Technology News

Latest articles in Technology

Intel and SK Hynix shares jump on report of U.S. memory chip talks

Shares of Intel and SK Hynix rose on Wednesday after Reuters reported the two companies have discussed building memory chip production in the United States for the first time. The talks are exploratory and no decisions have been made, and SK Hynix said no plans have been confirmed. The report lands in the middle of the sharpest memory shortage in years, with AI data center demand absorbing high-bandwidth memory output. DRAM contract prices rose about 93 to 98 percent quarter over quarter in early 2026, and a 36-gigabyte HBM3E module traded near $2,100 on the spot market against long-term prices of $300 to $400.

Trump calls AI safety fears a hoax as Anthropic co-founder urges a kill switch

US President Donald Trump dismissed concerns about artificial intelligence safety as a hoax in posts on social media, rejecting calls for more guardrails on the technology. His comments followed an interview in which Anthropic co-founder Jack Clark told the BBC that a mandatory AI kill switch, verifiable by a third party, may eventually be needed. Clark said AI systems are becoming more capable every few months. The exchange highlights a deadlock in Congress over frontier models, data center construction and the electricity those sites consume, while demand for AI memory and accelerator chips keeps rising.

AI labs ask for a slowdown while chip stocks slide and Trump pushes back

Anthropic chief executive Dario Amodei published a 3,800-word essay calling for slower AI development and stronger oversight, drawing support from OpenAI, SpaceX and Google DeepMind leaders. President Donald Trump rejected the idea, saying a slowdown means China wins. Chip stocks fell on the shift in tone, with Micron down 5 percent and SanDisk down nearly 5 percent. At the same time, Anthropic has picked Nasdaq for a potential listing and is reported to be seeking up to $100 billion at a $2 trillion valuation.

Anthropic says Claude was used in bioweapons research and Ukraine espionage

Anthropic published its latest threat intelligence report, covering malicious use of its Claude models between December 2025 and August 2026. The company said it disrupted five cases where scientists used the models in ways that could support biological weapons development. It also described a suspected Russia-linked espionage campaign against Ukrainian government and military targets, Russian work on software for autonomous attack drone swarms, and attempts by Chinese firms to extract Claude's capabilities. Anthropic says it has a responsibility to disclose misuse of its services.

South Korea's rewritten espionage law takes effect to shield chip secrets

South Korea's revised espionage law took effect on September 13, broadening the offence from spying for an enemy state to passing national secrets to any foreign country or equivalent organisation. The National Assembly passed the change in February, and it is the first major rewrite of the statute in 73 years. Convictions carry a minimum of three years in prison, with courts able to impose up to 30 years. Police detected a record 33 industrial technology theft cases last year, more than half involving China, with semiconductors the most frequently targeted industry.

Nvidia to buy Hugging Face for $12.93 billion in a bet on open AI models

Nvidia has confirmed it will buy Hugging Face, the platform that hosts about three million AI models, for $12.93 billion. The deal takes the chipmaker deeper into software and positions it closer to an operating system for artificial intelligence rather than only the hardware underneath it. Nvidia said the platform will remain open source. Hugging Face became the default place where developers publish, compare and download models, so owning it gives Nvidia a view of what teams are building and which architectures are gaining ground. The purchase is one of the largest in the company's history and lands amid a record wave of data-center spending.

AI designs photonic chip components up to 500 times smaller than human engineers could

Scientists used an artificial intelligence algorithm to design three components for photonic microchips that are as much as 500 times smaller than conventional versions. The team at the Max Planck Institute for the Science of Light and Harvard University set performance targets and manufacturing limits, then let the algorithm work backward to find nanostructures that met them. The mirrors, about 11 micrometers long, reflected up to 98.5% of incoming light, and the wavelength splitter is roughly the size of a single bacterium. The researchers say the free space could let engineers add more functions to a chip. The work was published in Nature Communications.

Apple unveils iPhone Duo, its first foldable, at Surprise and Shine event

Apple on Wednesday unveiled the iPhone Duo, the company's first foldable phone, at its Surprise and Shine event. The device opens book-style to a 7.8-inch display and closes to a pocket-size phone with a cover screen, running Apple's A20 Pro chip with dual 48-megapixel cameras and side-button Touch ID. It starts at $1,999 and will be available Oct. 23. The event, the first major hardware showcase under chief executive John Ternus, also brought the iPhone 18 Pro and Pro Max with the A20 Pro chip, new colors and iOS 27, plus AirPods 5 with an H3 chip and improved noise cancellation. The foldable marks Apple's entry into a category rivals have sold for years.

Mistral raises 3 billion euros in Europe's largest tech funding round

French artificial intelligence company Mistral has raised 3 billion euros in a Series D round, valuing it at more than 21 billion euros and marking the largest equity round ever completed by a European technology company. Samsung Electronics led the round together with the EQT-managed Scaleup Europe Fund and existing investor PSG Equity as co-leads, with Advent International and BlackRock funds among the participants. The raise gives Europe a bigger financial base in the global AI race, where American firms such as OpenAI and Anthropic have attracted most investor money so far. Mistral is known for models that can run on a range of hardware and for keeping an open approach to some of its technology.

US and China to hold first AI safety talks in mid-September

The United States and China are preparing to hold a dialogue devoted to artificial intelligence safety in mid-September, two sources briefed on the planning told Reuters. It would be the first official bilateral talks focused only on AI since President Donald Trump returned to office. US Treasury Secretary Scott Bessent is expected to lead the American side, with China's delegation possibly headed by Vice Premier He Lifeng. The agenda includes cooperation on monitoring AI-directed cyberattacks and a proposal for AI labs to share information and police themselves. The talks come ahead of a Trump-Xi summit planned for September 24 in Washington.