You’re receiving this email because you subscribed to the newsletter. If you’d like to unsubscribe, you can manage your preferences here with one click (or use the button at the bottom of the email).

Welcome to this edition of Loop!

To kick off your week, I’ve rounded-up the most important technology and AI updates that you should know about.

‏‏‎ ‎ HIGHLIGHTS ‏‏‎ ‎

Let's jump in!

1. Apple partners with Google, ahead of an updated Siri

We start this week with Apple, who have officially partnered with Google and will begin using their AI models. The announcement comes before an overhaul of Siri, which has barely changed over the last decade.

The tech giant confirmed that Google's Gemini models and cloud infrastructure will underpin future Apple Foundation Models. It’s believed that Apple is paying $1 billion for access to Google’s technology, although neither company has confirmed the figure.

This is a notable shift for Apple, which has historically prided itself on vertical integration and keeping everything in-house. The company does offer its own AI models, but they’re significantly less powerful than ChatGPT and Gemini.

Of course, Apple signed a similar deal with OpenAI in June 2024 and ChatGPT has since been integrated with Siri. That’s unlikely to change due to concerns around anti-trust - but Google’s models will now become the preferred option.

2. Google's AI can now create vertical videos

On the surface, this seems like a pretty small story. But I think this will have a huge impact on social media platforms and how advertisers create new campaigns.

Social media platforms have been prioritising video content for the last few years, as they hope to close the gap with TikTok and keep users engaged for longer. Facebook and Instagram have done it with Reels, while X is trying to do the same.

This has led to a rise in vertical videos being shared online, with advertisers taking note and doing the same. Many advertisers now pay influencers to create these videos for them and promote their product, with hundreds of variations used to encourage sales.

But if these companies can use Google’s AI video generator, this strategy will change over time as the technology improves. If you’re selling a new lipstick, you can easily create an AI video that looks pretty realistic and promotes the product.

Simply include two reference images - such as your product, and a real influencer - and Google Veo will use them to create a realistic video. It’s incredibly easy to use and people are already using the technology to create ads.

I’ve recently used Google’s tool to create a short, animated film - which I’ll share in the coming weeks. The technology is really impressive, but it still has its limitations so you need to carefully check the video that was generated.

I think we'll see a lot more AI-generated content on TikTok and Instagram in the coming months as teams experiment with these tools. The barrier to entry has dropped significantly, and businesses that were previously priced out of video marketing can now compete.

3. Taiwan invests $250 billion in US semiconductor manufacturing

The US and Taiwan have struck a significant trade deal, which will see Taiwanese chip companies invest $250 billion in American manufacturing.

Under the agreement, Taiwan's government will guarantee $250 billion in credit for these investments, while the US will cap reciprocal tariffs on Taiwanese goods at 15%.

TSMC has already invested $40 billion in its Arizona plant, where it produces chips for Apple and Nvidia, and plans to expand that operation even further. Overall, the US is aiming to relocate 40% of Taiwan's semiconductor supply chain to US soil.

It follows rising tensions between China and Taiwan, where the world’s most advanced semiconductors are produced. If China was to invade Taiwan, there would be a catastrophic impact on the global supply chain.

The US is trying to minimise this risk and encourage more domestic manufacturing, which is badly needed. However, it’s concerning that Europe isn’t doing the same - especially when there are growing calls for the EU to invest more in local technology companies.

4. UK company beats Peter Thiel and secures European satellite licence

A British satellite company has secured a highly contested spectrum licence from Liechtenstein, beating rivals that were backed by Peter Thiel and Chinese investors.

The Ka-band spectrum - which is an important frequency for high-speed internet and military intelligence applications - will give the company a significant foothold in the low Earth orbit satellite market.

Historically, European companies have lagged behind their American and Chinese counterparts in this area, but Open Cosmos is looking to close the gap.

The company, which has spent a decade launching satellites for telecoms and Earth observation, plans to launch two Ka-band satellites "in the coming weeks.”

The company has framed this as an important win for European "space sovereignty and digital resilience" - reflecting Europe’s concern about how dependent it is on American technology and the ways it can be used against them.

5. British police blame Microsoft Copilot for intelligence mistake

In an embarrassing admission, West Midlands Police said that Microsoft's AI chatbot was used to gather intelligence and included details about a football match that never happened.

The fictional West Ham versus Maccabi Tel Aviv game was cited in an intelligence report, which led to fans being banned from a match in November 2025.

Previously, the police force blamed social media for the mistake. Then they blamed a Google search, before finally admitting that AI was used instead.

It’s a reminder that we need to carefully review AI-generated results, especially when they’re used for important decisions. But organisations need to do more than simply reminding their staff about the risks.

People will naturally take the easiest path available, so this will always be a problem. If your business is planning to use AI in this way, you should have a firm process in place - or design one - that forces people to stop and verify, rather than simply accepting whatever the model produces.

Anthropic wants Claude to become a digital employee

Anthropic has unveiled a new tool for the workplace. Cowork is very similar to Claude Code, but it’s much easier to use and supports more general work - rather than just writing code.

With the Claude desktop app, you can enable Cowork and allow the AI to access files on your computer. For example, Claude can analyse a folder that contains your receipts - then log them as expenses on Workday.

I've been testing it for some time and I'm very impressed at what it can do. I asked Claude to analyse a file that contains all my newsletter posts, identify the key insights, and then create a dashboard to display them.

Before it accessed the file, Claude asked for my permission and then ran Terminal commands to analyse the data. Once it was ready, the AI opened a new Chrome tab and loaded the dashboard for me.

That’s incredibly powerful and I can see how non-technical people will really benefit from this. I can also see how this will disrupt the technology industry in the longer run.

For years, companies have been forced to rely on consulting companies, or paid thousands to analyse their data and extract key insights. If you can open a desktop app and ask an AI to do that instead, you can skip the middleman.

While this is fantastic to see, I think there are huge security concerns that will prevent businesses from adopting Cowork. What if Claude tries to extract confidential data from your machine and then upload it online? What if it misinterprets what you want to do and deletes important files?

Anthropic says there are restrictions in place and Claude will always ask for your permission, but the risks are still high. While humans should be double checking what the AI is doing, businesses should expect that people will become too comfortable with technology and won't always do that.

Despite all the incredible advances with Large Language Models, the core weakness remains: we cannot trust them for important actions. Until AI companies solve this, I'm not sure how people will become comfortable and allow it to access important files.

ChatGPT will soon have ads

OpenAI is bringing adverts to ChatGPT, with testing set to begin in the coming weeks for US-based free users and those on the new ChatGPT Go subscription tier ($8/month). If you already pay for ChatGPT, you won’t see any ads.

The company says that these "clearly labelled" ads will appear in a dedicated section at the bottom of chats, separate from the actual responses. They also say that no user data will be sold to advertisers, and that these ads won’t have an influence over ChatGPT's answers.

It’s been a long time coming for OpenAI, who have been working on this for many months. As I mentioned back in October, they have been laying the groundwork for this with their Sora app - a TikTok rival that shows AI videos and collects data about your interests.

This data is incredibly lucrative for advertisers, who want to specifically target people that are more likely to buy their products. In order to collect that data, OpenAI needs people to use their services for longer than they currently do - and include every detail that they can about their lives.

The company is aiming for $100 billion in annual revenues by 2028, while they currently generate ~$20 billion per year. That’s a huge task for OpenAI, which has committed to spending $1.4 trillion on AI infrastructure over the next decade.

With this new push into advertising, OpenAI is hoping that it can disrupt Meta and Google - which have an incredibly strong grip on the sector and earn tens of billions from advertisers.

🏥 Google releases a new AI model for medicine

💾 RAM prices are now over 4x higher than in September 2025

⚖️ Matthew McConaughey trademarks himself to prevent AI misuse

🥽 Meta has laid off 10% of its Virtual Reality staff

🔗 Google's Gemini can now access your data from Gmail, Search, and YouTube (if you give permission)

🚁 Wing expands its drone delivery service to another 150 Walmart stores

🇬🇧 UK pushes up a law that criminalises deepfakes, in response to X scandal

🤖 Neo's world model allows robots to learn what they see

🛒 Google announces a new protocol for agentic commerce

🎨 Apple launches "Creator Studio" bundle of apps for $12.99 per month

🛰️ SpaceX allowed to launch 7,500 more Starlink satellites

👥 Mira Murati's startup has lost two of its co-founders to OpenAI

This startup from Austin has just raised $12.7 million in Series A funding for its satellite imagery platform, which connects users to over 50 geospatial providers.

The company has carved out a niche by making satellite data easy to access, with its CEO even saying that his teenage daughters use it for their school projects.

But SkyFi is pivoting towards analytics and insights, essentially selling answers rather than just the raw data. For example, you can use their platform to monitor a specific factory and use that to better inform your investment decisions.

Or governments can use it to monitor crops being grown across the country, which allows them to monitor changes in real-time and means that farmers don’t have to report this themselves.

And naturally, there’s a lot of interest from the defence sector too. SkyFi’s funding round was initially targeting $8 million, but that quickly swelled when defence companies became involved.

For them, the advantages are quite clear. They can monitor new testing facilities or spot when troops are being built-up near a country’s border. They can also use SkyFi’s computer vision service to monitor warships and their movement patterns.

But defence companies are also integrating these insights into their own products. Anduril has recently built a VR headset that can take in satellite imagery and guide soldiers on the battlefield.

Satellite imagery is also being used for drone attacks, which allow operators to spot huge numbers of enemy troops and then target them.

Overall, SkyFi is definitely a startup that’s worth watching. For years, I’ve been really interested in the advances that are being made with satellite imagery and the level of insights that you can get from them. With this new funding round, investors clearly feel the same.

If you want to learn more about the company, I’ve included a link below.

This Week’s Art

We’ve covered quite a bit this week, including:

If you found something interesting in this week’s edition, please feel free to share this newsletter with your colleagues.

Or if you’re interested in chatting with me about the above, simply reply to this email and I’ll get back to you.

Liam

Share with Others

If you found something interesting in this week’s edition, feel free to share this newsletter with your colleagues.

Liam McCormick is a Senior AI Engineer and works within Kainos' Innovation team. He identifies business value in emerging technologies, implements them, and then shares these insights with others.