AI Accelerators

AI Accelerators

Top News

Samsung HBM4 sales cross $1 billion within four months as

Samsung HBM4 sales cross $1 billion within four months as AI chip demand accelerates

Samsung Electronics achieved over $1 billion in HBM4 sales just four months after launching mass production in February, becoming the first company globally to ship sixth-generation high bandwidth memory chips. The rapid revenue growth signals accelerating demand for next-generation hardware tailored for AI applications, particularly for platforms like Nvidia's Vera Rubin that power generative AI workloads.

Technology2 sources
SanDisk proposes stacking HBF NAND beneath GPUs to tackle

SanDisk proposes stacking HBF NAND beneath GPUs to tackle memory bottlenecks in AI systems

SanDisk has unveiled High-Bandwidth Flash (HBF), a memory architecture that stacks NAND flash directly beneath AI processors using advanced packaging. The technology could deliver up to 4 TB of storage capacity per stack, addressing HBM shortages and capacity constraints. While still in the patent stage, HBF represents a shift toward tiered memory solutions for AI workloads.

Technology2 sources
Intel hires former SK hynix chief Seok-Hee Lee to lead

Intel hires former SK hynix chief Seok-Hee Lee to lead advanced packaging push for AI chips

Intel has appointed Seok-Hee Lee, former CEO of SK hynix and SK On, as executive vice president of Intel Foundry to oversee advanced packaging, system integration, and back-end technology development. The hire comes as Intel splits advanced packaging into a dedicated business unit, aiming to capture billions in revenue from hyperscaler customers building next-generation AI systems.

Business3 sources
Architect Labs raises $24M to challenge Broadcom and

Architect Labs raises $24M to challenge Broadcom and Marvell with AI-powered chip design

Palo Alto startup Architect Labs emerged from stealth with $24 million in seed funding to transform custom chip design using AI. The company aims to reduce the two-year timeline and hundreds of millions in costs currently required to design custom chips, directly challenging industry giants Broadcom and Marvell who generate tens of billions annually from this business.

Startups3 sources
AMD FSR 4.1 support arrives early for RDNA 3 GPUs, bringing

AMD FSR 4.1 support arrives early for RDNA 3 GPUs, bringing AI upscaling to RX 7000 series

AMD has released FSR 4.1 support for RDNA 3 desktop GPUs ahead of schedule, bringing AI-powered upscaling to the RX 7000 series. The INT8 model is now available in over 300 games through driver updates, delivering up to 50% performance improvements with no quality loss. RDNA 2 support faces delays until early 2027 due to optimization challenges.

Technology4 sources
Qualcomm eyes $8-10 billion Tenstorrent acquisition in

Qualcomm eyes $8-10 billion Tenstorrent acquisition in major AI chip and RISC-V power play

Qualcomm is negotiating to acquire AI chip startup Tenstorrent for between $8 billion and $10 billion, according to recent reports. Led by renowned chip architect Jim Keller, Tenstorrent develops RISC-V-based AI accelerators and data center processors. The deal would represent one of Qualcomm's largest acquisitions and signal a significant commitment to RISC-V architecture.

Business4 sources
Dell and AMD expand on-premises AI platform with

Dell and AMD expand on-premises AI platform with next-generation MI350P GPU support

Dell and AMD are expanding their collaboration to deliver enterprise-grade on-premises AI infrastructure. Starting July 2026, Dell PowerEdge XE7745 and R7725 servers will support AMD Instinct MI350P PCIe GPUs, enabling organizations to scale generative and agentic AI workloads within existing data center infrastructure without costly redesigns.

Technology2 sources
HP and Ferrari unveil $5,599 limited edition AI PC with

HP and Ferrari unveil $5,599 limited edition AI PC with transparent engine bay and racing DNA

HP and Ferrari have partnered to create the HP Limited Edition Scuderia Ferrari AI PC, a striking red laptop limited to just 4,999 units worldwide. Priced at $5,599, this collector's item features a transparent Gorilla Glass panel revealing its cooling system, carbon fiber construction, and Ferrari-inspired design elements. The laptop launches June 12 in select markets.

Technology3 sources
3DMark Benchmark Previews Native 4K Path Tracing With AI

3DMark Benchmark Previews Native 4K Path Tracing With AI Rendering at Computex 2026

UL Solutions showcased its most demanding 3DMark benchmark yet at Computex 2026, featuring native 4K path tracing with AI upscaling and frame generation. The preview at Thermal Grizzly's booth reveals a shift in GPU testing toward AI-assisted rendering technologies alongside traditional graphics performance metrics, signaling how next-generation GPU architectures will be evaluated.

Technology2 sources
Microchip Technology launches XpressConnect Retimers to

Microchip Technology launches XpressConnect Retimers to tackle latency in AI Data Centers

Microchip Technology has unveiled XpressConnect PCIe 6.0 and CXL 3.1 retimers designed to solve critical latency and signal integrity challenges in AI data centers. The retimers achieve pin-to-pin latency of less than 12 ns—approximately 80% lower than PCIe 6.0 specifications—helping improve utilization of AI accelerators and GPUs while extending signal reach beyond conventional limits.

Technology2 sources
AMD Radeon RX 9070 GRE launches globally at $549, targeting

AMD Radeon RX 9070 GRE launches globally at $549, targeting 1440p gaming enthusiasts

AMD brings its China-exclusive Radeon RX 9070 GRE to global markets at $549, positioning it as a mid-range graphics card for 1440p gaming. Built on RDNA 4 architecture with 12GB GDDR6 memory, the card delivers roughly 19% better performance than NVIDIA's GeForce RTX 5060 Ti 16GB. The launch comes amid supply constraints and rising GPU prices driven by the AI gold rush.

Technology3 sources
Intel Crescent Island targets AI inference with cheaper

Intel Crescent Island targets AI inference with cheaper memory, air-cooling to challenge Nvidia

Intel unveils Crescent Island, an AI chip designed for inference workloads that uses cost-effective LPDDR5X memory instead of expensive HBM. The air-cooled GPU supports up to 480 GB of memory and ships later this year as Intel pivots from its failed Gaudi training chips to focus on the growing inference market.

Technology4 sources
TSMC targets 30% power savings as AI energy crisis forces

TSMC targets 30% power savings as AI energy crisis forces chip design rethink

TSMC says surging electricity demands from AI are making energy efficiency the main constraint shaping future chip development. The world's largest contract chipmaker expects to cut power consumption by up to 30% by 2028 while delivering over 20% higher computing performance, as customers from smartphones to AI data centers prioritize performance with lower power usage.

Technology2 sources
Broadcom partners with FuriosaAI on 2nm AI accelerator chip

Broadcom partners with FuriosaAI on 2nm AI accelerator chip with HBM4e memory for inference

Broadcom has partnered with South Korean startup FuriosaAI to develop a third-generation AI accelerator chip built on a 2nm process with HBM4/HBM4e memory. The custom AI silicon targets high-volume AI inference workloads and leverages Broadcom's advanced packaging technology to integrate multiple chiplets into a single system-on-package, with sampling expected in early 2028.

Technology2 sources
Rambus pushes DDR5 memory to 9600 MT/s, targeting agentic

Rambus pushes DDR5 memory to 9600 MT/s, targeting agentic AI workloads on next-gen PCs

Rambus announced a complete DDR5 9600 client memory module chipset designed to push PC memory speeds to 9,600 megatransfers per second. The chipset targets bandwidth and capacity demands of agentic AI workloads on desktops and laptops, setting the stage for Intel Nova Lake and AMD Zen 6 platforms launching later this year.

Technology2 sources
Prev1234…7Next
Organisationarrow_forward
OpenAIGoogleAnthropicMicrosoftMetaNvidiaApple
Peoplearrow_forward
Sam AltmanJensen HuangElon MuskSundar PichaiSatya NadellaMark zuckerbergDario Amodei
Productarrow_forward
ChatGPTGeminiClaudeCopilotGrokDeepSeekAndroid
LLM Modelarrow_forward
GPTClaudeGeminiLlamaGrokDeepSeekNanoBanana
Eventarrow_forward
Google I/OWWDCMicrosoft BuildNVIDIA GTCMWCNeurIPSCES
Conceptarrow_forward
Agentic AILLMsAGIAI safetyRAGVibe CodingAI regulation
Industryarrow_forward
FinTechRoboticsCybersecurityGamingLegalMediaAI learning
Geoarrow_forward
United StatesChinaEuropeIndiaUnited KingdomJapanMiddle East
TheOutpost.ai

Don’t drown in AI news. We cut through the noise - filtering, ranking and summarizing the most important AI news, breakthroughs and research daily. Follow topics that matter to you and stay ahead.

Instagram logo
LinkedIn logo
Youtube logo
Explore
Top StoriesPodcastsNewsletterBlogArchivesAbout us
News Categories
TechnologyBusinessPolicyStartupsHealthScienceEntertainment
© 2026 TheOutpost.AI All rights reserved
TermsPrivacyContentContact Us