- AI: The New Frontier
- Posts
- Size isn’t everything!
Size isn’t everything!
Just don’t mention AI
Time for a catch-up!
OK, a lot of the focus this week was on Apple’s latest product releases, and in particular the iPhone Air, which incredibly is only 5.6mm thick. From an AI perspective, it was interesting that whilst referenced, AI was not the key focus of any of the presentations by Apple, which is a complete change from a year ago. Given all of the challenges Apple have faced with Siri, it‘s clear that whilst the products still use AI extensively, Apple are shifting the focus away from AI and back to the products.
However, for me, the most significant news relates to how AI is changing the World of e-commerce and online search, both of which are now fusing together. I’ve already covered the fact that enquiries via LLMs is leading to a drop in traffic via traditional web search tools like Google, as well as AI created summaries resulting in less through-traffic to websites.
This situation is going to increase, with reports this week highlighting that travel websites such as Booking.com, Expedia, and Airbnb are now starting to face disruption from AI-powered agents. The challenge of course is how to optimise your website for searches, when it’s AI rather than a human looking?
There’s also the issue about Intellectual Property (IP) and what I perceive is a further shift towards the monetisation of data, an area where there was movement this week. Most people are already familiar with RSS feeds (Really Simple Syndication), which enables users to automatically receive updates from their favourite blogs, news sites, or podcasts in one place, without having to visit each site individually.
So, it caught my attention when it was announced that one of the two co-creators of RSS had launched DSNP-AI, a blockchain tag that lets publishers mark, price, and track how AI models use their content. Five news sites and three data brokers start a public test next quarter. The system supports per-token fees and can auto-remove unlicensed copies.
Disputes over AI using IP continue (see later), so I can envisage the situation were tools such as DSNP-AI effectively start to licence the transfer of data, adding another layer of costs to the use of AI.
Anyway, let’s dive right in on the other news from last week …
Big Sharks vs. small fish
There was a lot happening connected to the computer chips used to power AI this week, following news last week from Nvidia about its new chips coming out in 2026. First up was news that OpenAI is planning to produce its own AI chips in collaboration with Broadcom, aiming to meet surging demand while mitigating GPU supply constraints. Broadcom’s stock surged around 11–15% as the deal, potentially valued at $10 billion, signalled the continued importance of AI hardware and infrastructure issues. It also underlined the main players hedging their bets against over-reliance on Nvidia, as we saw with Meta’s similar move recently to develop their own chips.
ARM and Intel also announced new chips, all with higher performance, lower power consumption and lower latency, which is all important for AI workloads such as speech recognition, vision and generative output.
Infrastructure was also in the news, with Microsoft announcing a $17.4 billion deal with a company named Nebius, specifically to add 400 MW of AI - ready energy capacity across Europe, with this being 100% renewable energy by 2028. This will ease the growing pressure AI is placing on local energy sources, potentially leading to computing costs reducing by 10% - 15%.
Microsoft also announced that it would be licensing Anthropic’s Claude models for Azure and Microsoft 365 Copilot under a new multi-year pact. The move gives Microsoft an alternative AI supplier, easing dependence on OpenAI and giving Microsoft pricing leverage before contract renewals later this year.
Whilst we’re on the subject of Microsoft, their latest Windows 11 security update introduced a load of new features, most notably a Settings AI Agent, alongside a redesigned Windows Hello UI and Recall homepage.
The last bit of “Big Shark” AI news is that due to its involvement in AI infrastructure projects, Oracle’s share value increased by 41% in early trading this week, at one point adding more than $80 billion to the wealth of co-founder Larry Ellison, briefly making him richer than Elon Musk.
Legislation, policy and other news
In other news, OpenAI co-authored new research around AI hallucinating. The research suggested this is not reducing and may in fact get worse as models are made more capable, with the conclusion being that there needs to be improved benchmarks and controls.
Then, with almost perfect timing, a new study revealed that 61% of enterprises have faced insider security breaches, with AI complexities compounding file protection challenges. Why perfect timing? Well, this week we got some more details about the alleged “behaviour” of former xAI engineer Xuechen Li. According to court filings, when he left xAI to join Open AI a couple of weeks ago, he sold $7 million in shares in xAI and then proceeded to copy all of Grok's trade secrets and confidential files.
As mentioned, the publishing industry is struggling and stated this week that it is facing an “Existential Crisis”, as Google’s AI - driven summaries and “AI Mode” features are cutting referral traffic to news sites by as much as 89%, prompting publishers to seek licensing deals and regulatory recourse.
To put that into context, this week Anthropic agreed to pay $1.5 billion to settle a lawsuit with a group of authors, over unauthorised use of copyrighted texts for training AI models. Therefore, it’s not surprising to then hear that Apple are being sued for similar transgressions as part of a class-action lawsuit, and Warner Bros. are accusing Midjourney of infringing its creators’ rights with its AI - generated images.
That’s why I think that the DSNP-AI tool is so important, as data becomes even more valuable.
Finally, as an aside and in happier news for wine lovers, it was announced that vineyards in Napa Valley are now using self-driving tractors and yield‑detecting cameras to harvest grapes, improving quality and reducing costs.
That’s all for now from me at the end of yet another week … so stay informed, stay critical, and wherever possible - stay ahead.
Regards
Tom Carter