China readies 7nm chip production in efforts to boost tech self-sufficiency

China readies 7nm chip production in efforts to boost tech self-sufficiency



China’s Hua Hong Group has developed advanced chip manufacturing technologies that can be used to produce artificial intelligence chips, four people familiar with the matter said, a major milestone in Beijing’s efforts to boost tech self-sufficiency.

 


The group’s contract chipmaking business, Huali Microelectronics, is readying a 7-nanometre (nm) chipmaking process at its plant in Shanghai, the people said, which would make it the second Chinese chipmaker with such advanced technologies. Hua Hong is China’s second-largest chipmaker.

 


China’s largest contract chipmaker, SMIC, is at present the only domestic producer capable of making chips with 7 nm technologies.

 


The development comes after Washington eased some of its tech export controls since last year, allowing Nvidia to sell its second-most-powerful AI chips to China.

 
 


Despite the easing, Beijing has encouraged domestic firms to purchase homegrown alternatives, as it seeks to wean itself off foreign suppliers.


Reuters could not determine how Hua Hong achieved the advanced manufacturing capability, its manufacturing efficiency and which major equipment suppliers were involved in the development. Hua Hong’s development of a 7 nm chipmaking process has not been previously reported.

 


But Chinese tech giant Huawei Technologies has been in collaboration with the chipmaker for the 7 nm technologies, three of the sources said. All of the sources declined to be named, because the information is not meant to be public.

 


Hua Hong Group, Huali, its sister company Hua Hong Semiconductor and Huawei did not respond to requests for comment.

 


SMIC uses Dutch chip equipment maker ASML’s immersion machines to make 7 nm chips, but production yields – the number of good chips made per silicon wafer – have remained weak, analysts have said.

 


ASML said it does not comment on questions related to deliveries.


Test production under way


Huali’s research and development on 7 nm chips at its Hua Hong Fab 6 began last year, with support from domestic equipment suppliers including Huawei-backed SiCarrier, which tested its equipment at a facility in Shenzhen last year, a separate source said. SiCarrier did not respond to a request for comment.

 


The development followed an announcement by Hua Hong Semiconductor in December that it planned to acquire a controlling stake in Huali and raise a further 7.56 billion yuan ($1.10 billion) to fund technological upgrades and research at the foundry.

 


Huali is planning initial 7 nm chip production capacity of a few thousand wafers per month by the end of the year, with a goal to ramp up more later, two of the sources said.

 


Chinese graphics processing unit designer Biren is using Huali’s 7 nm line for tape-out, a process in which a chip design is committed to a physical prototype for testing before mass production begins, one of the sources said.

 


Placed on a U.S. trade blacklist in 2023, Biren lost access to TSMC’s contract manufacturing service shortly after. Biren did not respond to requests for comment.


The Hua Hong Fab 6 is the most advanced of seven foundries within the Hua Hong Group and currently manufactures logic chips using 22 nm and 28 nm process nodes, according to the company’s website.

 


By contrast, its Fab 5 produces chips using mature technologies ranging from 40 nm to 55 nm.



Source link

MacBook Neo emerges as Apple's most repairable laptop in more than a decade

MacBook Neo emerges as Apple's most repairable laptop in more than a decade



Apple’s MacBook Neo, the laptop it announced last week that starts at $499 for students, is the most repairable laptop the company has released since 2012, according to an analysis released Friday by iFixit.

 


iFixit publishes repair guides and sells parts and tools for consumer electronic devices, but also provides ratings for how easy items are to fix and keep running. Laptop makers such as Dell Tech and Lenovo Group have used those ratings to improve the repairability of their products.

 


In the teardown published on Friday, iFixit found that Apple had made key changes from previous laptops, such as attaching the computer’s batteries and keyboard with screws rather than glue or rivets, and making it easy to swap out parts such as the device’s camera and fingerprint sensor.

 
 


Apple is widely believed to be targeting the same education markets with its MacBook Neo that Google targets with its low-cost Chromebooks. Kyle Wiens, iFixit’s chief executive, said Chromebooks are frequently repaired, with some school districts such as those in Oakland, California even tapping student interns to fix them.

 


But Apple’s MacBook Neo still scored only a 6 out of 10 on iFixit’s scale, where other machines such as a recent Lenovo ThinkPad have scored 9s and 10s. Apple, which has prioritised thinner and lighter devices over the past decade, has made its products harder to repair.

 


Apple did not immediately respond to a request for comment. Wiens said one of the reasons is that MacBook Neo’s 8 gigabytes of DRAM are directly soldered to the circuit board of the machine as part of a package with the machine’s main processing chip, which is similar to all of Apple’s Mac designs in recent years but will make MacBook Neos impossible to easily upgrade with more memory.

 


Wiens said that could make it hard for the MacBook Neo to run artificial intelligence applications as they grow in complexity in the coming years, even as Apple has publicly cited the privacy benefits of running those applications on a laptop instead of in the cloud. He said Apple could improve its offerings by including an additional layer of memory chips that users can upgrade.

 


“Apple’s future for privacy-centred AI has to be local models,” Wiens said. “I would argue this is a flaw across Apple’s entire Mac product line.”

 


(Only the headline and picture of this report may have been reworked by the Business Standard staff; the rest of the content is auto-generated from a syndicated feed.)

 

 



Source link

Samsung's Choi says AI revolution is unavoidable, urges preparation

Samsung's Choi says AI revolution is unavoidable, urges preparation



The Artificial Intelligence (AI) revolution is ‘unavoidable’, and it is going to impact every aspect of human life, and people need to be prepared for this transformation, said Samsung Electronics Mobile Experience Business COO Won-Joon Choi.


This will be different from the previous revolution that happened through the internet and mobile, which impacted human life, said Choi in a media roundtable here recently, after the launch of S26.


“AI is different in the sense that it will impact every area, you name it, medical, law, everything. Everything will change through AI. It’s not just the IT sector. I think we need to be prepared. Not just IT companies, but legal, HR, medical, everything. Everyone needs to be prepared,” he said.

 


Choi also acknowledged the large investments being made by big tech companies into AI research and development.


“I think there can be some adjustments along the way, but at the macro level, we still need to invest. We still need to bring resources. And we still need to do a lot of research. I think it’s still the beginning, in my opinion. It’s not done,” he said.


While replying to a query about whether excessive marketing of AI features is creating a sense of unease among consumers, Choi said that every new technology looks overwhelming at first.


“I think as time goes on and as these AI features get integrated more deeply into the device and UI applications, people will not feel this (AI) is special, and we would not need to even call this AI. I think people will just naturally accept this is something that I can use, something that can help. I think that time will come,” he said.


Choi said Samsung is working on three big areas on AI, which include the democratisation of the technology.


As part of Samsung’s plans to bring AI features to 800 million devices by the end of this year, double from 400 million devices last year.


Samsung has adopted a Hybrid AI model, which is a combination of on-device and on-cloud. Choi said Samsung is open to integrating third-party AI solutions into its devices.


“We announced the partnership with Perplexity in addition to Gemini. At the end of the day, our goal is to provide the best AI technologies or solutions to Galaxy users,” he said.


Samsung is also ensuring that sensitive information is stored on devices, not in the cloud, to ensure privacy and security in the era of AI.


“We want to make sure that when you use Galaxy AI, you have peace of mind,” he said, adding, “At the platform level, at the hardware and software and even the application layer, we try to provide a mechanism so that people know how they are using AI and how all this information is being used. We will be very transparent and build the core hardware functionality.” 
Over Samsung’s future objectives, Choi said the brand is working to expand the support for local languages and is looking to partner with more brands for the newly-launched Agentic AI play.



Source link

India must build SLMs alongside LLMs to ensure linguistic inclusion: PSA

India must build SLMs alongside LLMs to ensure linguistic inclusion: PSA



India needs to place special emphasis on developing small language models (SLMs) with multimodal capabilities, alongside indigenous large language models (LLMs), to ensure linguistic inclusion, affordability, energy efficiency and public-sector suitability, the Office of the Principal Scientific Advisor (PSA) has suggested.

 


SLMs, which are more economical to train and run, are focused, domain-oriented models that can be fine-tuned for sector-specific tasks in agriculture, healthcare, education, climate and urban governance, and must therefore be developed in consonance with LLMs, the PSA’s office suggested in a white paper.

 


Further, the development of indigenous LLMs is crucial to build AI systems that are less biased, more trustworthy and remain locally relevant in a globally competitive AI ecosystem, the white paper suggested.

 
 


This can be achieved by allowing the development of indigenous LLMs trained on more diverse data, designed for India’s linguistic and social diversity, and governed through national frameworks, the PSA’s office suggested.

 


“Relying solely on foreign models risks under-representation of Indian languages and cultural contexts. Any biases in these models can cascade across all downstream applications that rely on them. This makes it critical to have a policy focus on these systems,” the PSA’s office suggested in the white paper.

 


At present, the central government has approved proposals from a dozen startups to develop indigenous LLMs and SLMs. These include proposals from Sarvam, which is developing a sovereign 105-billion-parameter LLM, referenced alongside 30-billion-parameter models designed for Indian languages, with a focus on governance, public service and high-stakes deployment.

 


Other proposals include an Indian Institute of Technology Bombay-led consortium, BharatGen, which is developing multilingual and multimodal AI models ranging from 2 billion to 1 trillion parameters.

 


On the other hand, Soket AI is developing a 120-billion-parameter open-source multilingual foundation model tailored for India’s linguistic diversity, while Gan AI is developing a 70-billion-parameter AI model targeting high-performance (“superhuman”) text-to-speech capabilities.



Source link

YouTube
Instagram
WhatsApp