Tag

Ai Hardware

All articles tagged with #ai hardware

technology2 days ago•6 min saved

AMD Strix Halo Brings 120B-Parameter AI to a $1,500 PC

AMD's Strix Halo Ryzen AI Max Plus enables on-device inference for models up to 120B parameters using up to 128 GB of unified memory in a $1,499 mini PC, offering strong cost efficiency and data control but with real-world memory bandwidth around 122 GB/s and a still-maturing ROCm/Windows ecosystem. It trails NVIDIA in prefill speed and lacks CUDA-level software maturity, though its cost per GB memory (~$25.77) is far cheaper than Apple M3 Ultra’s. A future Gorgon Halo with 192 GB and up to 300B parameters is planned for 2026, signaling AMD's ongoing AI hardware push. For workloads prioritizing local processing and agent workflows, Strix Halo is compelling; for CUDA-heavy or bandwidth-intensive tasks, NVIDIA or Apple may be preferable.

via Geeky Gadgets|

#ai-hardware #amd #cost-efficiency

business5 days ago•19 min saved

AI hardware momentum keeps chips in focus as rate bets wobble this week

Markets start a lighter week as investors eye PMIs and ISM services data after a June jobs report that added 57,000 payrolls and saw revisions lower, tempering bets on an imminent Fed rate hike while inflation remains a concern. Earnings highlights include PepsiCo and Delta Air Lines. In the AI space, memory and processor names (MU, INTC, AMD) led gains in the first half of 2026, helping the Philadelphia Semiconductor Index rally roughly 75% year-to-date, underscoring Bank of America’s view that AI’s growth now hinges on physical infrastructure—chips and power—as memory shortages persist.

via Yahoo Finance|

#ai-hardware #business #chips-memory

technology8 days ago•3 min saved

SpaceX denies rumors of an ultra-slim AI hardware prototype

The Wall Street Journal reports SpaceX is developing a slim AI hardware prototype thinner than an iPhone, but Elon Musk has denied the claim as utterly false; the piece notes SpaceX’s growing AI portfolio (SpaceXAI from xAI and Cursor) and a rumored device said to run a proprietary OS on a Qualcomm Snapdragon, though it may not reach market.

via 9to5Mac|

#ai-hardware #musk #spacex

business1 month ago•6 min saved

Nvidia Chief’s Praise Pushes Marvell Toward a trillion-dollar Dream

Nvidia CEO Jensen Huang hailed Marvell Technology as a potential “next trillion-dollar” company, sending MRVL higher after Nvidia disclosed a $2 billion stake. Marvell’s growth story hinges on its connectivity business and optical interconnects for AI clusters, with ~70% growth expected this year and revenue near $11.5 billion, but reaching a $1 trillion valuation would be difficult given margins and current multiples; the stock remains in a high-price landscape despite the endorsement.

via The Motley Fool|

#ai-hardware #business #marvell-technology

technology1 month ago•5 min saved

Microsoft’s Surface Laptop Ultra Debuts as RTX Spark-Powered Flagship

At Computex 2026, Microsoft unveiled the Surface Laptop Ultra, a flagship RTX Spark laptop powered by Nvidia’s RTX Spark SoC with a 20-core CPU and up to 128GB of unified memory, delivering AI-ready performance in a premium 15-inch form factor. It features a 15-inch mini-LED PixelSense Ultra display with up to 2000 nits, a premium chassis, robust cooling, and a full set of creator-friendly ports, plus a removable backplate for easier repairability. Several key questions remain about minimum RAM configurations, price (likely $2,500–$4,000+), battery life, and availability, with pre-orders expected later in the year.

via ZDNET|

#ai-hardware #computex-2026 #laptops

technology1 month ago•3 min saved

RTX Spark PCs Target Developers With Premium Pricing

Nvidia unveiled the RTX Spark processor at Computex 2026, powering new AI-focused PCs from major partners with a fall launch. Industry chatter points to prices around $2,000–$2,500 for the N1 and $2,500–$2,900 for the higher-end N1X, signaling a premium aimed at developers and early adopters rather than mainstream consumers as memory and storage costs loom.

via PCWorld|

#ai-hardware #computex-2026 #nvidia

technology1 month ago•15 min saved

Pinterest scales AI ambitions with a $4B AWS infrastructure deal

Pinterest announced a planned $4 billion, 2031-end infrastructure commitment with AWS—the largest in its history—to train and run AI models at scale for over 600 million monthly users, leveraging AWS Trainium and Graviton chips. The deal deepens a decade-long partnership, supports AI-driven visual search and discovery, and accompanies Pinterest’s infrastructure modernization to Kubernetes on EKS to boost developer velocity and efficiency.

via About Amazon|

#ai-hardware #artificial-intelligence #aws

technology1 month ago•7 min saved

Opal Bets Big on AI Hardware With OpenAI-Backed Audio Gadget

Opal, the webcam maker, is rebranding as Opal Electronics and shifting into AI-powered consumer devices after a $40 million Series B led by OpenAI (with Samsung, Peter Thiel, Seven Seven Six, and MKBHD among investors). The company is developing an AI-driven audio device to launch in 3–4 months, with testing by OpenAI, xAI, Thinking Machines, and Anthropic; it aims to become a design-first, Sony-like AI hardware brand and will manufacture in Taiwan. OpenAI is the largest shareholder but Opal retains its IP and will pursue partnerships with multiple AI labs for future models, with additional products expected within the next year.

via WIRED|

#ai-hardware #funding #opal-camera

technology1 month ago•4 min saved

Photons Could Speed AI Data Moves as Nvidia Bets Big

AI workloads are limited by how quickly data moves between chips and servers; photonics—using light rather than copper wires—could dramatically speed interconnects and reduce energy use, a technology Nvidia is betting on with billions in investments (Lumentum, Coherent, Marvell, Corning, Ayar Labs) to scale silicon photonics and GPU-to-GPU links, though mass deployment faces manufacturing ramp and redesign challenges to replace copper in existing product roadmaps.

via CNBC|

#ai-hardware #data-transfer #nvidia

technology1 month ago•11 min saved

Copilot+ AI Comes with 8GB RAM? Surface Laptop for Business 13" Priced at $1,299

Microsoft unveiled a Surface Laptop for Business 13-inch variant with 8GB RAM priced at $1,299, marketed as a Copilot+ PC, despite Microsoft's own 16GB minimum RAM guidance for Copilot+ and ongoing AI ambitions. The piece argues this pairing of premium design and low RAM is misaligned with enterprise needs, especially given Windows 11’s memory footprint and the 50 TOPS NPU, and contrasts it with Apple’s $599 MacBook Neo to suggest a flawed pricing/RAM strategy for Surface. It calls for a RAM-rich or more affordable entry-level option to uphold Surface’s reliability and enterprise value.

via Windows Latest|

#ai-hardware #copilot #ram

technology1 month ago•13 min saved

Memory Over Speed: The AI Inference Shift

Ben Thompson argues that the AI compute boom is moving from GPU-dominated training to memory-centric, agentic-inference architectures; Cerebras’ wafer-scale chips offer extraordinary on-chip memory and bandwidth for fast answer inference but face cost and scalability limits, while the long-term potential lies in memory hierarchies that support autonomous agentic work, potentially reducing Nvidia’s dominance and reconfiguring compute across training, inference, and even space data centers.

via Stratechery by Ben Thompson|

#agentic-inference #ai-hardware #cerebras

markets2 months ago•105 min saved

POET Shares Slump After Marvell Cancels Celestial AI Orders

POET Technologies’ stock fell after Marvell canceled all outstanding Celestial AI orders, with Marvell accusing POET of confidentiality breaches. POET says it’s pivoting to other priorities (including a $5 million order with another firm) and continues developing its POET Optical Interposer, but cautions there’s no guarantee it can reestablish the Marvell relationship. The company also faces upcoming May earnings and traders are watching key levels around support near $12 and resistance near $9 as the narrative unfolds.

via Benzinga|

#ai-hardware #celestial-ai #markets

markets2 months ago•16 min saved

Allbirds pivots to AI hardware, rebrands as NewBird AI amid sharp stock surge

Allbirds abandons its footwear roots to become AI hardware company NewBird AI, a move that sent its stock soaring about 660% in a day. The plan calls for leasing AI processing power, a pivot critics say could be a short-term FOMO play; the company has sold its shoe assets for $39M and aims to raise roughly $50M for initial chips. Leadership is still largely from the apparel side, and analysts have given BIRD a Hold, highlighting skepticism about building a true AI business without massive capital and expertise.

via TipRanks|

#ai-hardware #allbirds #amazon-aws

business2 months ago•4 min saved

Victory Giant Seeks HK IPO to Raise Up to $2.2 Billion Amid Iran-Driven Volatility

China's Victory Giant launched a Hong Kong share sale to raise up to HK$17.5 billion ($2.2 billion), aiming for the largest HK listing since September. Cornerstone investors including CPE Rosewood, Janchor Fund and Yunfeng Capital will buy about $997 million of shares; proceeds will largely fund expansion of production of printed circuit boards for AI hardware, with listing set for April 21. The deal comes as Middle East tensions add volatility to markets, but analysts see longer-term interest from AI investors.

via Reuters|

#ai-hardware #business #hong-kong

technology3 months ago•7 min saved

Arm Ventures Into Chipmaking With Its Own AGI CPU

Arm unveiled its first in-house semiconductors, introducing the Arm AGI CPU designed for agentic AI tasks in data-center servers. Built by TSMC on a 3nm process, the chip is pitched as highly energy-efficient to meet surging AI compute needs. Meta is among the first customers, with OpenAI, Cerebras, Cloudflare, SK Telecom, and Rebellions also lined up, and full production expected in the second half of 2026. The move marks a shift from Arm’s licensing model toward silicon ownership and could heighten competition with AMD and Intel in the AI-focused data-center market, while underscoring Arm’s emphasis on power efficiency.

via WIRED|

#agi-cpu #ai-hardware #arm