Think everyone is getting bored of waiting for a price sensitive announcementHmm. Very quiet.
This whole 'watch the financials' thing sounds just like the delphic oracle's ambigous answer to King Croesus who asked Should we go to war?Trust the CEO. He said 'watch the financials'. I have implicit trust in him. Follow his word, he stands by it. Long live BRN.
Not hardware though. No way 100b arm chips are going to be made + installed +sold between now and end 2025 (1.5 years).
Oh! By the way, someone said they heard a rumour that BRN had put production of Akida 2 + TeNNs SoC on hold so as not to upset a customer's applecart. I wonder if that was an arm's length deal?Arm to roll out 100 biliion devices for AI by 2025
Jun 3, 2024
https://www.facebook.com/sharer.php?u=https://www.gizmochina.com/2024/06/03/arm-to-roll-out-100-biliion-devices-for-ai-by-2025/
Arm Holdings, is aiming to roll 100 billion Arm devices out in 2025. This initiative is part of a broader strategy to establish Arm as a key player in the rapidly expanding AI chip market, projected to grow from $30 billion in 2024 to over $200 billion by 2032.
To achieve this, Arm will establish a dedicated AI chip division. The division aims to produce prototypes by spring 2025 and begin mass production by the fall of the same year. The manufacturing will be managed by contract manufacturers, including potential collaborations with Taiwan Semiconductor Manufacturing Corp (TSMC) and others. The development costs for these AI chips are expected to reach hundreds of billions of yen, with financial backing from Arm and contributions from SoftBank.
Strategically, there is a possibility that the AI chip business could be spun off and integrated under SoftBank once mass production is underway. This move aligns with SoftBank CEO Masayoshi Son’s vision of transforming SoftBank into an AI powerhouse. Son has earmarked a $64 billion investment to drive innovations across various sectors, including data centers, robotics, and power generation. He envisions integrating AI, semiconductor, and robotics technologies to revolutionize industries such as shipping, pharmaceuticals, finance, manufacturing, and logistics.
At the Computex forum in Taipei, Arm highlighted their target of having 100 billion Arm devices worldwide by the end of 2025, emphasizing the role of AI-ready devices. Furthermore, Arm unveiled new AI chip designs and software tools aimed at enhancing AI capabilities in smartphones. These new designs are expected to improve compute and graphics performance by over 30% and accelerate AI inference by 59% for machine learning and computer vision tasks.
Arm’s strategy includes providing ready-for-manufacturing blueprints, a departure from their previous approach of offering abstract designs. This shift aims to expedite the development process for their partners, such as Samsung and TSMC. Samsung is already collaborating with Arm on 3nm process technology to meet the growing demand for generative AI in mobile devices. TSMC is working with Arm to enhance AI performance and efficiency through their Open Innovation Platform ecosystem.
With these initiatives, Arm is not positioning itself as a competitor but as an enabler, helping its customers bring AI-driven chips and devices to market more swiftly. Chris Bergey, Arm’s General Manager, expressed the company’s vision of combining a platform for tightly coupled accelerators with customer Neural Processing Units (NPUs) to foster the rapid development of powerful AI chips and devices.
As Arm moves forward with its plans, it aims to capture a significant share of the burgeoning AI chip market, capitalizing on the unmet demand currently dominated by Nvidia. The company’s focus on AI chip development is a pivotal step in SoftBank’s broader strategy to lead the AI revolution and transform various industries through cutting-edge technology.
Extract
View attachment 64344
Imagine for a moment if this did include our IP. Even at a measly $1 per item...............................Oh! By the way, someone said they heard a rumour that BRN had put production of Akida 2 + TeNNs SoC on hold so as not to upset a customer's applecart. I wonder if that was an arm's length deal?
PS: "Applecart" - Freudian slip re mobile phones?
Even arm does not get that much royality from those devices. 100 billion dollars of royality will make arm atleast one trillion dollar company.Imagine for a moment if this did include our IP. Even at a measly $1 per item...............................
View attachment 64348
View attachment 64349
AMD CEO: Pricier Ryzens will probably get more powerful NPUs
AMD chief executive, Lisa Su, answers questions about how NPUs will scale, if we'll see Ryzens with more cores than there are now, and more at Computex.www.pcworld.com
AMD CEO: Pricier Ryzens will probably get more powerful NPUs
But will you see Ryzens with more than sixteen cores? It doesn't sound that likely, CEO Lisa Su says.
By Mark Hachman
Senior Editor JUN 2, 2024 10:39 PM PDT
Image: Mark Hachman / IDG
You’d expect that more expensive PC processors would feature faster clock speeds, more cache, more powerful integrated graphics. Future Ryzens will probably scale NPU TOPS, too.
“AI will be everywhere.” If there was a theme to Dr. Lisa Su’s post-keynote press conference, at the Computex 2024 show in Taipei, it was that. Su touched on the theme time and time again. And AI will be everywhere across AMD’s CPU lineup — but not necessarily with the same potency.
On stage, Su joked with Microsoft’s Windows and devices chief Pavan Davuluri that an NPU’s TOPS don’t come for free, echoing the old adage that whatever hardware a chipmaker builds, software will suck it up.
“That’s why I was sort of kidding with Pavan on stage,” Su said, in response to my question. “Nothing is for free, when you look at these products both from an overall power standpoint, as well as this overall cost standpoint.
“I think what we’re seeing is AI will truly be everywhere. Our expectation is that the current Copilot+ and (AMD Ryzen AI 300 Copilot+ PCs, or “Strix Point,”) at 50+ TOPS will start more at the higher end of the stack,” Su added. “But we would expect that you will see AI throughout our entire stack as we go forward.
“You’re going to see at the top end that we’re going to continue to scale the TOPS because we are big, big believers in the more local TOPS you have, the more capable your AI PCs are going to be,” Su concluded. “We believe people are going to value that and so it’s worth it, to put it on chip locally.”
Su didn’t describe how AMD will differentiate various Ryzens with NPU capabilities. But there’s a history here: In 2021, AMD mixed and matched parts from various Zen generations under the Ryzen 5000 name. AMD could conceivably do the same with future Ryzens, taking older NPUs and combining them with various CPUs and GPUs.
But that’s not to say we could see just an NPU, either. In response to another question about whether AMD could develop a neuromorphic chip like Intel’s Loihi, Su seemed open to the possibility. “I think as we go forward, we always look at some specific types of what’s called new acceleration technologies,” she said. “I think we could see some of these going forward.”
But could Ryzens get bigger, with more cores? Su demurred while answering a question about whether AMD would ever go beyond the current 16-core count. “There’s no physical reason that we couldn’t go more than 16 cores,” she said.
Su pointed out that software developers don’t always use all the core AMD already provides. “I think the key is just going at the pace that the software guys can utilize them,” she said.
View attachment 64348
View attachment 64349
AMD CEO: Pricier Ryzens will probably get more powerful NPUs
AMD chief executive, Lisa Su, answers questions about how NPUs will scale, if we'll see Ryzens with more cores than there are now, and more at Computex.www.pcworld.com
AMD CEO: Pricier Ryzens will probably get more powerful NPUs
But will you see Ryzens with more than sixteen cores? It doesn't sound that likely, CEO Lisa Su says.
By Mark Hachman
Senior Editor JUN 2, 2024 10:39 PM PDT
Image: Mark Hachman / IDG
You’d expect that more expensive PC processors would feature faster clock speeds, more cache, more powerful integrated graphics. Future Ryzens will probably scale NPU TOPS, too.
“AI will be everywhere.” If there was a theme to Dr. Lisa Su’s post-keynote press conference, at the Computex 2024 show in Taipei, it was that. Su touched on the theme time and time again. And AI will be everywhere across AMD’s CPU lineup — but not necessarily with the same potency.
On stage, Su joked with Microsoft’s Windows and devices chief Pavan Davuluri that an NPU’s TOPS don’t come for free, echoing the old adage that whatever hardware a chipmaker builds, software will suck it up.
“That’s why I was sort of kidding with Pavan on stage,” Su said, in response to my question. “Nothing is for free, when you look at these products both from an overall power standpoint, as well as this overall cost standpoint.
“I think what we’re seeing is AI will truly be everywhere. Our expectation is that the current Copilot+ and (AMD Ryzen AI 300 Copilot+ PCs, or “Strix Point,”) at 50+ TOPS will start more at the higher end of the stack,” Su added. “But we would expect that you will see AI throughout our entire stack as we go forward.
“You’re going to see at the top end that we’re going to continue to scale the TOPS because we are big, big believers in the more local TOPS you have, the more capable your AI PCs are going to be,” Su concluded. “We believe people are going to value that and so it’s worth it, to put it on chip locally.”
Su didn’t describe how AMD will differentiate various Ryzens with NPU capabilities. But there’s a history here: In 2021, AMD mixed and matched parts from various Zen generations under the Ryzen 5000 name. AMD could conceivably do the same with future Ryzens, taking older NPUs and combining them with various CPUs and GPUs.
But that’s not to say we could see just an NPU, either. In response to another question about whether AMD could develop a neuromorphic chip like Intel’s Loihi, Su seemed open to the possibility. “I think as we go forward, we always look at some specific types of what’s called new acceleration technologies,” she said. “I think we could see some of these going forward.”
But could Ryzens get bigger, with more cores? Su demurred while answering a question about whether AMD would ever go beyond the current 16-core count. “There’s no physical reason that we couldn’t go more than 16 cores,” she said.
Su pointed out that software developers don’t always use all the core AMD already provides. “I think the key is just going at the pace that the software guys can utilize them,” she said.
"the design of MB.OS demands a different approach because we are decoupling the hardware and software innovation cycles and integration steps. This will make software development and integration much faster, and it also facilitates the constant flow of innovation into the vehicle, resulting in better products for our customers."