Top related persons:
Top related locs:
Top related orgs:

Search resuls for: "Stephen N"


25 mentions found


While SoftBank had said in March it planned to list Arm in the U.S. stock market, the company's IPO registration shows that it is pressing ahead with the blockbuster offering despite adverse market conditions. Arm plans to sell its shares on Nasdaq later this year, seeking to raise between $8 billion and $10 billion, the sources said. SoftBank and Arm declined to comment. There are signs that the IPO market is beginning to thaw. Arm's IPO preparations are being led by Goldman Sachs Group Inc (GS.N), JPMorgan Chase & Co (JPM.N), Barclays (BARC.L) and Mizuho Financial Group (8411.T).
During a conference call with investors, the company projected that adjusted gross margins will climb above 40% in the second half, having hit historic lows in the first half of the year. Underscoring Intel's profitability slump in recent years, its first-quarter unadjusted gross margin fell to 34.2%, almost half of its multi-decade high of over 67% in 2010. The company forecast a further drop to an unadjusted gross margin of 33.2% for the second quarter. "While we understand investors may be disappointed in its 2Q23 gross margin outlook, we are confident that Intel's gross margin will recover in 2H23 as the burden of factory underutilization and new product start-up cost diminishes," said Kinngai Chan, analyst at Summit Insights Group. Intel said adjusted losses were 4 cents per share, above analysts' expectations of a 15 cent per share adjusted loss.
Bosch said it plans to invest $1.5 billion to retool TSI's chip production facilities in Roseville, California to start producing silicon carbide chips by 2026. Bosch said the TSI facility would become the "third pillar" of in-house semiconductor production, along with two sites in Germany. The silicon carbide chips Bosch said it will manufacture at the TSI Roseville site are increasingly in demand by electric vehicle manufacturers. U.S-based Wolfspeed Inc (WOLF.N) is building new plants to make silicon carbide chips in New York State and in Germany. Bosch plans to acquire the buildings, machines and infrastructure of TSI as well as well as its commercial semiconductor business.
Many of the automotive industry's biggest chip suppliers such as NXP Semiconductor (NXPI.O) and STMircoelectronics NV (STM.DE) tap TSMC to make their chips. But automotive chips must meet a higher bar for ruggedness and longevity than the chips that go into consumer electronics. TSMC has special manufacturing processes for the automotive industry that typically arrive a couple years after similar processes for consumer chips. In the past it has then taken automotive chip firms extra time to create chip designs for those specialized manufacturing lines. At a conference in Silicon Valley on Wednesday, TSMC unveiled new software that will enable automotive chip designers to start work on their designs about two years sooner.
Atomera said the deal allows STMicro to install Atomera's technology in its factories, and manufacture and distribute chips made with it. Atomera has developed a new material that can slot into existing semiconductor manufacturing processes to improve transistors, the tiny electrical switches at the heart of all chips. Atomera claims its material in some cases can improve transistors as much as jumping to a new-generation technology, but with lower costs. Atomera Chief Executive Scott Bibaud told Reuters that STMicro, with the new deal, is now Atomera's closest customer to moving to mass production. Bibaud said there is no firm timeline for when chips with Atomera technology will be in production, but it could be within 12 to 18 months.
April 25 (Reuters) - Nvidia Corp (NVDA.O), a major supplier of chips and computing systems for artificial intelligence, on Tuesday released a set of software tools aimed at helping chatbots watch their language. But the chatbots can still be unpredictable and say things their creators wish they did not. Nvidia's software tools, provided free of charge, are designed to help companies guard against unwanted responses from chatbots. U.S. lawmakers have called for regulations around AI systems as apps like ChatGPT have surged in popularity. Few legal rules or industry standards exist on how to make AI systems safe.
Generative AI gobbles up reams of computing power, amplifying the urgency of Meta's capacity scramble, said five of the sources. Executives also that spring set about reorganizing Meta's AI units, naming two new heads of engineering in the process, including Janardhan, the author of the September memo. According to four of the sources, Meta did not prioritize building generative AI products until after the launch of ChatGPT in November. Zuckerberg announced a new top-level generative AI team in February that he said would "turbocharge" the company's work in the area. Carvill, the Meta spokesperson, said the company has been building generative AI products on different teams for more than a year.
April 20 (Reuters) - Two companies that make tools for manufacturing chips said that they expect sales to China to boom later this year despite U.S. export restrictions on equipment used to make cutting-edge semiconductors. Lam is subject to the U.S. export restrictions, and ASML will face new rules from the Dutch government on China sales later this year. But those rules so far have only affected equipment used in making the most advanced chips. In Lam's case, it had originally estimated that the China restrictions would cost it between $2 billion and $2.5 billion of revenue in 2023. That's a jump from November, when ASML said China made up 18% of its then 38 billion euro backlog.
April 18 (Reuters) - Broadcom Inc (AVGO.O) on Tuesday released a new chip for wiring together supercomputers for artificial intelligence (AI) work using networking technology that is already in wide use. Broadcom is a major supplier of chips for Ethernet switches, which are the primary way the computers inside conventional data centers are connected to one another. But the rise of AI applications like OpenAI's ChatGPT and Alphabet Inc's (GOOGL.O) Bard have presented new challenges for the networks inside data centers. Broadcom on Tuesday announced a new chip, Jericho3-AI, which can connect up to 32,000 GPU chips together. The Jericho3-AI chip will compete with another supercomputer networking technology called InfiniBand.
[1/2] An example of Adobe Inc generative artificial intelligence features for video editing is seen in this undated screenshot of the company's software obtained by Reuters on April 13, 2023. April 17 (Reuters) - Adobe Inc (ADBE.O) on Monday said it plans to introduce new artificial intelligence (AI) features into its video editing software used by the film and television industries. The tools build on a new system called Adobe Firefly that the company introduced last month for generating still images and text. But after Getty Images sued Stability AI alleging that the startup misused Getty's copyrighted images in training its AI system, legal questions clouded whether the output of such AI systems can be used in commercial work. Adobe is testing the system now and plans to release the video tools later this year.
Amazon Web Services (AWS), the world's biggest cloud computing provider, on Thursday jumped into that race with a suite of its own proprietary AI technologies, but it is taking a different approach. ChatGPT creator OpenAI, for example, offers a similar service, letting customers fine-tune the models behind ChatGPT to create a custom chatbot. The Bedrock service lets AWS customers test-drive those technologies without having to deal with the underlying data center servers that power them. "It's unneeded complexity from the perspective of the user," Vasi Philomin, vice president of generative AI at AWS, told Reuters. Those underlying servers will use a mix of Amazon's own custom AI chips as well as chips from Nvidia Corp (NVDA.O), the biggest supplier of chips for AI work but whose chips have been in tight supply this year.
April 12 (Reuters) - Intel Corp (INTC.O) on Wednesday said its chip contract manufacturing division will work with U.K.-based chip designer Arm Ltd to ensure that mobile phone chips and other products that use Arm's technology can be made in Intel's factories. Intel's turnaround strategy hinges in part on opening up its factories to other chip companies, particularly those in mobile phones. It has said firms such as Qualcomm Inc (QCOM.O) are planning to use its factories for future chip designs. For its part, Arm, owned by Japanese technology investor SoftBank Group Corp (9984.T) and which plans to go public later this year, is a major supplier of intellectual property to many chip companies, especially in mobile phones. Arm has partnerships with major chip contract manufacturers in place to ensure that its designs will work well on their manufacturing processes.
The new RTX 4070 chip, which Nvidia will start shipping on Thursday, will cost $599, putting it near the middle of the company's range of graphics processing units (GPUs), which list for up to $1,600. The chip that it updates, the RTX 3060, is the fourth most popular gaming chip on the market, according to survey data in March from game distribution platform Steam. Nvidia's chips help PC video games render images on to high-resolution screens more quickly to make games look more realistic. The 4070 chips will be the cheapest available that use Nvidia's newest AI technology to do so. I have to really understand the motion between the two frames," Justin Walker, senior director of Nvidia's GeForce products, said in an interview.
Databricks sells software tools for building AI systems. Ghodsi told Reuters that the company is releasing the free training data in the hope that other companies will use it to make their own AI systems, possibly using Databricks to do so. But it could not be used in commercial products because the data used to train the model was generated by OpenAI's ChatGPT, whose terms of service forbid using its data to develop commercial AI systems that could compete with OpenAI. Using data generated by AI to train other AI systems has become common. Users will be able to examine the training data themselves, which they cannot do for models such as ChatGPT or Alphabet Inc's (GOOGL.O) Bard, whose training data wasn't released.
Companies Apple Inc FollowApril 5(Reuters) - Apple Inc (AAPL.O) on Wednesday said that more of its supply chain is committing to use renewable energy in producing the company's iPhones, Macs and other products. Apple said suppliers are supporting 13 gigawatts of active renewable energy projects, up from 10 gigawatts last year. Apple suppliers have made commitments to support an eventual total of 20 gigawatts of projects, up from a total of 16 gigawatts a year before, the company said. Apple on Wednesday said that 250 of its suppliers have pledged to use renewable energy for Apple production, up from 213 suppliers the year before. In China, 70 suppliers have made clean energy pledges for Apple production, up from 55 last year, the company said.
Google has designed its own custom chip called the Tensor Processing Unit, or TPU. The Google TPU is now in its fourth generation. Google said its supercomputers make it easy to reconfigure connections between chips on the fly, helping avoid problems and tweak for performance gains. Google said that startup Midjourney used the system to train its model, which generates fresh images after being fed a few words of text. Google said it did not compare its fourth-generation to Nvidia's current flagship H100 chip because the H100 came to the market after Google's chip and is made with newer technology.
Google has designed its own custom chip called the Tensor Processing Unit, or TPU. The Google TPU is now in its fourth generation. Google said its supercomputers make it easy to reconfigure connections between chips on the fly, helping avoid problems and tweak for performance gains. "Circuit switching makes it easy to route around failed components," Google Fellow Norm Jouppi and Google Distinguished Engineer David Patterson wrote in a blog post about the system. Google said that startup Midjourney used the system to train its model, which generates fresh images after being fed a few words of text.
Nvidia dominates the market for training AI models with huge amounts of data. Qualcomm's chips hit 197.6 server queries per watt versus 108.4 queries per watt for Nvidia. Neuchips, a startup founded by veteran Taiwanese chip academic Youn-Long Lin, took the top spot with 227 queries per watt. Qualcomm also beat Nvidia at object detection with a score of 3.2 queries per watt versus Nvidia's 2.4 queries per watt. Nvidia hit 10.8 samples per watt, while Neuchips ranked second at 8.9 samples per watt and Qualcomm was in third place at 7.5 samples per watt.
April 3 (Reuters) - Micron Technology Inc (MU.O) on Monday said that its business operations in China are normal while it is cooperating with a Chinese government cybersecurity review of its products. Last week, the Cyberspace Administration of China said it would conduct a security review of Micron's products sold in the country. The move comes amid a deepening rift between the United States and China over chip technology that has left companies caught in the crossfire. Micron is the only U.S.-based player in the global market for memory chips and is building a new $15 billion factory in upstate New York. Chinese companies have also been working to break into the memory market, but the U.S. last year restricted export of chipmaking tools to the country.
Synopsys spreads AI throughout its chip design tools
  + stars: | 2023-03-29 | by ( Stephen Nellis | ) www.reuters.com   time to read: +2 min
Synopsys first released an AI tool for one part of the chip design process three years ago, and with customers like Samsung Electronics Co Ltd (005930.KS) and ST Microelectronics (STM.DE) using the system. The tools Synopsys released Wednesday at its annual user conference in Santa Clara, California, spread much further across the chip design process. Synopsys is in a race with Cadence Design Systems (CDNS.O), its largest competitor, to add AI to chip design software. While some of the Synopsys tools released Wednesday are catching up to Cadence, Karl Freund, principal analyst with Cambrian AI research, said Synopsys is ahead, with more than 100 chips by customers using its AI tools coming to market. "I think they'll probably be at 1,000 (completed chip designs) by the end of the year."
[1/2] A smartphone with a displayed NVIDIA logo is placed on a computer motherboard in this illustration taken March 6, 2023. REUTERS/Dado Ruvic/IllustrationMarch 27 (Reuters) - Nvidia Corp (NVDA.O), the world's leading designer of computer chips used in creating artificial intelligence, on Monday showed new research that explains how AI can be used to improve chip design. Chip design engineers use complex design software from firms like Synopsys Inc (SNPS.O) and Cadence Design Systems Inc (CDNS.O) to help them optimize the placement of those transistors. On Monday, Nvidia released a paper showing that it could use a combination of artificial intelligence techniques to find better ways to place big groups of transistors. The Nvidia research took an existing effort developed by University of Texas researchers using what is called reinforcement learning and added a second layer of artificial intelligence on top of it to get even better results.
RUTHERFORD, California, March 27 (Reuters) - The head of the U.S. National Security Agency's cybersecurity directorate on Monday said TikTok represents a "strategic issue" rather than an immediate "tactical" threat to the United States. Joyce said China could use its influence on TikTok to suppress information that might make the nation look bad to Americans. His remarks echo earlier warnings by Federal Bureau of Investigation Director Christopher Wray and National Security Agency Director Paul Nakasone, who told U.S. lawmakers that TikTok could be used to carry out sweeping influence campaigns. U.S. President Joe Biden has signaled his support for legislation introduced by a dozen Senators that would give him the power to ban TikTok and other foreign technologies if they pose a national security risk. Reporting by Stephen Nellis in Rutherford, California; Editing by Chris Reese and Marguerita ChoyOur Standards: The Thomson Reuters Trust Principles.
SAN FRANCISCO, March 24 (Reuters) - Intel Corp (INTC.O) co-founder Gordon Moore, a pioneer in the semiconductor industry whose "Moore's Law" predicted a steady rise in computing power for decades, died Friday at the age of 94, the company announced. Intel and Moore's family philanthropic foundation said he died surrounded by family at his home in Hawaii. In recent years, Intel rivals such as Nvidia Corp (NVDA.O) have contended that Moore's Law no longer holds as improvements in chip manufacturing have slowed down. He went to work at the Shockley Semiconductor Laboratory where he met future Intel cofounder Robert Noyce. In 1968, Moore and Noyce left Fairchild to start the memory chip company soon to be named Intel, an abbreviation of Integrated Electronics.
March 24 (Reuters) - The United States and Canada said on Friday they would work together to create a bilateral semiconductor manufacturing corridor, as International Business Machines (IBM.N) signaled its intent to expand in Canada. The Canadian government will spend C$250 million ($181.94 million) on its domestic semiconductor industry to boost research and development and manufacturing, the prime minister's office said in a statement. The countries did not disclose the amount of IBM's investment in Canada. IBM maintains semiconductor research and manufacturing operations in upstate New York. ($1 = 1.3741 Canadian dollars)Reporting by Stephen Nellis in San Francisco; Editing by Richard ChangOur Standards: The Thomson Reuters Trust Principles.
The code is an AI model, an algorithm that is trained on sets of data and can then learn from new data to perform a variety of tasks. Databricks CEO Ali Ghodsi said the release was aimed at demonstrating a viable alternative to training a kind of AI model called a large language model with enormous resources and computing power. OpenAI, valued at $29 billion, trains its AI models with huge troves of data on a supercomputer from investor Microsoft Corp (MSFT.O). Databricks wants enterprises to train their own AI models using its software. "My belief is that in the end, you will make these models smaller, smaller and smaller, and they will be open-sourced," Ghodsi said.
Total: 25