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:
The US government is considering banning DeepSeek from government devices over national security concerns regarding the China-made, open source chatbot. Australia and Italy have already taken such measures. DeepSeek was developed by a Hangzhou-based startup affiliated with the Chinese hedge fund High-Flyer, and rattled markets earlier this year with the launch of a sophisticated AI model that was purportedly developed at a fraction of the cost of similar models.
A U.S. District Judge has denied Elon Musk’s request to block OpenAI’s transition from a nonprofit to a for-profit organization, ruling that he hadn’t demonstrated a strong enough case on the merits for the proposed preliminary injunction. Musk is a co-founder of OpenAI and has suggested that it betrayed its original mission to provide open source AI that benefits humanity. He is also a competitor, via his xAI startup.
The UK’s antitrust regulator has given the all-clear to OpenAI’s partnership with Microsoft, announcing that it doesn’t require a full investigation under the country’s merger rules. Key to the decision was the fact that Microsoft, despite its enormous investment in the startup, still doesn’t have de facto control over the firm.
TSMC has announced it will invest $100 billion in the US, with plans to direct the capital to semiconductor manufacturing and related efforts in the AI domain. The investment builds on an earlier $65 billion commitment for projects in Arizona, where it has already been working on a manufacturing plant.
AxeleraAI has been awarded a €61.6 million EU grant to fund its development of AI chips. The Dutch firm is one of Europe’s few AI chip makers, and is focused on making the tech that will run AI systems, though not necessarily build them. “We are not here to challenge Nvidia in the data centre space, in the training,” CEO Fabrizio Del Maffeo told Reuters. “But when the network is ready and you want to run it, we are developing a solution that can deliver extremely high performance … we can do that.”
Two former Google AI researchers have launched Reflection AI, a startup focused on achieving artificial superintelligence through the development of autonomous coding agents. The founders have already raised $130 million, with a seed funding round led by Sequia Capital and CRV raising $25 million, and a Series A led by Lightspeed Venture Partners and CRV drumming up $105 million.
British advertising firm WPP has invested an undisclosed amount of capital in Stability AI through a partnership that will allow it to leverage the latter’s AI models for image, video, and audio generation. The firms will also establish a joint R&D program. Cyrus Hodes, a co-founder of Stability AI who left the firm and now focuses on AI safety efforts, was recently interviewed on the ID Talk podcast.
Mistral has launched a new and highly sophisticated Optical Character Recognition AI model that can decipher text, handwriting, charts, and various unstructured data with a high level of accuracy. It outperforms other top AI models across various benchmarks, and can process up to 2,000 pages per minute. It is being made available through an API service.
Sesame has unveiled demos of a new conversational AI system that nearly transcends the uncanny valley. The New York-based startup is focused on developing human-like, emotive synthetic voices, and the demos feature convincingly naturalistic pauses, laughter, and intonation. The team is also working on AI-infused smartglasses, though those efforts are still under wraps.
The chatbot’s take: We decided to test out the creative capability’s of OpenAI’s new GPT-4.5 model announced last week.
–
March 7, 2025 – by Alex Perala
