🏢 OpenAI to open first permanent London office in 2027 ↗
OpenAI locked in its first permanent London office, which feels both expected and faintly symbolic. London has already been central to its overseas push, but this makes the expansion look less like a trial run and more like a proper stake in the ground.
The broader read is scale. OpenAI is positioning London as a major research base outside the US, which says a great deal about where it thinks talent, demand, and likely influence are clustering right now. Not exactly subtle, that. (Reuters)
🛒 Tesco partners with Adobe to ramp up AI-driven personalised marketing ↗
Tesco is leaning further into AI with Adobe, using customer data analysis to sharpen personalised marketing. It is a very retail move - practical, sales-first, not flashy - but that is where a great deal of AI spending is landing now.
The point is not some sci-fi shopping bot. It is better targeting, tighter campaigns, and nudging customers with more relevant offers. Familiar, yes. Still important, also yes. (Reuters)
🧠 Chinese AI startup StepFun to unwind offshore structure to pave way for IPO, sources say ↗
StepFun is reportedly restructuring to prepare for a Hong Kong IPO, unwinding the offshore setup many Chinese startups used for overseas fundraising. It sounds technical - because it is - but it hints at how AI financing is being refashioned by Beijing’s tighter oversight.
What makes this one stand out is the signal beyond StepFun itself. Reuters says other Chinese AI firms are weighing similar changes, so this could be less a one-off and more the new choreography for getting AI companies public in China. Dryly bureaucratic, but still pretty major. (Reuters)
⚙️ TSMC likely to book fourth straight quarter of record profit on insatiable AI demand ↗
TSMC looked set for another record quarter, driven by relentless demand for advanced AI chips. It is one of those stories that keeps repeating, yet somehow keeps getting louder - every fresh earnings beat turns the whole AI supply chain into a giant humming transformer.
The simple version is that model builders and cloud giants still need more compute, and TSMC sits right in the middle of that rush. When demand for AI infrastructure spikes, TSMC does not just benefit - it becomes the weather vane. (Reuters)
🛡️ Anthropic talking to the Trump administration about its next AI model, co-founder says ↗
Anthropic said it is in talks with the Trump administration about its frontier model Mythos, even after its clash with the Pentagon. That is the interesting tension here - public friction on one side, active engagement on the other. Fraught, perhaps, but very real.
The backdrop is a dispute over guardrails for military use of AI, which led to the Pentagon cutting off business with Anthropic. Even so, the company is making clear it still wants to stay in the national security conversation rather than storm off entirely. (Reuters)
⚠️ AI-boosted hacks with Anthropic's Mythos could have dire consequences banks ↗
Anthropic’s Mythos also showed up in a more alarming context - cybersecurity. Reuters reported that experts see the model as capable enough to intensify complex attacks, especially against banks still running on older systems. Not great, obviously.
What stands out is the split-screen effect. The same model being discussed with government officials is also being framed as a potential accelerant for cyber risk. Frontier AI keeps doing this - dazzling in one room, setting off alarm bells in the next. (Reuters)
FAQ
Why does OpenAI opening a permanent London office matter?
It suggests OpenAI sees London as more than a temporary outpost. The move points to London becoming a major research base outside the US, implying confidence in local talent, customer demand, and strategic influence. In practical terms, it makes the company’s international expansion look more permanent and deliberate.
How is Tesco using AI with Adobe in practice?
This appears focused on marketing rather than flashy, consumer-facing AI. Tesco is using AI-driven customer data analysis to improve targeting, tighten campaigns, and send more relevant offers. The goal is better personalization and stronger retail performance, not a futuristic shopping assistant.
What does StepFun’s restructuring say about the Chinese AI market?
StepFun’s reported plan to unwind its offshore structure ahead of a Hong Kong IPO signals how AI financing in China may be changing under tighter Beijing oversight. The article also suggests that other Chinese AI startups are considering similar moves. That makes this feel less like an isolated case and more like a potential financing pattern.
Why does TSMC keep showing up in AI news?
TSMC sits at the center of demand for advanced AI chips, so rising AI infrastructure spending shows up in its results quickly. Reuters says it was likely heading for a fourth straight quarter of record profit. That makes TSMC a strong signal of how robust compute demand remains across model builders and cloud companies.
Why is Anthropic talking to the Trump administration while also clashing with the Pentagon?
That split shows how complicated frontier AI policy has become. Anthropic is reportedly discussing its next model, Mythos, with the Trump administration even after a dispute over military-use guardrails led the Pentagon to cut off business. The key takeaway is that policy engagement is continuing despite serious disagreements over defense use.
Could Anthropic’s Mythos make cyberattacks worse for banks?
According to the article, experts think Mythos could intensify complex attacks, especially against banks still relying on older systems. The concern is not only that the model is powerful, but that it could lower the effort required for sophisticated offensive activity. That puts cybersecurity and AI capability on the same track, which is why the story stands out.