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:
Siri might finally get smart. Apple is reportedly finalizing a deal with Google that will give its iPhone voice assistant a Gemini-powered brain – which is expected to be a 1.2 trillion-parameter custom model. Apple is reportedly prepared to pay a billion dollars a year for the privilege. But Google Search will not be folded into Siri queries; Apple wants to keep control over that domain.
OpenAI CEO Sam Altman says his firm will finish the year with annualized revenue topping $20 billion, a significant jump over the predicted $10 billion ARR from June. He also confirmed that OpenAI has about $1.4 trillion in committed infrastructure obligations over the next eight years or so to build data centers and secure chips. And he clarified that OpenAI isn’t looking for any government bailouts in the future after comments earlier in the week from the firm’s CFO provoked some controversy.
Microsoft says it will now pursue “humanist superintelligence” via an internal “MAI Superintelligence Team” led by AI chief Mustafa Suleyman. The company is looking to achieve breakthroughs in medicine, materials science, and clean energy, developed under explicit safety guardrails. The move comes after Microsoft finalized a restructuring agreement with OpenAI last week that effectively loosened the bond between the two firms (Microsoft is a huge investor in OpenAI), though they say they will continue to work together closely.
The European Commission is reportedly considering softening aspects of its wide-reaching AI Act after Big Tech lobbying. While the overall law is expected to remain in place, the EC is reportedly considering a one-year grace period for companies that are already selling generative AI tools in the region, delaying fines related to transparency obligations, and making guidance for the monitoring of high-risk AI applications a bit more flexible. The AI Act already entered into force in August of 2024, but has a staggered implementation plan mapped out through August of next year.
China has banned the use of American AI chips in state-funded data centers, with a new guidance requiring that the latter only use domestically-produced chips. Projects under 30 percent completion must rip out foreign chips or cancel orders; more advanced builds will be reviewed case-by-case. The move accelerates Beijing’s push for AI-hardware self-reliance, and actually aligns with the ostensible interests of the US government, which doesn’t want American firms like Intel and Nvidia selling their chips to China anyway.
1X Technologies has unveiled Neo, its first humanoid robot for the home. The Norway-founded, Palo Alto-based startup is pitching its robot as a kind of AI housekeeper that can fold laundry, load the dishwasher, and put away groceries. Except it can’t actually do that yet – not autonomously. At present, Neo often needs to be remotely piloted, and 1X Technologies is hoping to use the data collected from its built-in sensors to train an autonomous system. So its first customers are essentially being asked to trade some privacy for early access. Preorders have opened at $20,000 per unit, or a $500-per-month subscription.
Finally, a friendly reminder that our media production studio has developed an AI-powered daily news podcast. The Morning News Update isn’t aimed specifically at the digital identity space, but instead delivers the latest major headlines and market moves. The script, music, and host are entirely AI-generated, on an automated basis.
The chatbot’s take: We got GPT-5 to put its thinking cap on.
–
By Alex Perala
