Latest AI News
IT, IITs, and GCCs: The Trifecta Fuelling India’s Next AI Innovations
The partnerships between IITs and GCCs focus on accelerating emerging AI and robotics innovations from academic research into real-world enterprise deployments.
View
Zhipu AI’s GLM-5.1 Becomes Top Model on SWE-Bench Pro, Beats GPT-5.4, Claude Opus 4.6
The Chinese AI model was only trained on 100,000 Huawei Ascend without relying on NVIDIA chips.
View
Intel Shares Rise by 4% After Joining Tesla’s Terafab Chip Venture
The partnership with Tesla, SpaceX, and xAI positions Intel at the centre of a $25 billion semiconductor project.
View
AWS Summit Bengaluru Returns with a Focus on AI, Data & Cloud
The two-day event will feature keynote addresses, breakout sessions, hands-on labs, and an expo floor showcasing AWS technologies and partner solutions.
View
The New Teddy Bear: Children Love AI Toy Robots, But Do They Have a Hidden Cost?
AI robots are transforming childhood experiences, but psychologists and policy experts caution against overdependence and data risks.
View
Anthropic Unveils Claude Mythos, Says Model Too Dangerous to Launch
Anthropic said the model has already identified “thousands of high-severity vulnerabilities,” including flaws in major operating systems and web browsers.
View
Uber to Use AWS Chips for Trip Processing and AI Training
Uber said its AI models are trained on data from billions of trips and deliveries.
View
I can’t help rooting for tiny open source AI model maker Arcee
Arcee, a tiny 26-person U.S. startup that built a massive, 400B-parameter open source LLM ona $20 million shoestring budget, has released its new reasoning model. Arcee calls the model Trinity Large Thinking — and it’s the most capable open-weight model “ever released by a non-Chinese company,” claims CEO Mark McQuade to TechCrunch. As that comment implies, Arcee has a goal that I can’t help but root for: It wants to give U.S. and Western companies a model that gives them no reason to use a Chinese-based one. While Chinese models are extremely capable, they areperceived as risky, putting power, and perhaps data, into the hands of a government that doesn’t share all of the Western world’s ideals. With Arcee, companies can download the model, train it to their own needs, and use it on premises. Companies can also use Arcee’s cloud-hosted version, accessible via API. While Arcee’s models are not outperforming the closed source models from the big labs like Anthropic or OpenAI, they’re not being held hostage by the whims of those giants, either. For instance, Claude, with its exceptional abilities to code, has been a popular choice for users of open source AI agent tool OpenClaw. But Anthropic pulled the rug out from them last week when it told users that their Anthropic subscriptionswill no longer cover OpenClaw usage— they will have to pay additionally for that. (In February, OpenClaw creator Peter Steinberger said he wasjoining Anthropic’s biggest rival, OpenAI.) In contrast, McQuade proudly points todata from OpenRouterthat says it has become one of the top models used with OpenClaw. So, how good is Trinity Large Thinking? It is comparable to some of the other top open source models, according to the benchmark results it shared with TechCrunch. As we previously reported, it is not a head-to-head threat to the big cheese among U.S.-built open models: Meta’s Llama 4. But it also doesn’t have the odd, not-reallyopen source license issuesof Meta’s model. All of Arcee’s Trinity models are released under the gold standard for OS licenses, Apache 2.0. Just to be clear, there are also countless other U.S. startups offering open source models and, as a fan of the ingenuity of startups, I’m rooting for them, too.
View
Uber is the latest to be won over by Amazon’s AI chips
On Tuesday, Amazonannouncedthat Uber was expanding its contract for AWS cloud services to run more of its ride-sharing features on Amazon’s chips. Uber will particularly expand its use of AWS’s Graviton (a low-power, ARM-based server CPU) and start a new trial testing Trainium3, AWS’s Nvidia competitor AI chip. This deal is a bit less about a long-term threat to Nvidia than it is a thorough thumbing of the nose by Amazon at AWS’s cloud competitors, Google and Oracle. While Uber historically ran its own data centers, back in 2023, the ride-hailing company famously signed giant, multi-year cloudcomputing deals with Oracle and Google. The idea was to move the majority of its IT infrastructure off its own data centers and onto these two clouds, it said. Even in December, Uber publicly reiterated that goal, writing ina blog post: In February 2023, Uber began transitioning from on-premise data centers to the cloud using OCI and Google Cloud Platform, taking on the dual challenge of shifting massive workloads and introducing Arm-powered compute instances into a previously x86-dominated environment. Uber particularly called out in that post the use of the ARM chips made by Ampere in Oracle’s cloud. This is where things get interesting. If you want a crash course in how inter-tangled Silicon Valley can be,take a look at the history of Ampere. Ampere was founded by former Intel bigwig Renee James after she was not promoted to CEO at the chipmaker. She pulled all her strings, including her power at her then-job as an investor at private equity firm Carlyle and her board seat position at Oracle, to raise the cash to start this company. Oracle owned about one-third of the company, and James had to give up her status as an independent Oracle director because of that investment. (James was, by the way, a key board personwho helped vote in Oracle’s $9.3 billion purchase of NetSuitein 2016, a company where Larry Ellison was a major stockholder. That dealsparked an unsuccessful shareholder lawsuitalleging Oracle overpaid for it.) In December, Ampere’s major competitor SoftBank acquired it, and Oracle sold its stake for ahandsome $2.7 billionpre-tax gain. James left Oracle’s board at the end of 2024 and is no longer working at Ampere. Oracle israising money as fast as it can to build data centersfor OpenAI and Stargate.Ellison said Oracle sold Amperebecause he believed designing chips in-house for its data centers was no longer a competitive advantage. It prefers to buy the chips and has signed massive deals with Nvidia. It’s worth noting that Oracle, SoftBank, and Nvidia are also part of OpenAI’sorbit of circular dealsthat are supposed to fund the model maker’s massive data center build-out. But now AWS is announcing it has nabbed a bigger contract from one of Oracle’s star customers, Uber, because it has in-house-designed chips. Uber joins Anthropic, OpenAI, and Apple as Big Tech companies that have signed on or increased their usage of AWS because of these AI chips. In December, Amazon CEO Andy Jassy saidTrainium was already a multibillion-dollar business. (Fora look at the team and lab that design these chips, check out our exclusive tour of the facility.)
View
Anthropic debuts preview of powerful new AI model Mythos in new cybersecurity initiative
Anthropic on Tuesday released a preview of its new frontier model, Mythos, which it says will be used by a small coterie of partner organizations for cybersecurity work. In apreviously leaked memo, the AI startup called the model one of its “most powerful” yet. The model’s limited debut is part of a new security initiative, dubbed Project Glasswing, in which 12 partner organizations will deploy the model for the purposes of “defensive security work” and to secure critical software, Anthropic said. While it was not specifically trained for cybersecurity work, the model will be used to scan both first-party and open source software systems for code vulnerabilities, the company said. Anthropic claims that, over the past few weeks, Mythos identified “thousands of zero-day vulnerabilities, many of them critical.” Many of the vulnerabilities are one to two decades old, the company added. Mythos is a general-purpose model for Anthropic’s Claude AI systems that the company claims has strong agentic coding and reasoning skills. Anthropic’s frontier models are considered its mostsophisticated and high-performance models, designed for more complex tasks, including agent-building and coding. The partner organizations previewing Mythos as part of Project Glasswing include Amazon, Apple, Broadcom, Cisco, CrowdStrike, the Linux Foundation, Microsoft, and Palo Alto Networks. As part of the initiative, these partners will ultimately share what they’ve learned from using the model so that the rest of the tech industry can benefit from it. The preview is not going to be made generally available, Anthropic said, though 40 organizations will gain access to the Mythos preview aside from the partnership. Anthropic also claims that it has engaged in “ongoing discussions” with federal officials about the use of Mythos, although one would have to imagine that those discussions are complicated by the fact that Anthropic and the Trump administration are currently lockedin a legal battleafter the Pentagon labeled the AI lab a supply-chain risk overAnthropic’s refusalto allow autonomous targeting or surveillance of U.S. citizens. News of Mythos was originally leaked in a data security incidentreported last month by Fortune. A draft blog about the model (then called “Capybara”) was left in an unsecured cache of documents available on a publicly inspectable data lake. The leak, which Anthropic subsequently attributed to “human error,” was originally spotted by security researchers. “‘Capybara’ is a new name for a new tier of model: larger and more intelligent than our Opus models — which were, until now, our most powerful,” the leaked document said, adding later that it was “by far the most powerful AI model we’ve ever developed,” according to the report. In the leak, Anthropic claimed that its new model far exceeded performance areas (like “software coding, academic reasoning, and cybersecurity”) met by its currently public models and that it could potentially pose a cybersecurity threat if weaponized by bad actors to find bugs and exploit them (rather than fix them, which is how Mythos will be deployed). Last month, the companyaccidentally exposednearly 2,000 source code files and over half a million lines of code via a mistake it made in the launch of version 2.1.88 of its Claude Code software package. The company thenaccidentally causedthousands of code repositories on GitHub to be taken down as it attempted to clean up the mess. Correction April 7, 2026: An earlier version of this article erroneously stated how many partners are working with Anthropic on Project Glasswing. There are 12 partner organizations, though 40 organizations total will have access to the Mythos preview.
View
Intel signs on to Elon Musk’s Terafab chips project
Intel will join SpaceX and Tesla in an effort to build a new U.S. semiconductor factory in Texas, although the scope of its contributions are unclear. “Our ability to design, fabricate, and package ultra-high-performance chips at scale will help accelerate Terafab’s aim to produce 1 TW/year of compute to power future advances in AI and robotics,” Intelsaidin a corporate post on X. Intel hasn’t shared any more information. Elon Muskannouncedin March a team-up between the two tech companies he leads to develop chips for AI compute, satellites, and SpaceX’s mooted space data center and to support the possibility of autonomous Tesla vehicles and robots. However, building a chip fab is one of the most difficult and expensive corporate infrastructure projects out there, typically requiring years of time and more than $20 billion to create a facility with a huge clean room for thousands of ultra-precise machines to carve silicon. It wasn’t obvious how SpaceX and Tesla, two companies with no experience in the sector, could team up to execute the project efficiently. Now we have a better idea: Intel will do it. The company has been hunting for large anchor customers to support its foundry business, and now it has two. Still, if investors thought that Terafab would be a greenfield approach based on SpaceX’s and Tesla’s unique approach to engineering, that may not play out. Once the leading U.S. silicon producer, Intel has seen rivals Nvidia and AMD take the lead in developing advanced processors and adopt the “fabless” business model where chip designers outsource the manufacturing of their semiconductors. Intel’s stock rose more than 3% on the news today. It was trading at $52.28, about 2.9% higher than its opening bell price, at 2 p.m. ET. Intel declined to comment on the partnership, while SpaceX didn’t respond to TechCrunch’s query.
View
Firmus, the ‘Southgate’ AI data center builder backed by Nvidia, hits $5.5B valuation
Asia AI data center provider Firmus on Mondayannounceda fresh $505 million raise led by Coatue at a $5.5 billion post-money valuation. With this round, Firmus has raised $1.35 billion in six months, it says. The Singapore-based data center company previously raisedAU$330 million(approximately $215 million) at anAU$1.85 billion($1.2 billion) valuation from investors, including Nvidia. Firmus is developing an energy-efficient “AI factory” network of data centers in Australia and Tasmania, a project it dubs Project Southgate. It is using Nvidia’s reference designs for building these efficient data centers. These new data centers will use Nvidia’s Vera Rubin platform — the chip giant’s next-gen AI computing system succeeding its Blackwell architecture, expected to ship in the second half of 2026. Firmus originally provided cooling technologies for Bitcoin mining and hasbecome yet anothercrypto-roots-turned-AI providercompany that investors love.
View