Samsung embraces chip-making tech favoured by SK Hynix amid intensifying competition - Business Guardian
Connect with us


Samsung embraces chip-making tech favoured by SK Hynix amid intensifying competition



Samsung Electronics plans to use a chip making technology championed by rival SK Hynix, five people said, as the world’s top memory chipmaker seeks to catch up in the race to produce high-end chips used to power artificial intelligence.

The demand for high bandwidth memory (HBM) chips has boomed with the growing popularity of generative AI. But Samsung, unlike peers SK Hynix and Micron Technology, has been conspicuous by its absence in any dealmaking with AI chip leader Nvidia to supply latest HBM chips.

One of the reasons Samsung has fallen behind is its decision to stick with chip making technology called non-conductive film (NCF) that causes some production issues, while Hynix switched to the mass reflow moulded underfill (MR-MUF) method to address NCF’s weakness, according to analysts and industry watchers.

Samsung, however, has recently issued purchase orders for chipmaking equipment designed to handle MUF technique, three sources with direct knowledge of the matter said.

“Samsung had to do something to ramp up its HBM (production) yield. Adopting MUF technique is a little bit of swallow-your-pride type thing for Samsung, because it ended up following the technique first used by SK Hynix,” one of the sources said.

Samsung’s HBM3 chip production yields stand at about 10-20 per cent while SK Hynix has secured about 60-70 per cent yield rates for its HBM3 production, according to several analysts.

The HBM3 and HBM3E, the newest versions of HBM chips, are in hot demand. They are bundled with core microprocessor chips to help process vast amounts of data in generative AI.

Samsung is also in talks with material manufacturers, including Japan’s Nagase, to source MUF materials, one source said, adding mass production of the high-end chips using MUF is unlikely to be ready until next year at the earliest, as Samsung needs to run more tests.

The three sources also said Samsung plans to use both CF and MUF techniques for its latest HBM chip.

Samsung said its internally developed NCF technology is an “optimal solution” for HBM products and would be used in its new HBM3E chips. “We are carrying out our HBM3E product business as planned,” Samsung said in a statement.

Nvidia and Nagase declined to comment. All sources spoke on condition of anonymity as the information is not public. Samsung’s plan to use MUF underscores growing pressure it faces in the AI chip race, with the HBM chip market, according to research firm TrendForce, seen more than doubling this year to nearly $9 billion amid AI related demand.

NCF versus MUF

The non-conductive film chip manufacturing technology has been widely used by chipmakers to stack multiple layers of chips in a compact high bandwidth memory chipset, as using thermally compressed thin film helps minimise space between stacked chips. But there are often problems linked to adhesive materials as manufacturing gets complicated as more layers are added.

Samsung says its latest HBM3E chip has 12 chip layers. Chipmakers have been looking for alternatives to address such weaknesses. SK Hynix successfully switched to the mass reflow moulded underfill technique ahead of others, becoming the first vendor to supply HBM3 chips to Nvidia.

SK Hynix’s market share in HBM3 and more advanced HBM products for Nvidia is estimated at above 80 per cent this year, according to Jeff Kim, an analyst at KB Securities. Micron joined the high bandwidth memory chip race last month, announcing that its latest HBM3E chip will be adopted by Nvidia to power the latter’s H200 Tensor chips which will begin shipping in the second quarter.

Samsung’s HBM3 series have not yet passed Nvidia’s qualification for supply deals, according to one of the four sources and another person with knowledge of the discussion.

The Daily Guardian is now on Telegram. Click here to join our channel (@thedailyguardian) and stay updated with the latest headlines.

For the latest news Download The Daily Guardian App.


For equipment upgrades, China eyes fresh $69 bn credit in tech sector



The Hong Kong-based ‘South China Morning Post’ (SCMP) newspaper, China has unveiled a plan to reintroduce two relending mechanisms previously utilized to mitigate the economic effects of Covid-19. The People’s Bank of China (PBOC) will facilitate loans through 21 banks to support small and medium-sized technology firms at an interest rate of 1.75 per cent. These loans can be extended twice, each extension lasting up to one year, as per the report. The move, announced on Sunday, comes amid challenges posed to the Chinese economy by a property crisis and geopolitical tensions with key trading partners. China’s policymakers are aiming to enhance liquidity and bolster confidence in the world’s second-largest economy.

Relending mechanisms these measures will allocate a combined 500 billion yuan (US $69.1 billion) to incentivise loans supporting technological innovation and large-scale equipment upgrades – two sectors that have been explicitly prioritised by the country’s leadership, said the report. The refinancing programme will cover 60 per cent of the principal amount for eligible loans extended to technology-focused small and medium-sized enterprises (SMEs) and can be renewed twice, each time for an additional year.

By the end of last year, the PBOC had 17 active structural support tools with a cumulative outstanding size of 7.5 trillion yuan – equivalent to 16.4 per cent of central bank assets. What are China’s relending programmes? These targeted monetary instruments gained prominence in 2014 when pledged supplementary lending was first utilised to directly provide loans to commercial banks to renovate outdated residential buildings. Among the tools, 13 were introduced as temporary measures during the pandemic to support small businesses, toll roads, private firms, property delivery, logistics, and carbon emissions reduction. Seven of them have already expired.

SCMP said the move has sparked speculation among market participants regarding the extent to which Chinese authorities are willing to implement monetary easing, in light of the US Federal Reserve postponing anticipated interest rate adjustments and the Chinese economy concluding the first quarter of 2024 on a stronger footing. The previous relending mechanism for technology, with a quota of 400 billion yuan (US $55.2 billion), was initiated in April 2022 and has since concluded. Similarly, the earlier equipment renovation program, with a quota of 200 billion yuan (US $27.6 billion), was active from September to December 2022.


This relending programme aligns with Beijing’s guidelines for domestic banks, encouraging them to provide funding for five essential finance categories outlined by President Xi Jinping: technology finance, green finance, inclusive finance, pension finance, and digtal finance.

Furthermore, it corresponds with the objective of large-scale equipment upgrades mentioned during the February meeting of the Central Financial and Economic Affairs Commission. This objective serves dual purposes: Leveraging the country’s substantial fixed-asset investment to stimulate economic growth and advancing its vast manufacturing sector, the Hong Kong-daily said. The relending program and other structural measures aim to aid China amidst ongoing challenges in the property market and fragile investor sentiment.

These hurdles will scrutinize the country’s aspirations to attain a 5 percent economic growth rate this year, as reported by SCMP.

Continue Reading


OpenAI transcribed Google’s YouTube videos to train AI models, says report



OpenAI transcribed more than one million hours of YouTube videos to gather training data for its advanced GPT-4 model, disregarding copyright regulations set by the Google-owned platform. Utilizing an indigenous speech recognition tool named Whisper, supported by Microsoft, OpenAI converted audio from YouTube videos into conversational text. This text served as training material for the AI model behind ChatGPT.

According to the report, makers of ChatGPT internally discussed how the use of YouTube data for training might be against the platform’s policy. The company reportedly opted to use YouTube videos’ data as it had exhausted the reservoir of publicly available data. The report stated that OpenAI’s president, Greg Brockman, personally assisted in selecting videos for transcription. Google prohibits the use of videos posted on YouTube for applications that are “independent” of the video platform.

In a statement to The Verge, OpenAI spokesperson Lindsay Held said that the company uses “unique” datasets for each of its models to “help their understanding of the world.” She added that the company uses “numerous sources including publicly available data and partnerships for non-public data.”

Commenting on the topic, Google spokesperson Matt Bryant told The Verge that Google has “seen unconfirmed reports” related to OpenAI using YouTube videos for training AI models. He added that the streaming platform’s “Terms of Service and robots.txt files prohibit unauthorized scraping or downloading of YouTube content.”

Earlier this week, YouTube CEO Neal Mohan in an interview said that “he has seen reports” related to OpenAI using YouTube videos to train their text-to-video generator Sora. He said that he has no information about the same, but it would be a “clear violation” of the platform’s policies if it did.

Google has also used transcribed texts from YouTube videos for training its AI model Gemini. If true, this violates the copyright to the videos, which belongs to the creator who posts the video to the platform.

The report stated that Google broadened its terms of service to allow the company to be able to use publicly available Google Docs files, restaurant reviews on Google Maps, and more for training AI models.

Continue Reading


Realme targets top spot in Rs 15-25k segment with P series launch



Realme India aims to garner the highest share in the Rs 15,000-25,000 mobile phone segment this year with the launch of a new portfolio of devices under the P series, a senior company official said.

While sharing the plans on the P series, Realme India’s business strategy lead, Tarini Prasad Das, told reporters that in 2024, the company aims to achieve 50 million smartphone sales cumulatively under its partnership with Flipkart.

“We aim to lead the Rs 15,000-25,000 smartphone segment this year overall with the launch of the P series. The P series itself is expected to lead the segment,” Das said.

Realme is among the top five smartphone brands in terms of volume market share. The company had a 12 percent market share in 2023.

The company plans to launch the P series in the second week of April in the sub-Rs 20,000 price range.

Das said that both offline and online sales contribute equally to the overall business of Realme, and the company is looking to consolidate its leadership in the Rs 15,000-25,000 segment with the P series.

Without revealing the specifications of P series smartphones, Das said that it will focus on performance, design, better display, and charging facility in the target segment.

“We have exclusively partnered with Flipkart for the P series. We aim to achieve the 50 million sales milestone on Flipkart with the launch of the new P series this year,” Das said.

Realme captured the top spot on Flipkart in the price band of Rs 20,000-30,000 in February 2024 with a 29.2 percent share, as per the Counterpoint February 2024 trend report.

Continue Reading


Microsoft warns of AI misuse in Global polls by China



Countries like India, South Korea and the US, tech giant Microsoft has warned. Voting for 543 Lok Sabha seats in India will take place.

Microsoft has cautioned that China may utilize AI-generated content on social media platforms to influence public opinion and advance its geopolitical agenda during elections in nations such as India, South Korea, and the US. The Lok Sabha elections in India are scheduled to be conducted in seven phases between April 19 and June 4, involving the voting for 543 parliamentary seats. South Korea is set to hold its general election on April 10, while the US is gearing up for its Presidential election on November 5.

“With major elections taking place around the world this year, particularly in India, South Korea, and the United States, we assess that China will, at a minimum, create and amplify AI-generated content to benefit its interests,” Clint Watts, General Manager of Microsoft Threat Analysis Center, stated in a blog post.

Despite the chances of such content affecting election results remaining low, China’s increasing experimentation in augmenting memes, videos, and audio will likely continue and may prove more effective down the line, he said.

China will do it along with North Korea, he wrote.

These are among the Microsoft Threat Intelligence insights in the latest East Asia report published on Wednesday by the Microsoft Threat Analysis Center (MTAC).

China is using fake social media accounts to poll voters on what divides them most to sow division and possibly influence the outcome of the US presidential election in its favour.

China has also increased its use of AI-generated content to further its goals around the world.

Continue Reading


Google tests AI email replies in Gmail Android App



Google is reportedly testing Gemini AI integration into the Gmail app for Android. According to a report by Android Central, Google is testing a new feature for the Gmail Android app that will allow users to use AI-generated replies to respond to emails.

Google has already made Gemini Advanced-powered AI tools available on its services like Google Drive and Docs through Google One AI Premium subscription. It appears that the company is now targeting AI integration into the mobile application of its services.

Gemini AI-powered tools are also available on Gmail. Google One AI subscribers can get suggestions for writing an email, although this functionality is only available in beta version.

According to the report, once the new feature gets enabled, Gemini will be able to analyze the content of the email that the user has received and come up with suggestions. The AI-powered tool will provide users with three potential replies that make sense in the context of the received email. The report states that in certain scenarios, the AI-generated responses can be as long as full sentences or multiple sentences. The user can then tap on any one of the suggested replies, which then moves automatically to the compose field. The message can then be edited or sent as it is.

The feature is currently available to select users; however, it is expected to roll out soon to Google One AI subscribers globally.

In related news, Google is reportedly planning to make the Search Generative Experience (SGE), its generative AI-powered search tool, exclusive to the paid-tier. According to a report by the Financial Times, Google is considering charging for generative AI features for Search, including the experimental AI-powered search service. According to the report, Google’s traditional search service will remain free to use, and the generative AI search experience will be added to its One AI Premium subscription plan. However, it is likely that the company will run ads in search results, irrespective of the subscription.

Continue Reading


Google restricts generative AI search to paid tier



Google is reportedly considering making its Search Generative Experience (SGE), a generative AI-powered search tool, exclusive to its paid-tier services. According to a report by the Financial Times, Google is contemplating charging for access to generative AI features in Search, including the experimental AI-powered search service.

The traditional search service provided by Google will continue to be free to use, while the generative AI search experience may be added to its premium subscription services. These premium subscriptions already offer access to advanced features like Gemini Advanced and AI-powered tools for services such as Gmail and Docs. However, it’s expected that Google will continue to display ads in search results, regardless of the subscription status.

In response to inquiries from the Financial Times, Google stated, “With our generative AI experiments in Search, we’ve already served billions of queries, and we’re seeing positive Search query growth in all of our major markets. We’re continuing to rapidly improve the product to serve new user needs.” However, Google clarified that there are currently no new announcements regarding this matter. Additionally, Google emphasized that it is not considering making Search ad-free and will instead focus on developing new premium capabilities and services to enhance its subscription offerings across Google.

Google introduced the AI-powered SGE in August last year, and it has been available to users as an opt-in experiment in Google’s Search Lab. Once enabled, SGE provides relevant information directly on the search page, instead of displaying links to multiple websites. It generates an AI-powered overview of key information along with relevant links to the source of information.

Continue Reading