Edge & On-device AI Silicon

ASUS announced that the ASUS Googlebook 14 (CX9406CAA) is now available for purchase in the United States, priced at $1,299.00, online at the ASUS Store and Best Buy. The 14-inch ultraportable is the first-ever ASUS laptop designed for Gemini Intelligence, built on a shared Android foundation for seamless integration with Android phones. It weighs just 2.18lbs and measures 0.43 inches thin, with a CNC-machined magnesium-aluminum chassis finished in Dusk Ash and protected by Nano Ceramic Technology, tested to US military-grade MIL-STD 810H and ASUS Superior Durability standards. The laptop runs an Intel Core Ultra 5 325 processor with 16GB LPDDR5X memory and up to a 512GB PCIe 4.0 SSD, and features a 14-inch 16:10 3K OLED HDR touchscreen with up to 1000 nits peak HDR brightness and a 120Hz refresh rate, a six-speaker audio system, and a 70Wh battery rated at up to 18 hours of battery life. Every purchase includes up to $500 in additional value through a curated bundle of apps and services, most notably 12 months of Google AI Pro and 5TB of cloud storage, plus three months of YouTube Premium, Adobe Premiere Pro, Adobe Photoshop, and CapCut Pro, along with 12 months of Luminar and GeForce NOW.

Qualcomm Incorporated unveiled new Snapdragon platforms designed to run AI agents directly on consumer devices at its Snapdragon Summit, as the company positions agentic AI as its next growth phase beyond smartphones. CFO Akash Palkhiwala said the two new chips target the high-end smartphone market with best-in-class gaming and productivity performance and set up the platform for agentic AI. The company is expanding across AI-powered devices, automotive systems, data centers, and physical AI applications, and earlier this month announced a multi-generation collaboration with Amazon to develop customised silicon for large-scale AI data centers focused on AI inference. Qualcomm's forward GAAP P/E of 15.59x sits 6.76% below its 5-year average of 16.72x and below the sector's forward GAAP P/E of 29.51. Hedge fund ownership slipped from 78 funds at the end of Q4 2025 to 71 at the end of Q1 2026, and short interest reached 3.53% of the float as of August 31, 2026.

MediaTek announced that its flagship Dimensity 9500 platform powers the new Samsung Galaxy Tab S12 Ultra and Samsung Galaxy Tab S12+, bringing flagship performance and on-device intelligence to Samsung's premium tablet lineup. The chip is built on TSMC's 3nm process and features an All Big Core CPU architecture, the MediaTek NPU 990 and Generative AI Engine 2.0, and a flagship 12-core GPU with hardware-accelerated raytracing and Unreal Engine 5.5+ support. Jerry Yu, Corporate Senior Vice President and Head of Smart Life Business Group at MediaTek, said the platform brings the company's latest advances in flagship performance, on-device AI and power efficiency to the Galaxy Tab S12 Series, while Drew Blackard, SVP of Product Management at Samsung, said the partnership pairs flagship performance with on-device Galaxy AI. The Dimensity 9500-powered Galaxy Tab S12 series will be available at all major retailers starting Monday, October 7.

Samsung Electronics Co., Ltd. today unveiled the Galaxy Tab S12 series, comprising the Galaxy Tab S12 Ultra and Galaxy Tab S12+, at 5.1mm the company's thinnest Galaxy Tab and, Samsung says, its most powerful to date. The 3nm processor delivers up to 61% faster NPU processing, 19% faster CPU performance and 17% improved GPU power compared to Galaxy Tab S11 Ultra, with battery life of up to 23 hours on the Ultra and up to 21 hours on the Tab S12+, recharging from 0% to 100% in about 95 minutes with the 45W adapter. The Ultra carries a 14.6-inch display and the Tab S12+ a 12.6-inch display, both Dynamic AMOLED 2X panels reaching up to 1,600 nits of peak brightness, and both ship with an S Pen and run Android 17 with One UI 9. Adobe Photoshop comes pre-loaded on the series, a first for Android tablets, alongside Notability, and Samsung says the bundled app benefits are worth approximately $397.70 USD. The Galaxy Tab S12 series goes on sale starting Oct. 7 in select markets in Grey and Silver, with support for up to seven times of OS upgrades and security updates.

Dosilicon released an investor relations activity record announcement stating that the company's integrated storage-compute-connect chip R&D project has a planned cycle of about 42 months, and the project is currently progressing in an orderly manner. Through advanced packaging, the project integrates storage, connectivity, and computing units into a single chip to create a low-power embedded intelligent computing chip featuring local intelligent processing, data privacy and security, low latency, and low power consumption, with research and development undertaken by the company's own team. The announcement said that the chip needs to go through multiple stages including architecture design, tape-out, hardware and software debugging, and customer verification, so there is uncertainty in the R&D rollout. Please refer to the company's disclosed information for specific progress.

Qualcomm showcased phone-based AI agents at its Snapdragon Summit, unveiling the Snapdragon 8 Elite Extreme Gen 6 and Gen 6 chips that combine processing, graphics and sensing so agents can respond to context and act across apps. The wireless and edge computing chipmaker also reported more than 20 partners working with Adreno Neural Fusion, its AI rendering technology for mobile games. Shares rose about 0.3% to $188.04 at around 11:23 a.m. ET Tuesday, after closing at $187.48 on Monday. At that price the stock sits 7.34% above the chart's $175.18 GF Value estimate, meaning investors are already paying for some of the promise. The next proof point is widespread use in shipping phones, followed by evidence that it lifts chip revenue.

Apple is pitching its new desktops as a way for businesses to run AI without paying for every cloud request, a move that comes as the company's shares fell about 1.5% to $333.33 around 10:10 a.m. EST Tuesday. Apple says the M6 Mac mini delivers up to four times faster AI performance than its predecessor, and in a demonstration reported by Reuters, four connected Mac Studios ran a trillion-parameter model using one wall outlet. The company argues that owning the hardware could make sense for steady, private AI workloads, though it also leaves customers paying upfront and managing capacity themselves. At $333.33, the shares sit 16.03% above the chart's $287.27 GF Value estimate, meaning investors are already paying a premium. The bigger question for investors is whether that pitch can turn Macs into a meaningful business expense for AI teams.

SiMa.ai, a startup developing chips and software that let robots, drones, cameras, and other devices run AI directly on the device, has raised a $150 million Series C at a $1.45 billion valuation. The round was co-led by Fidelity Management & Research Company and Amplify, with participation from Alter Venture Partners, Dell Technologies Capital, and StepStone Group. Founded in 2018 by Krishna Rangasayee, previously COO of chipmaker Groq, the company provides energy-efficient chips that eliminate the need to send data back and forth to the cloud. SiMa.ai hopes its low-latency performance and more affordable chips, compared to Nvidia's GPUs, will help it capture the growing market for physical AI devices, including humanoid robots. The new round brings SiMa.ai's total capital raised to over $500 million, and the startup was previously valued at $960 million after raising an $85 million Series B in July 2025, according to PitchBook.

Spectrum has begun activating computing power across its network, placing distributed compute capacity within 10 milliseconds of 500 million connected devices in homes and businesses across the U.S. through more than 1,000 smaller owned and operated edge data centers already built into its infrastructure. The company said it is enabling new AI applications with Cast AI, HP and Hydra Host, marking a commercial adoption milestone for its Edge Compute Infrastructure platform powered by NVIDIA. Cast AI, HP and World Wide Technology will join Spectrum at SCTE TechExpo 26 to demonstrate low-latency AI workloads, including a robotics demonstration from WWT featuring a Unitree G1 humanoid robot in Spectrum's booth B335. Cast AI will highlight its OMNI technology integrating NVIDIA Blackwell Ultra-based accelerated computing platforms into hyperscale cloud environments, while HP will showcase HP Z Boost and its ZGX Nano and ZGX Fury AI Stations for enterprise edge AI development. Spectrum also has an agreement with Hydra Host allowing Hydra's AI Factory platform to use capacity owned, hosted and operated by Spectrum on its ECI platform.

Taiwan Semiconductor Manufacturing is expanding 2-nanometer production faster than previously expected as major customers increase their commitments to the new technology. The company could reach monthly 2nm output of about 120,000 wafers by the end of 2026, compared with earlier estimates of 90,000 to 100,000, according to Taiwan's EDN. Apple, Nvidia, Advanced Micro Devices, Qualcomm and MediaTek are reportedly lifting 2nm orders by 10% to 20%, giving TSMC additional visibility as customers prepare next-generation processors and AI accelerators. In the second quarter, 2nm represented 3% of TSMC's wafer revenue, while technologies at 7nm and below accounted for 77%, and high-performance computing made up 66% of revenue, up from 60% a year earlier. TSMC has projected 2026 2nm capacity growth as it ramps multiple facilities in Taiwan.

The global computer vision market is projected to grow from $17.75 billion in 2025 to $20.52 billion in 2026 and reach $37.1 billion by 2030, expanding at a compound annual growth rate of 15.9%, according to the Computer Vision Market Report 2026 added to ResearchAndMarkets.com's offering. The 2025-to-2026 increase represents a CAGR of 15.6%, with growth driven by declining computing costs, wider adoption of automation technologies, early machine vision systems and increased deployment of industrial robots. Continued investment in AI-based image recognition and real-time video analytics is expected to sustain growth through the end of the decade, supported by improvements in deep learning models, the development of autonomous vehicles, the expansion of smart factories and rising demand for real-time operational insights, with edge computer vision, 3D vision systems and vision-guided automation also shaping product development. Retail digitalization is a significant factor supporting demand, as retailers invest in cashierless checkout, artificial intelligence, big data analytics, cloud computing, Internet of Things infrastructure and mobility solutions, while the expansion of e-commerce reinforces demand for automated warehousing, visual product identification and streamlined fulfillment. In May 2025, Amniscient Inc. launched AmniSphere, a computer vision AI platform offering real-time operational monitoring, reported accuracy of 99.9% and a low-code interface, and in March 2025, SightLine Applications LLC acquired Athena AI for an undisclosed amount to combine AI-enabled vision, geospatial analysis and neural network-based decision-support capabilities with SightLine's edge video processing platform. Asia-Pacific was the largest region in the computer vision market in 2025, while North America is expected to record the fastest growth during the forecast period, and the report benchmarks companies including Microsoft Corporation, Sony Semiconductor Solutions Corporation, Intel Corporation, Texas Instruments Incorporated and MediaTek Inc.

China's Ministry of Industry and Information Technology has indicated it may allow companies including Alibaba Group Holding Limited and ByteDance Limited to purchase Nvidia's new RTX Pro 5500 chips, according to a report by The Information. The ministry reportedly asked domestic firms to report how many RTX Pro 5500 units they plan to buy and what they would be used for, signaling an intent to approve the transactions. The RTX Pro 5500 is a Blackwell-generation professional workstation GPU released this month, with 84 GB of GDDR7 memory and 21,760 CUDA cores, and a starting price reported above $6,000. Because the chip sits outside the data-center accelerator category that has been the primary target of successive rounds of U.S. export controls since October 2022, Beijing may have greater regulatory latitude to greenlight purchases without triggering fresh friction with Washington. The timing follows last week's Xi-Trump summit in Washington, at which chip export controls were conspicuously absent from both governments' official readouts; USTR Jamieson Greer told CNBC that export controls deemed national security issues were taken off the table during trade talks, and China's Ministry of Foreign Affairs did not mention export controls in its announcement of eight summit deliverables.

Apple introduced its first foldable smartphone, the iPhone Duo, at its first product event under new CEO John Ternus, with the device powered by the company's latest A20 Pro chip and a starting price of $1,999. The company also plans to bring Apple Pay to India initially through Axis Bank credit cards, following regulatory changes that allow biometric authentication for card payments, though the rollout will not cover India's UPI system, which processes around 84% of the country's digital-payment volume. Apple is reportedly exploring a return to the enterprise server market with an AI-focused server designed primarily for inference workloads, aimed at local inference needs of governments, businesses and AI developers. The redesigned Siri AI Beta will initially be unavailable in the European Union on iPadOS, iOS and watchOS, and regulatory constraints also limit access in China, while Apple faces intense scrutiny over App Store practices and digital markets legislation. Among institutions, hedge fund ownership slipped from 170 funds in Q1 2026 to 169 funds in the following quarter, short interest remains below 1%, and BlackRock is the largest stakeholder with 1.16 billion shares, or 7.97% ownership, followed by Vanguard Capital Management at 6.57% and State Street at 4.21%.

Advanced Micro Devices announced a collaboration with Advantech to accelerate Edge AI workloads using AMD embedded and GPU platforms, while MindWalk has moved OpenFold3 into production on AMD Instinct GPUs. The Advantech alliance focuses on scalable, partner-led Edge AI solutions, with Advantech's WEDA ecosystem putting AMD Ryzen AI Embedded processors into pre-validated edge platforms that can be rolled out across industrial fleets. MindWalk's OpenFold3 work on Instinct MI325X GPUs reaches into life sciences prediction workloads, supported by inference microservices for regulated, production-grade scientific use cases. The company said the announcements line up with its existing narrative of expanding AI infrastructure across accelerators, software and full systems, and that MindWalk's validation of AMD Inference Microservices and the Kubernetes native path into regulated production supports the view that ROCm and related tooling can lower switching costs for third party models. AMD investors are watching how many additional production case studies using AMD Instinct GPUs and inference microservices appear through 2027 across sectors like life sciences, security and edge industrial systems.

Apple has launched its first foldable iPhone Duo alongside the iPhone 18 Pro lineup, emphasizing enhanced on-device Siri AI and privacy as it enters the competitive foldable smartphone market. The move broadens Apple's premium hardware range and tests whether AI-focused differentiation can sustain its brand premium amid intensifying global competition. Among the latest announcements, a US$250,000,000 Siri and Apple Intelligence settlement is most relevant to the investment narrative, because it directly touches expectations around the AI features that underpin Apple's premium pricing; Apple denies wrongdoing, and the payout highlights how misaligned marketing and delivered AI capabilities can create both legal and reputational risks. Apple's narrative projects $616.5 billion revenue and $164.4 billion earnings by 2029, yielding a $325.66 fair value, a 5% downside to its current price, while the most pessimistic analysts assume revenue of about US$561.5 billion and earnings around US$147.4 billion by 2029. The key near-term risk remains margin pressure from higher component costs and legal or regulatory outcomes, though the new Siri-related class action settlement looks manageable rather than thesis changing.

Apple Inc. is undertaking a major overhaul of its Health app and Apple Watch strategy, the Wall Street Journal reported on September 9, 2026. The company will debut the Apple Watch Series 12 and Ultra 4 with an upgraded Health Sensing System alongside an Apple Intelligence-driven software redesign coming to iOS 27 that introduces a daily readiness score synthesizing activity, sleep, and vitals into a single 0-to-10 rating with plain-language guidance. The redesign adds personalized insights, long-term health analysis, at-home physical assessments, lab testing through Quest Diagnostics, and longevity features such as Health Age. Apple scaled back its planned AI health-coaching service, code-named Mulberry, after management decided the original concept did not compete strongly enough with products from Oura and Whoop, and Services chief Eddy Cue took over Apple's health organization after longtime executive Jeff Williams retired. Oura reported a 74% increase in revenue to $1.21 billion for the nine months through June and earned $60.8 million as it filed for a U.S. IPO, while Apple plans to launch the new Health experience first in U.S. English with additional languages coming later and some key features requiring newer hardware such as the Apple Watch Series 12 or Ultra 4.

Intel has expanded its Edge AI reach through a new collaboration with Advantech Co., Ltd. focused on Advantech's Edge AI ecosystem, deploying Intel hardware and software alongside Advantech platforms to support industrial and embedded AI workloads at the network edge. Intel also highlighted ongoing product development in brain-inspired computing, including neuromorphic style processors aimed at more energy-efficient AI inference. Intel is a large US semiconductor manufacturer with a US$644.2b market cap, supplying computing hardware and related services worldwide. The company's alignment with Advantech and its brain-inspired chip work back management's effort to refocus the portfolio on AI workloads, though organizational complexity and capacity constraints remain a flagged risk to execution. The clearest proof point from here is whether Intel starts disclosing concrete wins tied to WEDA and neuromorphic projects, such as named industrial deployments, container catalog usage, or dedicated brain inspired SKUs in upcoming earnings materials over the next few reporting cycles.

Apple is making the case for running demanding AI models on Macs rather than paying cloud providers per use, and its shares climbed to $337.48 around 11:11 a.m. ET Friday, up about 0.46% from Thursday's close. The hardware pitch rests on Apple's claim that its M5 Ultra Mac Studio supports up to 512 gigabytes of unified memory and 1.2 terabytes per second of memory bandwidth, while the M6 Mac mini starts at $899. In a demonstration reported by Reuters, four Mac Studios ran a trillion-parameter model using one wall outlet, an eye-catching display of capacity, though businesses still need to know what the machines cost to buy and run at scale. The chart adds a valuation check: Apple's $337.48 share price sits 17.62% above its $286.93 GF Value. Local AI could give customers a reason to buy more powerful Macs, especially for steady workloads where cloud usage bills pile up, but investors are already paying a premium and Apple now has to turn a convincing demonstration into repeatable sales.

Meta Platforms unveiled Meta VR Glasses, a roughly 100-gram virtual-reality device that will go on sale in spring 2027 for $1,299.99. The headset features a 5K Infinite Display with micro-OLED panels, Dolby Vision and Dolby Atmos, runs on Qualcomm's Snapdragon Reality Elite processor, and integrates Meta AI with eye, voice and hand-gesture controls; a two-part design houses compute, battery and storage in a separate tethered puck. Meta VR Glasses are the first IMAX Enhanced-certified VR device, offering 3D movies and more than 100 immersive live U.S. sports events annually, plus multi-screen workspaces and more than 75 games at launch. The launch comes as Meta's Reality Labs revenues rose 16% year over year to $431 million in the second quarter of 2026, growth that was partially offset by lower Quest headset sales, and as Apple expands Apple Vision Pro into accessibility applications such as eye-based control for power wheelchair drive systems and Alphabet pushes Android XR smart glasses built with Samsung, Gentle Monster and Warby Parker. Meta shares have rallied 17.8% year to date, trailing the broader Zacks Computer & Technology sector's 21.7% return, and carry a forward 12-month Price/Sales of 6.77X versus the Internet Software industry's 4.42X.

JPMorgan analyst Doug Anmuth sees Meta Platforms' eclectic new AI hardware lineup as part of a much larger AI platform that could address a market worth tens of trillions of dollars. Meta plans to offer more than 100 glasses across Ray-Ban, Oakley and Meta by year-end, targeting the more than 2 billion people who already wear glasses, and has shifted Reality Labs toward wearables designed around Muse and personal superintelligence. The lineup spans a hearing-enhancement feature in the U.S., the Muse Charm keychain device Zuckerberg called a "joyful little device" that is expected to ship in December, and Meta VR Glasses, described as a "private cinema" at roughly 100 grams and expected in spring 2027 for $1,299. Anmuth says Muse is already reaching millions of people and could eventually become a personal agent for billions, with JPMorgan expecting Meta to eventually monetize transactions between users, businesses and AI agents through fees or commissions. Direct Muse monetization remains limited today, and the bigger test for investors is whether Meta can turn the collection of products into a single AI ecosystem capable of expanding the company beyond advertising.

Qualcomm introduced its Snapdragon 8 Elite Gen 6 and Elite Extreme Gen 6 processors, bringing advanced on-device AI features to premium smartphones while extending the company's reach into automotive systems, edge devices, and data center workloads using agentic AI. Management linked the launch to Qualcomm's automotive digital chassis platform as it pushes AI processing closer to the device. The company, a US semiconductor group with a market value of about $210.7b, is targeting a US$40b non-handset QCT business, with non-handset AI compute expected to carry more of the load by 2029. Qualcomm said the clearest sign the shift is gaining traction will be how it breaks out QCT revenue from automotive, IoT, and data center in upcoming results and whether those segments move closer to longer-term targets flagged out to fiscal 2029.