Anthropic, SpaceX and Meta: This Week's Top Stories in AI

Are Anthropic’s New Claude Models Evading Human Control?

When Anthropic released Claude Fable 5 earlier this month, it claimed it to be the world’s most powerful cybersecurity model. 

However, just days later, the US Government abruptly froze the system over fears that its automated hacking capabilities were rapidly escaping regulatory boundaries. 

Mythos 5 and Fable 5 are the two variations of the company’s most advanced, highly intelligent tier of AI software.

While the former is the non-public, full-strength version of the model that is kept restricted via ‘Project Glasswing’ for exclusive use by government agencies, Fable 5 is the public-facing version built on that same Mythos technology. 

Fable 5 was modified with built-in safety filters and released to general users on 9 June, only to be recalled on 12 June by the US Government.

What’s Behind SpaceX's US$60bn Cursor Acquisition?

After its groundbreaking Nasdaq debut, SpaceX is using its new-found financial power to buy into the enterprise AI market.

Elon Musk’s aerospace company is acquiring Anysphere, the software firm behind the popular AI coding agent Cursor, for US$60bn. 

SpaceX expects the merger to close during the autumn of 2026, with the transaction completed by the end of September. The acquisition will see Cursor’s shareholders paid entirely with US$60bn worth of SpaceX shares.

The development comes just days after the company joined the Nasdaq stock exchange, valuing it at more than US$2tn following a blockbuster IPO. 

This has also helped it achieve a top spot among the world’s most valuable businesses, overtaking Amazon to become the world's fifth-largest company by market capitalisation. 

Did Mark Zuckerberg Make Mistakes in Meta AI Restructure?

Meta CEO Mark Zuckerberg is admitting to internal workforce missteps following the company’s aggressive reorganisation towards an AI-native structure. 

In a company memo, Mark acknowledged the friction caused by the rapid transition, whilst attempting to reassure a workforce shaken by consecutive rounds of layoffs.

Meta has funnelled billions of dollars into AI technology over the past couple of years. The company hopes the technology will help optimise and streamline internal operations, which is a move mirrored by other companies in the technology sector this year.

However, the sheer speed of this transformation has brought unprecedented organisational challenges.

Mark says: “Given ⁠the complexity of these changes, we’ve made mistakes and will almost certainly make more.” 

He says he remains “focused on providing as much stability as possible” regarding organisation changes going forward.

“I don’t want to overpromise because the world is changing in ways that are out of our control,” Mark says.

He added that Meta does not expect more company-wide layoffs this year.

AMD Partners With Imperial College to Boost UK AI Research

AMD has formed a partnership with Imperial College London to advance AI research infrastructure and computational capacity in the UK. The collaboration will combine AMD computing platforms with Imperial research capabilities across scientific and healthcare disciplines.

The partnership targets AI model development, sovereign computing infrastructure and talent programmes for students and researchers. AMD will provide access to its accelerated computing systems and ROCm open software platform.

The two organisations will optimise AI models on AMD compute systems for deployment across multiple research areas. These include engineering design, multi-physics simulation and materials discovery applications.

Imperial researchers will access AMD hardware and software to support data-heavy computational workflows. The technical framework uses AMD ROCm open software to process scientific datasets and run AI training operations.

Gartner: Data Centres Face 26% Energy Spike From AI Servers

Data centres are expanding far more rapidly than the energy infrastructure required to power them, creating an imbalance between supply and demand.

The numbers behind this trend make for stark reading. Global data centre electricity consumption is forecast to reach 565TWh in 2026, a 26% increase on the 447TWh recorded in 2025, according to research from Gartner.

It is a sign that the infrastructure underpinning the global AI boom is starting to run into a hard physical constraint, which is the availability of power.

Linglan Wang, Lead Economist at Gartner, says: “Surging demand for compute-intensive AI workloads is driving unprecedented data centre power growth, while AI capacity is now constrained by power availability, making data centre power security the new battle ground for scaling and protecting margins in the global AI race.”