Artificial Intelligence

SK Hynix is capitalizing on what analysts call a memory chip supercycle driven by AI training demand that shows no signs of slowing. The Korean giant's high-bandwidth memory products command premium pricing as hyperscalers compete for scarce capacity. Unlike previous memory cycles driven by consumer electronics, AI demand appears structurally higher and less cyclical. For investors, this represents a fundamental shift in semiconductor economics where memory becomes a strategic constraint rather than a commodity input.

Trump's claims about saving Iranian women from execution involve both real cases and AI-generated content, complicating verification.

SpaceX secured an option to acquire AI coding assistant Cursor for $60 billion later this year, marking the largest pre-revenue technology acquisition in history. The deal suggests Musk's xAI desperately needs developer infrastructure as competition intensifies with OpenAI and Anthropic. Cursor's revenue remains under $100 million annually, making this a 600x revenue multiple that values potential over performance. The timing aligns with xAI's struggle to build enterprise developer tools while OpenAI slashes o3 pricing by 80 percent.

Anthropic launched an investigation after unauthorized users gained access to its unreleased Mythos AI model, potentially exposing proprietary training methods to competitors. The breach occurred through API endpoints that should have been restricted to internal testing teams. This marks the third major AI model leak in six months, following incidents at OpenAI and Google. Enterprise clients now face uncomfortable questions about data security when these models process sensitive corporate information.

Memory semiconductor stocks trade at 47x forward earnings, sparking debate whether AI demand justifies valuations reminiscent of the dot-com peak. Samsung and SK Hynix have doubled since October, driven by enterprise AI server buildouts requiring high-bandwidth memory. The supercycle thesis assumes data center memory demand will triple by 2027, but supply additions from Chinese manufacturers could crater pricing. Investors betting on sustained margins face the same cyclical reality that has crushed chip stocks every five years for three decades.

The most telegraphed succession in tech finally happened. Tim Cook steps down as Apple CEO after 13 years, with hardware chief John Ternus taking over and Cook moving to executive chairman. Ternus ran the iPhone, iPad, and Mac divisions through Apple's most profitable decade, but he inherits a company facing its first real AI lag since the iPhone launch. The transition timing suggests Apple wants fresh leadership for its next platform shift, not a gentle handover to preserve legacy.

Amazon just wrote the biggest AI infrastructure cheque on record. The $100 billion deal with Anthropic guarantees compute capacity through 2030 and positions Amazon Web Services as the primary cloud for Claude development. This is Amazon's insurance policy against Microsoft's OpenAI monopoly, but it also locks Anthropic into AWS architecture at precisely the moment when model training is shifting toward specialized chips. The bet works if Anthropic can challenge GPT performance. If not, Amazon just overpaid for second place.

The KOSPI closed at an all-time high as Samsung and SK Hynix rode AI chip demand to double-digit gains. Samsung's memory division reported 40% revenue growth quarter-over-quarter, driven entirely by high-bandwidth memory for AI training. The rally reflects South Korea's position as the only real alternative to Taiwan for advanced semiconductors, but it also prices in perfect execution on next-generation chip architectures. Any stumble in 3-nanometer production will hurt.

The NSA is running Anthropic's Mythos model for classified operations while the Defense Department has the same technology on its prohibited list. This bureaucratic contradiction exposes how different agencies are applying AI security standards, with the intelligence community apparently deciding Anthropic's safeguards are sufficient for national security work. The irony cuts deep: the agency responsible for digital security is using AI that military procurement won't touch.

Taiwan's stock market hit $4 trillion in value this week, edging past the UK for the first time. TSMC alone now trades at £615 billion, making it worth more than Shell, BP, and Vodafone combined. The island nation's semiconductor monopoly has turned geopolitical risk into a premium, not a discount. Every ChatGPT query and Tesla autopilot decision runs through Taiwanese fabs, creating the strangest economic moat in modern history.

OpenAI will start selling ads inside ChatGPT responses within days, marking its first serious revenue diversification beyond subscriptions. The move puts it in direct competition with Google's $307 billion search ad business. Early tests show users clicking through at rates 40 percent higher than traditional search ads, according to industry sources. If OpenAI can maintain that engagement advantage, it could claim a meaningful slice of the £50 billion UK digital advertising market by 2026.

Figma's new AI design tools have triggered what insiders call its 'iPhone moment', with enterprise adoption accelerating faster than any product launch since Slack. Professional designers are completing projects 60 percent faster using the AI features, according to early user data. Meanwhile, struggling shoe brand Allbirds has pivoted entirely to AI-powered personalized footwear, betting its survival on algorithmic fit prediction. The contrast illustrates how AI winners extend existing advantages while losers use it as a lifeline.

ASML just raised its 2026 revenue guidance, confirming that AI chip demand isn't a bubble — it's a structural shift. The Dutch monopolist that makes the machines that make advanced semiconductors is seeing orders pile up as every tech giant races to build AI capacity. When ASML gets bullish, it means the entire chip supply chain from Taiwan to California is about to get busier. This is the closest thing to a leading indicator for tech capex spending.

ByteDance just launched Seedance 2.0 everywhere except America, creating a parallel innovation ecosystem that sidesteps TikTok ban anxiety. The new platform gives global developers access to ByteDance's AI tools and recommendation algorithms — the same tech stack that made TikTok unstoppable. This isn't retreat; it's ByteDance building influence through infrastructure while Washington focuses on banning apps.

OpenAI is simultaneously attacking Anthropic and backing away from Microsoft — a two-front war that suggests the AI leadership is feeling cornered. The public spat with Anthropic isn't about technology; it's about positioning for the next funding round and enterprise contracts. Meanwhile, the Microsoft relationship that saved OpenAI from bankruptcy is becoming a liability as regulators circle and competitors cry monopoly. Sam Altman built the most valuable AI company by playing all sides. That strategy is expiring fast.

Corporate AI deployment is happening faster than governments can write the rulebooks — and that's exactly how Silicon Valley planned it. Companies are embedding AI into core business processes while regulators are still debating definitions, creating a fait accompli that makes future restrictions nearly impossible. The European AI Act and proposed US frameworks assume they can catch up to technology that's already reshaping every knowledge job. They can't. By the time meaningful oversight arrives, AI will be so integrated that rolling it back would crash entire industries.

Someone just tried to murder the world's most prominent AI CEO at his home — which means the technology that's reshaping civilisation now comes with assassination attempts. The charges of attempted murder against the attacker signal this wasn't random violence but targeted action against Altman specifically. Tech executives have faced protests and criticism, but physical violence represents a dangerous escalation. If AI leaders need security details just to live in their own neighbourhoods, it suggests the public debate around artificial intelligence has moved far beyond policy disagreements into something much darker.

Someone just crossed the line from keyboard warrior to actual arsonist in the AI debate. A 20-year-old threw a Molotov cocktail at Sam Altman's San Francisco home early Friday morning, then showed up at OpenAI's headquarters threatening to burn the building down. Police arrested the suspect outside the office, but the incident signals a dangerous escalation in anti-AI sentiment. This isn't just about one unhinged individual — it's about the real-world security costs now facing AI executives as the technology becomes more polarizing. Expect insurance premiums and security budgets to spike across the sector.

The gloves are off in AI's subscription wars. Anthropic just banned OpenClaw — the viral open-source tool with 247,000 GitHub stars — from using Claude Pro and Max subscriptions, forcing users onto expensive pay-as-you-go APIs that can cost $5,000 daily for autonomous agents. The timing reeks of retaliation: OpenClaw's creator Peter Steinberger just joined OpenAI for millions after negotiations with Anthropic failed. This isn't just about one tool — it's Anthropic choosing enterprise control over open-source innovation, pushing developers toward its $100 million Partner Network instead of community-built alternatives.

The bond market's new kings aren't wearing suits. AI hyperscalers borrowed $121 billion last year — more than four times their historical average — and analysts expect them to hit $140 billion annually through 2028, potentially exceeding the Big Six banks' $157 billion. Meta's $30 billion October deal was the largest non-M&A high-grade bond ever, while Oracle's credit default swaps tripled after its $18 billion September raise. The supply shift is reshaping credit markets: hyperscaler capex needs are driving net corporate issuance up 30% to $945 billion this year, forcing traditional bank issuers to step back from their record Q1 pace.