Tech News Digest: Wednesday, 5 August 2026

Today in tech, we’re looking at SpaceX’s massive revenue growth fuelled by the global AI boom and some serious concerns regarding the safety of the latest open-weight models. We also have updates on Google’s next-gen Pixel features and a major legal spat between Apple and OpenAI that’s getting rather personal.

SpaceX doubles revenue on Anthropic and Google compute deals

In its first quarterly report since going public in June, SpaceX has revealed that revenue has doubled compared to last year. While Starlink remains a pillar of the business, the growth is largely attributed to massive compute deals with AI heavyweights like Anthropic, proving that Musk’s aerospace firm is now a key player in the AI infrastructure race.

Open-weight AI models catch up to the frontier, but safety lags

A new report from SaferAI suggests that Z.ai’s latest GLM-5.2 model is approaching the capabilities of 'frontier' AI while lacking essential safety mitigations. For the UK tech community, this reignites the crucial debate over whether the accessibility of open-source AI is worth the potential security risks as these models become more powerful.

The Pixel 11’s ‘HiLight’ feature looks like a productivity winner

Leaked details for the upcoming Pixel 11 have confirmed that the much-rumoured "glow" feature is actually called HiLight. This AI-driven tool is designed to intelligently surface and summarise the most important information on your screen in real-time, which could be a genuine game-changer for mobile multitasking and productivity.

Apple widens trade secrets investigation into OpenAI

The legal battle between Apple and OpenAI is intensifying, with Apple now claiming in court filings that even more ex-employees may have taken confidential data to the AI startup. This isn't just a corporate squabble; it could fundamentally change how intellectual property and talent are managed across the industry’s biggest hitters.

AI-supervised exam disaster forces 58,000 student retakes

A massive failure in an AI-powered remote proctoring system has resulted in nearly 60,000 students being told their exam results are void. It serves as a stark cautionary tale for UK educators and firms about the dangers of over-relying on automated systems for high-stakes assessments without sufficient human oversight.

That’s your lot for today—stay safe out there and we’ll see you back here tomorrow.