VeriSilicon's Scalable High-Performance GPGPU-AI Computing IPs Empower Automotive and Edge Server AI Solutions
Article content
Provide AI acceleration with high computing density, multi-chip scaling, and 3D-stacked memory integration
Article content
SHANGHAI — VeriSilicon (688521.SH) today announced the latest advancements in its high-performance and scalable GPGPU-AI computing IPs, which are now empowering next-generation automotive electronics and edge server applications. Combining programmable parallel computing with a dedicated Artificial Intelligence (AI) accelerator, these IPs offer exceptional computing density for demanding AI workloads such as Large Language Model (LLM) inference, multimodal perception, and real-time decision-making in thermally and power-constrained environments.
Article content
VeriSilicon's GPGPU-AI computing IPs are based on a high-performance General Purpose Graphics Processing Unit (GPGPU) architecture with an integrated dedicated AI accelerator, delivering outstanding computing capabilities to AI applications. The programmable AI accelerator and sparsity-aware computing engine accelerate transformer-based and matrix-intensive models through advanced scheduling techniques. These IPs also support a broad range of data formats for mixed-precision computing, including INT4/8, FP4/8, BF16, FP16/32/64, and TF32, and are designed with high-bandwidth interfaces of 3D-stacked memory, LPDDR5X, HBM, as well as PCIe Gen5/Gen6 and CXL. They are also capable of multi-chip and multi-card scale-out expansion, offering system-level scalability for large-scale AI application deployments.
Article content
VeriSilicon's GPGPU-AI computing IPs provide native support for popular AI frameworks for both training and inference, such as PyTorch, TensorFlow, ONNX, and TVM. These IPs also support General Purpose Computing Language (GPCL) which is compatible with mainstream GPGPU programming languages, and widely used compilers. These capabilities are well aligned with the computing and scalability requirements of today's leading LLMs, including models such as DeepSeek.
Article content
'The demand for AI computing on edge servers, both for inference and incremental training, is growing exponentially. This surge requires not only high efficiency but also strong programmability. VeriSilicon's GPGPU-AI computing processors are architected to tightly integrate GPGPU computing with AI accelerator at fine-grained levels. The advantages of this architecture have already been validated in multiple high-performance AI computing systems,' said Weijin Dai, Chief Strategy Officer, Executive Vice President, and General Manager of the IP Division at VeriSilicon. 'The recent breakthroughs from DeepSeek further amplify the need for maximized AI computing efficiency to address increasingly demanding workloads. Our latest GPGPU-AI computing IPs have been enhanced to efficiently support Mixture-of-Experts (MoE) models and optimize inter-core communication. Through close collaboration with multiple leading AI computing customers, we have extended our architecture to fully leverage the abundant bandwidth offered by 3D-stacked memory technologies. VeriSilicon continues to work hand-in-hand with ecosystem partners to drive real-world mass adoption of these advanced capabilities.'
Article content
Article content
Article content
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Globe and Mail
25 minutes ago
- Globe and Mail
ScanTech AI Systems Appoints Security Industry Veteran Brad Buswell as Chairman of the Board
Atlanta, GA, June 09, 2025 (GLOBE NEWSWIRE) -- ScanTech AI Systems Inc. (the "Company" or "ScanTech AI") (Nasdaq: STAI), a next-generation provider of AI-powered CT screening systems for aviation, customs, and critical infrastructure, today announced the appointment of Brad Buswell as Chairman of the Board of Directors, effective June 9, 2025. Mr. Buswell has served as a member of the Board of Directors since January 2025. 'On behalf of the entire ScanTech AI team, I would like to thank Karl Brenza for his steady leadership and instrumental role in navigating the Company's transformation into a publicly traded enterprise,' said Dolan Falconer, Chief Executive Officer and President at ScanTech AI. 'We are equally honored to welcome Brad Buswell as our new Chairman. Brad brings unparalleled experience in both public and private sector security innovation, and his strategic acumen will be of great value as we continue to expand our market share and deepen our global footprint.' Mr. Buswell is a seasoned executive with more than 30 years of leadership in national security, aviation screening, and advanced detection technologies. He most recently served as Senior Vice President of Security Enterprise Solutions at Leidos (NYSE: LDOS), where he helped drive innovation in aviation and critical infrastructure protection. Prior to Leidos, Mr. Buswell held executive leadership roles, including President and CEO of Morpho Detection and President of Rapiscan North America. His public sector experience includes his appointment as Deputy Under Secretary for Science and Technology at the U.S. Department of Homeland Security, where he oversaw strategic R&D efforts supporting the Transportation Security Administration, Customs and Border Protection, and other DHS components. Mr. Buswell is also a 20-year veteran of the U.S. Navy and holds degrees from the U.S. Naval Academy and George Washington University. 'I am deeply honored by the appointment as Chairman of the Board of Directors at such an exciting juncture in the Company's growth,' said Mr. Buswell. 'ScanTech AI's cutting-edge AI and CT-based technologies are redefining the standards for non-intrusive security screening. I look forward to supporting the executive team and fellow board members as we build long-term value for stockholders, customers, and global partners.' About ScanTech AI ScanTech AI Systems Inc. (Nasdaq: STAI) has developed one of the world's most advanced non-intrusive 'fixed-gantry' CT screening technologies. Utilizing proprietary artificial intelligence and machine learning capabilities, ScanTech AI's state-of-the-art scanners accurately and quickly detect hazardous materials and contraband. Engineered to automatically locate, discriminate, and identify threat materials and items of interest, ScanTech AI's solutions are designed for use in airports, seaports, borders, embassies, corporate headquarters, government and commercial buildings, factories, processing plants, and other facilities where security is a priority. For more information, visit and Forward-Looking Statements This press release contains forward-looking statements within the meaning of Section 27A of the U.S. Securities Act of 1933, as amended (the 'Securities Act'), and Section 21E of the U.S. Securities Exchange Act of 1934, as amended ('Exchange Act'), including statements regarding ScanTech AI's management team's expectations, hopes, beliefs, intentions, plans, prospects or strategies regarding the future, including possible business partnerships, revenue growth and financial performance, product expansion and services. Any statements contained herein that are not statements of historical fact may be deemed to be forward-looking statements. Additionally, any statements that refer to projections, forecasts, or other characterizations of future events or circumstances, including any underlying assumptions, are forward-looking statements. The words 'may,' 'will,' 'could,' 'would,' 'should,' 'expect,' 'intend,' 'plan,' 'anticipate,' 'believe,' 'estimate,' 'predict,' 'project,' 'potential,' 'continue,' 'ongoing,' 'target,' 'seek' or the negative or plural of these words, or other similar expressions that are predictions or indicate future events or prospects, may identify forward-looking statements, but the absence of these words does not mean that a statement is not forward-looking. The forward-looking statements contained in this press release are based on the current expectations and beliefs made by the management of ScanTech AI, in light of their respective experience and their perception of historical trends, current conditions and expected future developments and their potential effect on ScanTech AI, as well as other factors they believe are appropriate under the circumstances. There can be no assurance that future developments affecting ScanTech AI will be those that it has anticipated. These forward-looking statements involve a number of risks, uncertainties (some of which are beyond the control of the parties) or other assumptions that may cause actual results or performance to be materially different from those expressed or implied by these forward-looking statements, including product and service acceptance, regulatory oversights, research and development success, and that ScanTech AI will have sufficient capital to operate as anticipated. Should one or more of these risks of uncertainties materialize, or should any of the assumptions of ScanTech AI prove incorrect, actual results may vary in material respects from those projected in these forward-looking statements. Additional factors that could cause actual results to differ are discussed under the heading 'Risk Factors' and in other sections of the filings of ScanTech AI (and its predecessor, Mars) with the U.S. Securities and Exchange Commission (the 'SEC'), and in the current and periodic reports filed or furnished by ScanTech AI (and its predecessor, Mars) from time to time with the SEC. All forward-looking statements in this press release are made as of the date hereof, based on the information available to ScanTech AI as of the date hereof, and ScanTech AI assumes no obligation to update any forward-looking statement, whether as a result of new information, future events or otherwise, except as may otherwise be required under applicable securities laws. Attachment Brad Buswell, Chairman of the Board of Directors
Globe and Mail
25 minutes ago
- Globe and Mail
Pre-Markets Flat Ahead of CPI, PPI & Possible Trade Deals
Monday, June 9, 2025 In terms of scheduled economic reports, we're starting off a new trading week slowly. Pre-market futures are up marginally across the board — Dow +44 points, S&P 500 +9 and the Nasdaq +17 points. Overall, these indexes are roughly flat year over year, although the S&P 500 grew above 6K for the first time since late February. Trade negotiations between China and the U.S. have remained ongoing, and while we have no new details emerging at this hour, we still see a pending rare-earth minerals deal between China and the top three automakers in the U.S., which will help facilitate auto building and deliveries this year. That said, a comprehensive trade deal between the world's two biggest economies remains elusive. The current deadline before announced tariffs manifest themselves is July 9th. What to Expect from the Stock Market This Week Our week's biggest market data comes mid-week, when on Wednesday morning the May print for Consumer Price Index (CPI) is expected to remain steady at +0.2%, with year-over-year up 20 basis points (bps) from a month ago to +2.5%. On the core side — stripping out volatile food and energy prices — analysts see headline up 10 bps to +0.3%, +2.9% on core year over year, also up 10 bps. Year-over-year headline CPI is also known as the 'Inflation Rate,' and April's +2.3% marks the low point over the past year. We hadn't been lower on this metric since way back in February on 2021, at the foothills of the Great Reopening. This is as close as we have been to the Fed's optimal +2% inflation rate. May's Producer Price Index (PPI) comes out Thursday morning, and this is expected to revert back to positive +0.2% from April's multi-year low -0.5%. (This was the deepest cut to PPI since April 2020 — near ground-zero of the Covid pandemic.) A +0.2% gets us back to where we were in February of this year. Year-over-year PPI on headline reached +2.4% in April, with core year over year +3.1%. This sounds bad relative to the near-2% levels of these other price indexes, but keep in mind core PPI year over year back in March was up to +4.0% — notching a 2-year high. Much of this information contains some tariff 'noise' that makes these indexes harder to view clearly. Earnings Reports Continue This Week While we acknowledge that calendar Q1 earnings season has wrapped up, are earnings reports ever really 'done'? This week, on Wednesday afternoon, brings us Oracle ORCL results for fiscal Q4, which expects modest earnings growth on +8.8% in quarterly revenues. Thursday after the bell we'll hear from Adobe Systems ADBE and RH RH, formerly Restoration Hardware. Adobe is expected to fetch double-digit earnings growth, while RH's loss per share is expected to come down considerably. (You can see the full Zacks Earnings Calendar here.) Questions or comments about this article and/or author? Click here>> Zacks Names #1 Semiconductor Stock It's only 1/9,000th the size of NVIDIA which skyrocketed more than +800% since we recommended it. NVIDIA is still strong, but our new top chip stock has much more room to boom. With strong earnings growth and an expanding customer base, it's positioned to feed the rampant demand for Artificial Intelligence, Machine Learning, and Internet of Things. Global semiconductor manufacturing is projected to explode from $452 billion in 2021 to $803 billion by 2028. See This Stock Now for Free >> Want the latest recommendations from Zacks Investment Research? Today, you can download 7 Best Stocks for the Next 30 Days. Click to get this free report Adobe Inc. (ADBE): Free Stock Analysis Report RH (RH): Free Stock Analysis Report
CTV News
38 minutes ago
- CTV News
Getty Images and Stability AI face off in British copyright trial that will test AI industry
The desktop and mobile websites for Stable Diffusion are pictured, Oct. 24, 2023, in New York. (AP Photo/John Minchillo, File) LONDON — Getty Images is facing off against artificial intelligence company Stability AI in a London courtroom for the first major copyright trial of the generative AI industry. Opening arguments before a judge at the British High Court began on Monday. The trial could last for three weeks. Stability, based in London, owns a widely used AI image-making tool that sparked enthusiasm for the instant creation of AI artwork and photorealistic images upon its release in August 2022. OpenAI introduced its surprise hit chatbot ChatGPT three months later. Seattle-based Getty has argued that the development of the AI image maker, called Stable Diffusion, involved 'brazen infringement' of Getty's photography collection 'on a staggering scale.' Tech companies have long argued that 'fair use' or 'fair dealing' legal doctrines in the United States and United Kingdom allow them to train their AI systems on large troves of writings or images. Getty was among the first to challenge those practices when it filed copyright infringement lawsuits in the United States and the United Kingdom in early 2023. 'What Stability did was inappropriate,' Getty CEO Craig Peters told The Associated Press in 2023. He said creators of intellectual property should be asked for permission before their works are fed into AI systems rather than having to participate in an 'opt-out regime.' Getty's legal team told the court Monday that its position is that the case isn't a battle between the creative and technology industries and that the two can still work together in 'synergistic harmony' because licensing creative works is critical to AI's success. 'The problem is when AI companies such as Stability AI want to use those works without payment,' Getty's trial lawyer, Lindsay Lane, said. She said the case was about 'straightforward enforcement of intellectual property rights,' including copyright, trademark and database rights. Getty Images 'recognizes that the AI industry is a force for good but that doesn't justify those developing AI models to ride roughshod over intellectual property rights,' Lane said. Stability AI had a 'voracious appetite' for images to train its AI model, but the company was 'completely indifferent to the nature of those works,' Lane said. Stability didn't care if images were protected by copyright, had watermarks, were not safe for work or were pornographic and just wanted to get its model to the market as soon as possible, Lane said. 'This trial is the day of reckoning for that approach,' she said. Stability has argued that the case doesn't belong in the United Kingdom because the training of the AI model technically happened elsewhere, on computers run by U.S. tech giant Amazon. Similar cases in the U.S. have not yet gone to trial. In the years after introducing its open-source technology, Stability struggled to capitalize on the popularity of the tool, battling lawsuits, misuse and other business problems. Stable Diffusion's roots trace to Germany, where computer scientists at Ludwig Maximilian University of Munich worked with the New York-based tech company Runway to develop the original algorithms. The university researchers credited Stability AI for providing the servers that trained the models, which require large amounts of computing power. Stability later blamed Runway for releasing an early version of Stable Diffusion that was used to produce abusive sexual images, but also said it would have exclusive control of more recent versions of the AI model. Stability last year announced what it described as a 'significant' infusion of money from new investors including Facebook's former president Sean Parker, who is now chair of Stability's board. Parker also has experience in copyright disputes as the co-founder of online music company Napster, which temporarily shuttered in the early 2000s after the record industry and popular rock band Metallica sued over copyright violations. The new investments came after Stability's founding CEO Emad Mostaque quit and several top researchers left to form a new German startup, Black Forest Labs, which makes a competing AI image generator. —— Kelvin Chan And Matt O'Brien, The Associated Press O'Brien contributed to this report from Providence, Rhode Island.
