Welcome to the newest edition of ID Tech’s AI Update. Here’s the latest big news on the shifting landscape of AI and identity technology:
China has built a prototype extreme ultraviolet (EUV) lithography machine in a high-security Shenzhen facility, according to a new Reuters report. The aim is to replicate the advanced chip manufacturing tech of ASML, which is uniquely positioned right now as the only EUV lithography specialist in the world. Indeed, the China project – which began in secrecy in 2019 – has recruited former ASML engineers. Sources told Reuters a Shenzhen team completed a working EUV prototype in early 2025. It can’t yet make chips at scale, but it is “successfully generating” EUV light, which is a non-trivial subsystem achievement given the complexity of stable EUV sources and contamination control.
OpenAI is reportedly on track to pass $13 billion in revenues for 2025, up from about $4 billion a year ago. It has also reportedly held preliminary talks for a funding round that could raise $100 billion, and value OpenAI at about $750 billion. This all comes amid speculation that OpenAI may be aiming for an IPO in 2026 that could value it at $1 trillion. And the startup is serving a considerable chunk of the planet, asserting that ChatGPT has over 800 million users a week.
Amazon could be one of OpenAI’s next major investors, with reports that the tech giant is in “very fluid” early-stage talks about contributing at least $10 billion. They already have an infrastructure partnership after OpenAI committed to a $38 billion, seven-year compute arrangement with Amazon Web Services last year.
Meta is reportedly developing an image and video generation AI model dubbed “Mango”. And it’s worth noting some context: Meta has already launched pretty sophisticated visual and audio models for things like segmenting people in videos and isolating a person’s speech in recordings. Alongside Mango, Meta has a project called “Avocado” that is focused on coding as well as work toward “world models” in which AI can learn about environments through visual inputs. Meta is reportedly aiming to launch both AI models in the first half of 2026.
The founder of Figure AI has launched a mysterious AI lab side project called Hark, with $100 million of his own money. Figure is known for being a pioneer in AI-powered robotics, and particularly humanoid robots. Hark, meanwhile, will reportedly focus on developing “human-centric” AI that can “think proactively, recursively, improve, and care deeply about people,” according to an internal memo. The serial founder, Brett Adcock, will serve as CEO for both firms.
Oracle shares fell steeply after reports emerged that it had lost a key financial backer for a major AI-focused data center project in Michigan. Blue Owl had previously been seen as Oracle’s primary financing partner for this kind of thing, but it backed out of a plan to finance the roughly $10 billion project, reportedly because it was getting less favourable terms than it had in previous deals. The fallout briefly produced a worrying signal about the precariousness of financing for major AI projects, but Oracle has insisted that the project is expected to move ahead without Blue Owl, and its stock price closed the week up 2.27 percent.
Manus says it has passed $100 million in annual recurring revenue (ARR) about eight months after launch, and that its total revenue run rate is now over $125 million (including usage-based and other revenue). It’s an encouraging sign for the AI agent space, with Manus being a pioneer in developing agentic AI platforms that can perform complex tasks over a long context window. Manus says the latest version of its system can build “production-ready” web apps end-to-end, including backends, databases, and authentication mechanisms.
The chatbot’s take: It really is incredible technology.
–
By Alex Perala
