Share
Linkedin
Twitter
Facebook
Whatsapp
Copy Link
onsemi launched its Embedded Power Platform that replaces traditional PCB substrates with silicon wafers, achieving 3-5x higher power density. The platform integrates silicon carbide and gallium nitride technologies for AI data centers and electric vehicles, with Subaru as an early partner evaluating the technology for future EVs.
A consortium of 10 banks led by Goldman Sachs and Barclays is providing $22 billion in debt financing to Crux AI, the cloud joint venture between Blackstone and Alphabet. The funds will purchase Google's Tensor Processing Units and support 500 megawatts of data center capacity launching in 2027.
AMD is developing its own neural rendering technology to rival Nvidia DLSS 5, targeting a launch with next-gen RDNA 5 GPUs in late 2027 or 2028. Leaker Kepler_L2 reveals AMD's approach aims to minimize AI-generated artifacts by giving AI direct access to 3D game data, but the delayed timeline may leave AMD trailing Nvidia's rapidly evolving DLSS stack.
Apple is developing an enterprise AI server powered by its future M8 Ultra chips, marking a potential return to the server market nearly two decades after discontinuing Xserve in 2011. The project could incorporate Nvidia NVLink Fusion technology and targets AI developers, businesses, and governments with a planned 2029 release.
AnalogAI has chosen memBrain SAGE neuromorphic hardware intellectual property from Microchip Technology's Silicon Storage Technology subsidiary for its edge AI processors. The hardware-aware solution enables simultaneous training and inference at or below one watt, targeting applications like environment-adapting humanoid robots, drones and autonomous vehicles.
Nvidia CEO Jensen Huang will join President Trump's state dinner for Chinese President Xi Jinping on September 24, alongside tech leaders including Sam Altman, Elon Musk, and Tim Cook. The high-profile event underscores AI's central role in US-China relations as both nations compete for dominance while industry insiders warn about existential risks from rapid AI advancement.
Microsoft hosts its first major Windows and Surface event in over two years on October 7 in San Francisco, focusing on local AI computing. CEO Satya Nadella and NVIDIA's Jensen Huang will unveil details about the Surface Laptop Ultra powered by RTX Spark, alongside broader AI-ready PC ecosystem innovations.
Netherlands-based startup Euclyd raised over $230 million in Series A funding led by Samsung Electronics, Somerset Capital Partners, and Scaleup Europe Fund. The company is developing craftwerk, an AI-specific chip designed to reduce energy consumption for AI inference workloads and tackle the memory wall challenge in the AI hardware ecosystem.
Bank of America strategists argue that tighter AI regulation will redirect spending from frontier AI models toward inference, monitoring, and control infrastructure. Nokia and STMicroelectronics are positioned as potential European winners, with Nokia's high-speed switches and optical transport solutions gaining traction as regulated AI agents require distributed infrastructure.
Broadcom CEO Hock Tan doubled down on the chipmaker's ambitious AI revenue targets after Anthropic CEO Dario Amodei called for moderating AI development pace. Despite a 4.8% stock drop and market concerns, Tan insists demand for AI compute infrastructure remains strong and durable, projecting $230 billion in AI semiconductor revenue by fiscal 2028.
MediaTek launched its Dimensity 9600 Pro, the first smartphone chip built using TSMC's cutting-edge 2nm process. The flagship processor delivers 51% improved AI performance with dedicated neural processing unit, 185fps gaming, and 17% faster single-core speeds. MediaTek aims to capture more premium smartphone market share against Qualcomm.
Pinterest announced an expanded partnership with Nvidia to supercharge its AI-driven discovery engine with 7.3 times faster search speeds. Built on Nvidia Blackwell GPUs and Dynamo, the new multimodal AI infrastructure processes over 80 billion monthly searches while cutting infrastructure costs and enabling Pinterest Assistant to handle 25 times more visual context per request.
Nvidia's RTX 5090 has vanished from major US retailers in under two weeks, with third-party sellers now charging $6,500 to $9,500—more than triple its $1,999 launch price. Photos show pallets of consumer RTX 5090 cards being deployed in AI server builds, highlighting how AI demand has transformed Nvidia's flagship GPU from a gaming card into an AI workhorse.
Nvidia unveiled the RTX PRO 5500 Blackwell workstation graphics card featuring 84GB of GDDR7 memory and 21760 CUDA cores—matching the GeForce RTX 5090's core count but delivering 2.6X more VRAM. Built on Blackwell architecture with the GB202 GPU, the card targets large-model inference, rendering, and AI workloads while filling the gap between RTX Pro 5000 and RTX Pro 6000.
Nvidia is reshaping AI infrastructure financing by partnering with BlackRock, Apollo, and KKR to raise over $500 billion in capital. The chip giant forecasts 70% revenue growth but faces a $100 billion supply-demand gap as it becomes the world's most active venture capital investor in large funding rounds.
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.