CEO and co-founder ofGamma;Robby Stein, head of product atGoogle; andMo Jomaa, partner atCapitalG. For founders trying to move faster with fewer mistakes, this is one of the most tactical environments at Disrupt. Lock in your Disrupt ticketso you can tackle your biggest growth challenges. As fintech markets mature and investor scrutiny increases, startup success hinges on knowing which financial technologies are still creating durable growth and which models are losing momentum. TheSmart Money StageatDisruptfocuses on how financial infrastructure is evolving beyond the hype cycle and toward digital financial systems. Sessions examine where real-time payments are gaining traction, why some embedded finance models struggled, and where founders are still building durable fintech businesses despite tighter investor scrutiny. Programming centers on: These conversations are grounded in what’s actually surviving in a more skeptical market, not speculation. Speakers on this stage includeJack Zhang, founder and CEO ofAirwallex, andLotti Siniscalco, general partner atEmergence Capital. For fintech founders and investors, the value is understanding where capital still sees long-term opportunity and where market enthusiasm is starting to disappear. Register for Disruptand get direct access to the investors, founders, and operators shaping the next generation of financial infrastructure. As AI expansion accelerates, demand for data center capacity, energy, grid connectivity, and industrial systems is increasing just as quickly. TheSmart Systems StageatDisruptfocuses on one of the biggest constraints facing the technology industry: physical infrastructure needed for energy, climate, and industrial systems. This stage focuses on the operational systems that modern software companies increasingly depend on but often overlook. Sessions explore: A couple of leaders in this sector to take center stage includeJeff Lawson, co-founder and CEO ofInertia, andDavid Kirtley, CEO ofHelion. For founders and investors building in energy, robotics, logistics, infrastructure, or climate tech, this stage offers a clearer understanding of where physical-world constraints could create the next major opportunities — or new barriers. Get your Disrupt ticketto see where infrastructure innovations are becoming the next competitive advantage. Once AI enters physical systems, reliability becomes a business issue, not just a technical one. TheAI in the Real World StageatDisrupttackles what happens when AI systems move beyond demos and into environments where reliability matters, from robotics and autonomous systems to manufacturing and drug discovery. Programming explores how AI is being deployed across: The focus on this stage shifts away from AI hype and toward operational reality, including how trustworthy systems are built, what happens when cloud access is limited, how physical AI products scale, and where deployment failures create real financial and operational risk. For founders and investors evaluating the next generation of AI companies, this stage offers a clearer picture of which businesses can actually survive the transition from prototype to production. Explore the Disrupt ticket optionsfor founders, investors, operators, and startup teams. TheAI StageatDisrupt,presented byGoogle Cloud, focuses on how generative AI and AI agents are changing software companies at every level. Programming on this stage, like “Rewriting SaaS: Why AI Breaks the Old Business Model,” reflects a growing reality across the software industry: Traditional SaaS advantages are eroding quickly as AI changes user expectations and product economics. Sessions explore: For founders and operators, the value is understanding how software companies are adapting right now — and where competitive advantages are likely to disappear next. Find your ticket matchand experience Disrupt in person before prices increase. Three jam-packed days featuring200+ sessions across six stages, led by250+ tech leadersshaping the industry today, only atTechCrunch Disrupt 2026. Save up to $410 on your pass and get 50% off a second ticket so you can bring a plus-one and experience Disrupt together.Register now.
View
Amazon launches an AI shopping assistant for the search bar, powered by Alexa+
Whether you like itor not, Amazon continues to put AI at the center of the shopping journey. The companyannouncedWednesday “Alexa for Shopping,” its new personalized AI shopping assistant, powered by Alexa+. Notably, the experience will replaceRufus,its generative AI shopping assistant that launched in 2024. According to the company, Alexa for Shopping is designed to offer a voice- and touch-enabled shopping experience across mobile, desktop, and Echo Show smart displays. While Rufus focused on helping customers discover and compare products, Alexa for Shopping is meant to provide more personalized recommendations and automate the shopping experience across Amazon and other online retailers. Now available to U.S. customers, Alexa for Shopping can answer anything from “What’s a good skincare routine for men?” to “When did I last order AA batteries?” Users can type their question into the main search bar or the dedicated Alexa for Shopping chat window, and Alexa will provide tailored answers and recommendations, and even create custom shopping guides. The company says the assistant understands customers’ habits, preferences, and purchase history to bring “that connected, personalized assistance to how you shop” and to make the assistant “more personal and more helpful over time.” Beyond answering questions, Alexa for Shopping can compare products, track prices, and even schedule recurring orders for essentials like pet food or paper towels. If you want to automatically add something to your cart when it goes on sale, you can just tell Alexa, “Add this sunscreen to my cart if the price drops to $10.” Additionally, the assistant can go beyond Amazon’s marketplace, shopping other online stores and using its “Buy for Me” feature to handle the purchase for you, which could be seen as convenient but also a littlecontroversial, given the growing concern around AI autonomy and privacy. The launch of Alexa for Shopping comes on the heels of Amazon introducing its 30-minute delivery service, “Amazon Now,” in dozens of U.S. cities and anew AI-powered featurethat generates real-time conversational audio responses to customer product questions.
View
Anthropic courts a new kind of customer: small business owners
Anthropic is looking to court smaller companies. To that end, the company announced Wednesday the launch of Claude for Small Business, a new suite of services designed for customers who less resemble Walmart and Starbucks and more resemble the local hardware store or coffee shop. So far, much of the most intensive AI adoption has occurred at the enterprise level. In the recent past,studies have shownthat most companies that scaled AI systems beyond experimental or pilot-level integration tended to be large companies with expansive budgets. This appearsto be changing somewhat, as smaller and midsized businesses are seeing greater adoption. Anthropic’s new bundle of features are designed to serve those new AI converts. They are available via a newly introduced toggle within Claude Cowork, the company’s task-automation platform for business users that can browse the web, manage files, and execute multistep workflows on a user’s behalf. By toggling it on, paying users gain access to a host of automated services, including bookkeeping functions, business insights, and generative tools for ad campaigns. The new suite also includes integrations between Claude Cowork and a number of software products — like QuickBooks, Canva, Docusign, HubSpot, and PayPal. “Small businesses account for 44% of U.S. GDP and employ nearly half the private-sector workforce, but their adoption of AI has lagged behind larger enterprises,” the company said. “Tools and training are rarely tailored to the ways small businesses operate, and as a result their use often stops at the chat window.” For founders and investors, the move signals that the AI platform wars are expanding downmarket and that the next major battleground for user acquisition isn’t the Fortune 500; it’s the36 millionsmall businesses that make up the backbone of the U.S. economy. Anthropic is a little behind its competitor, OpenAI, which launchedEnterprise ChatGPTat the end of 2023, including an integration for smaller teams called ChatGPT Business. Anthropic is planning to aggressively promote its new features with a coast-to-coast promotional tour, starting in Chicago and hitting 10 cities in total. At each stop, the company plans to offer a free AI training workshop that will be available to 100 local small business leaders.
View
Origin Lab raises $8M to help video game companies sell data to world-model builders
As AI begins to interact with the physical world, new types of labs are working to buildworld modelsthat could be used to operate physical robotics or model objects in physical space. Unlike large language models, there isn’t an easy source of data for those models, which has left many labs scrambling to assemble the necessary training sets. Now, one startup is emerging with an unlikely data source: the video game industry. That’s the premise ofOrigin Lab, which just announced an $8 million seed funding round led by Lightspeed Ventures. SV Angel, Eniac, Seven Stars, and FPV also participated, with angel funding from Twitch co-Founder Kevin Lin and Cruise founder Kyle Vogt. “The AI systems that are being built now need to understand how the physical world works and how things move,” co-CEO and co-founder Anne-Margot Rodde told TechCrunch. “That data essentially lives in video games.” In simple terms, Origin Lab will serve as a marketplace where world-model-focused labs such asYann LeCun’s AMI LabsorFei-Fei Li’s World Labscan buy high-quality licensed data. On the other side of the trade, video game companies can squeeze additional revenue out of the digital assets they’ve already created. In the middle, Origin Lab will convert the video game assets into a form that works as training data — something that could be as simple as a rendering run or as complex as automating hours of walkthrough footage. “It became clear that the video game industry was sitting on some incredibly valuable data, but there was no real way or infrastructure to basically connect AI labs and the video game industry,” says Rodde. “So essentially, we built that bridge.” Labs have long been interested in video game footage as a data source, but licensing and data quality issues have often gotten in the way.In December 2024, OpenAI caused a minor scandal when the first version of its Sora video-generation model seemed to regurgitate footage of popular video games and streamers — presumably because it had been trained on Twitch streams. Amazon has been open aboutits interest in using Twitch footageto train models. Origin’s success in fundraising is a sign of a growing market — not just for training data, but for startups that can serve as essential suppliers to major AI labs. Faraz Fatemi, a partner at Lightspeed who led the Origin investment, says the success of companies like Scale.AI has made the opportunity impossible to ignore. “We’ve seen how sharp the revenue scaling can be for data vendors that are serving the major labs,” Fatemi told TechCrunch. “These are very well-capitalized businesses, and the bottleneck for all of them is data.”
View
Who trusts Sam Altman?
In May 2023, OpenAI CEO Sam Altman was sworn in and testifying before Congress about the regulation of artificial intelligence. Senator John Kennedy of Louisiana heard his ideas about licensing advanced models and asked if Altman might be qualified to run a hypothetical AI regulatory agency. “I love my current job,” Altmansaid, to titters. “You make a lot of money, do you?” Kennedy asked him. “No, I’m paid enough for health insurance, I have no equity in OpenAI,” Altman assured him. “You need a lawyer,” Kennedy replied. Now Altman has many lawyers, who watched as their client suffered a withering interrogation, sworn in to a California federal court on Tuesday. They were investigating much the same matter as Kennedy—is Altman qualified to to control the most advanced AI models? “You didn’t disclose to the United States Senate that you had an interest in OpenAI through a share in a Y Combinator fund, did you?” barked Steve Molo, the combative attorney leading Elon Musk’s effort to shut down OpenAI’s for-profit business. Altman had admitted that he did have economic exposure to OpenAI through his LP position in the Y Combinator fund. "I didn't mention it in that testimony, but, again, I think it is well understood of what it means to be a passive owner of many venture funds," Almtan said."You thought Senator Kennedy was a very sophisticated investor when he asked you that question?" Molo replied. Altman's decision to volunteer that he had no equity when he could have simply side-stepped the question was interesting one. It's technically true, but Altman—who emphasized his expertise in investing in early-stage startups—surely understood his economic exposure to OpenAI through Y Combinator, and through investments in other AI companies that worked with OpenAI. Altman's credibility was on trial yesterday, at least in the eyes of the plaintiffs. OpenAI's attorney's maintained that little was done to advance Musk's case, accusing their counterparts of character assassination. But the juroy and Judge Yvonne Gonzalez Rogers are weighing Altman's credibility as a pivotal character in the events they are examining. Molo ran through a litany of people who accused Altman of lying or misleading them while under oath in the courtroom, including former OpenAI board members Helen Toner and Tasha McCauley, Elon Musk and OpenAI cofounder Ilya Sutskever. He also brought up the recentNew Yorker storydetailing concerns about his honesty. The "blip"—when OpenAI's board briefly fired Altman and OpenAI president Greg Brockman for failing to be candid with them—has been a subject of significant discussion at this trial. Then-board members Toner and McCauley testified that Altman had misled them, with McCauley referring to "a toxic culture of lying." "I do have doubts that was the full reason" for his firing, Altman said. Asked again if to acknowledge that the board said he had not been candid with them, Altman replied "they asked me to come back the next morning." The focus on his firing is not just about questioning Altman's credibility. One key question of the trial is whether OpenAI's structure lives up to its mission, and specifically whether the non-profit board can exercise true control over the for-profit. From the point of view of Musk's lawyers, the 2023 episode offers evidence that Altman's influence over the company exceeded that of its board of directors. Witnesses brought by OpenAi and Microsoft have insisted that the current non-profit board does exercise control over the for-profit. Microsoft CEO Satya Nadella called Altman's firing "amateur city." Bret Taylor, who joined OpenAI's board as chair in the wake of Altman's rehiring, said he found nothing that warranted his termination and that Altman has been "forthright with me." Dr. Zeko Kolter, the OpenAI board member focused on AI safety, said no one had interfered with that work since he started in 2024. But Taylor also made clear that the choice to rehire Mr. Altman in 2023 was because his departure would effectively ended OpenAI as a going concern, with most of the employees intent on following him out the door. Now, as the jury and Judge weigh whether the current structure lives up to the organization's mission, they will wonder whether the board can really fire or discipline its CEO. Asked if he would ever fire himself as CEO, Altman said he had no plans to do so. Asked if he could be trusted, he replied "I believe I am an honest and trustworthy business person."
View
Uber to Build First Data Centre in India With Adani Group
The facility is expected to be operational later this year and will support Uber’s global technology operations.
View
Adaption aims big with AutoScientist, an AI tool that helps models train themselves
For years, AI researchers have anticipated the moment when AI systems will be able to improve themselves better than humans could. With investors pouring money into a new generation of research-driven AI labs, there are more resources than ever available to pursue the goal. Now, one of those neolabs has taken a major step towards making it real. On Wednesday,Adaptionintroduced a new product calledAutoScientistthat helps models learn specific capabilities quickly by using an automated approach to conventional fine-tuning. The techniques are applicable to a wide range of fields, but the Adaptation team is particularly focused on the potential for speeding up and easing the process of training and fine-tuning a frontier-level AI model. According to co-founder and CEO Sara Hooker, who previously worked as VP of AI research at Cohere, AutoScientist represents a new way to approach the AI training process. “What’s super exciting about it is that it co-optimizes both the data and the model, and learns the best way to basically learn any capability,” Hooker told TechCrunch. “It suggests we can finally allow for successful frontier AI trainings outside of these labs” AutoScientist builds on the company’s existing data offering,Adaptive Data, which aims to make it easier to build high-quality datasets over time. AutoScientist, meanwhile, is designed to turn those continuously improving datasets into continuously improving AI models. “Our view at Adaption is that the whole stack should be completely adaptable, and should basically optimize on the fly to whatever task you have,” Hooker says. Of course, that approach will only be as good as the results. In its launch materials, Adaption boasts that AutoScientist has more than doubled win-rates across different models — impressive numbers, but difficult to put into context. Since the system is built to adapt models to specific tasks, conventional benchmarks like SWE-Bench or ARC-AGI aren’t applicable. Still, Adaption is confident that users will see the difference once they try AutoScientist out — so confident that the lab is making the tool free to use for the first 30 days after its release. “The same way that code generation unlocked a lot of tasks, this is going to unlock a lot of innovation at the frontier of different fields,” Hooker says.
View
Zoho Commits ₹70 Crore to ONDC to Empower MSMEs with Accessible Sovereign Tech
With this investment, Zoho seeks to support MSMEs in their digital transformation and contribute to India's economic growth.
View