Time for a catch-up!

Well, we’ve had the usual slew of new releases and upgraded tech, some of which is really, really impressive. We’ve also had lots of developments around investment and legislation, which I will discuss in greater depth later.

However, the big news for me (and I suspect everyone else) was Open AI buying “io”, the company owned by Jony Ive, the man behind all of those iconic Apple designs. I mentioned the rumours 6 weeks ago in my newsletter, when the talk was of a $500 million USD price tag for the partnership. How wrong I was!!! Open AI just bought the company for a jaw-dropping $6.5 billion (yes with a B), even though it currently has no products and no customers. The deal doesn’t even include Ive’s services, although there is a separate contract between Ive’s design company LoveFrom and Open AI.

Sam Altman has described one of the prototype products as “the coolest piece of tech the world will have ever seen”. That’s a lot of hype to live up to, but given Ive’s track record, it’s not something I would discount, and clearly OpenAI feel the same way.

I’d like to say that was all that’s happened this last week, but there’s been loads more things, so let’s dive right in …

Big Sharks vs. small fish

Last week we had already had Saudi Arabia unveil “Humain”, a state-backed AI company designed to centralise the Kingdom’s research and infrastructure efforts, aiming to transform it into a global AI innovation leader. That was then followed by the United Arab Emirates opening Phase 1 of a new 5 GW AI research campus in Abu Dhabi, This includes an annual import quota of 500,000 Nvidia AI chips, which did wonders for Nvidia’s stock price.

This also linked to analysts predicting the TSMC would be doubling its manufacturing to meet surging demands from Nvidia, Google, and AMD, solidifying Taiwan as the world’s semiconductor hub.

This week wasn’t good news for everyone though, with Meta postponing the rollout of its flagship “Behemoth” model that they had announced only a few weeks ago, quoting “performance concerns”, and triggering a 2.4 % drop in its share price as markets weighed on the setback.

Amazingly, OpenAI didn’t release any new models this week – probably too busy spending that $6.5 billion. However, they did release a research preview for Codex, a parallel-tasking AI coding assistant targeting software developers. Codex could drastically reduce software development cycles, enabling startups and enterprises to launch products faster.

Not to be outdone, at Google I/O, Alphabet spent hours unveiling loads of new products, including Gemini Live (which lets you point your camera and ask questions), Google AI Ultra, and Project Astra’s universal AI assistant that is breaking new grounds in intuitive AI.

Google is obviously gunning for dominance in the AI search wars, competing head-on with OpenAI's ChatGPT and Microsoft’s Copilot. They showcased the blending of Gemini 2.5 with personalised tools for shopping, creating, and more. Interactive AI like AI Mode signals a shift in how we source and use information, with a focus on retention and immersion within Google’s ecosystem. The aim is to keep users locked into Alphabet’s portfolio longer, disrupting search behaviours globally. 

By integrating everything so seamlessly into its ecosystem, Google is banking on customers not even realizing they’re interacting with it. Given the fact that Google’s AI Overviews already handles searches for 1.5 billion users monthly, I think that it will be increasingly difficult to know when you’re actually using AI.

The two attention grabbers though were Veo 3, an updated AI video generator that can also generate accompanying sound effects, background noise, and dialogue. I would recommend checking out some of the videos online, which are frankly amazing. They also showed off their Google Meet Translations and a 3D conferencing platform, which translates different languages in real-time during meetings, matching the speaker's voice and tone - equally stunning, and something that changes business meetings forever.

Then we also had Anthropic’s release of Claude Opus 4 and Sonnet 4. Opus is an AI agent that codes, reasons, and works autonomously for hours, which is definitely going to have a massive disruption impact in software development cycles. The larger context window means that both versions can run longer and more complex tasks, broadening their ability to carry out complex tasks. Perhaps not as attention grabbing as the Google releases, but equally as significant in terms of how it could impact on things moving forward.

Legislation, policy and other news

Meta was also having a bad week more generally. After pausing due to privacy laws, Meta announced that it was back to using European user data for training AI. No sooner had they announced this when they were hit with legal action by a group stating their GDPR rights were being breached.

Intellectual Property was also in the news in the UK, with legislators considering an amendment to require AI firms to disclose usage of copyrighted works. Elton John also entered the fray, expressing concern about the impact of AI on the creative industry.

However, even as there were indications that regulators in the US and UK would prefer more controls on AI, a provision in Trump’s major tax bill would block state and local governments from enacting AI regulations for ten years, something which echoed Elon Musk’s call for IP rights to be scrapped.

There was more negative news on the job front, at least for those working as software coders, with Microsoft announcing it would be laying off roughly 7,000 employees (3% of its global workforce), and reallocating resources into its advancing AI initiatives.

There were also concerns about the likely long-term environmental impact of AI, with Google revealing that its emissions have raised by 48% since 2019, with most of that down to AI development, which accounts for 20% of global data-centre power demand

However, to end on a more positive note, Microsoft’s AI Agents discovered a new chemical in 200 hours, which heralds the possibility of being able to develop and synthesise new drugs more quickly. The Ottawa Hospital also announced that by using an AI-powered ambient voice to interact with patients, it reduced physician burnout by 70%, whilst achieving 97% patient satisfaction rates. Perhaps if they could get AI to help doctors write prescriptions that we could actually read, then we’d all be happier.

Stay informed. Stay critical. And wherever possible—stay ahead.

Regards

Tom Carter