Intel's SambaNova acquisition was seen as a way to spearhead the company's AI strategy in inference, but it appears Team Blue has settled for much less. Intel's CEO Now Intends to Invest Directly in SambaNova's Funding Round After Previously Backing It Through Walden Capital When we talk about Intel, the company is one of the only compute providers that hasn't managed to capitalize on the AI frenzy at all, and this has been a persistent problem for several quarters. Intel's former CEO, Pat Gelsinger, has acknowledged the lack of intent in AI, and it appears that missing out on the […]

Read full article at https://wccftech.com/intel-misses-another-ai-opportunity-with-sambanova/

Fortnite gamers have acquired computing power that rivals that of some of the world's largest hyperscalers, and that is just from gaming PCs and consoles. Fortnite Gamers Have Contributed Significantly Towards Modern-Day Hardware Advancements In today's world, computing power is a major bottleneck, which is why hyperscalers are investing "hundreds of billions" in the AI frenzy to build their capabilities. One recent deal we discussed was the AMD-Meta agreement, which involved compute commitments worth up to 6 gigawatts. Interestingly, when talking about the deal, Epic Games' CEO Tim Sweeney revealed that Fortnite gamers alone consume up to 30 gigawatts of […]

Read full article at https://wccftech.com/fortnite-players-alone-generate-30gw-of-compute-power/

AMD has signed a major agreement with Meta, which involves the deployment of customized infrastructure that could be worth 'hundreds of billions'. AMD To Provide Next-Gen Venice & Verano CPUs to Meta, Alongside MI450 GPUs In a 6GW Commitment The race for next-gen AI infrastructure is more competitive than ever, with manufacturers like AMD/NVIDIA rushing to enter exclusive partnerships to secure customer capacity for their upcoming product lineups. A few days ago, we reported on NVIDIA's agreement with Meta for Vera Rubin AI racks, and the headline then was that Team Green snagged one of AMD's biggest customers. However, AMD […]

Read full article at https://wccftech.com/amd-puts-up-a-hefty-20-of-its-company-just-for-ai-deals/

The Japanese government is eager to expand onshore production, and one way to attract investment is to offer incentives to giants like SK hynix and Samsung. Japanese Government's Support & Incentives Reportedly Reduce TCO of Memory Fabs by More than 50% The global semiconductor dynamics have evolved rapidly since nations like Japan and the US have made chip production a matter of national security. Incentives like the CHIPS Act have contributed significantly to the growth of American manufacturing, and the current Japanese administration is now looking to attract private investment as well. A report by Chosun Biz reveals that memory […]

Read full article at https://wccftech.com/japan-is-actually-offering-incentives-to-samsung-and-sk-hynix-to-build-memory-fabs/

ASML plans to enable fab manufacturers to aggressively increase production through its latest breakthrough in EUV, as the Dutch chipmaker ramps up its light-source capabilities. ASML's Enhanced EUV Light Source Projected to Bring in a Huge Increase in Fab Output, Only With Equipment Upgrades The semiconductor industry is currently in a supercycle driven by demand from fabless manufacturers, not just for consumer products but also for enterprise/AI. We have extensively reported on how chip giants like TSMC are under significant supply constraints, and one way the supply chain is addressing this is by building an extensive fab network. However, ASML […]

Read full article at https://wccftech.com/asml-set-to-boost-chip-output-by-ramping-euv-power-to-a-kilowatt/

Intel's unified core strategy is something the world anticipates as the next pivot towards the P/E-core duopoly, and recent job postings indicate that the shift is being developed internally. Intel Is Looking Towards Pivoting Away From Current P/E-Core Designs, But the Idea is In Early Stages The idea of having one 'big core' has been discussed several times in the past, but it remains unclear how major CPU manufacturers will implement this approach in their next-gen lineups. However, Intel's recent job posting reveals that the company has a dedicated "Unified Core" team working on the venture, further bolstering the credibility […]

Read full article at https://wccftech.com/intels-unified-core-ambitions-with-next-gen-cpus-remain-intact/

Samsung has reclaimed the top spot in the DRAM industry, as the Korean giant's efforts to restructure its HBM business have proven fruitful. Samsung's DRAM Market Share Now Surpasses SK hynix, Credited to Recent HBM & Memory Breakthroughs For those unaware, Samsung has been leading the memory market for several years now, and the Korean giant was known to have one of the largest DRAM production capacities; however, last year, the company failed to achieve key objectives. This allowed competitors to gain market share from Samsung, which is why the firm lost its lead to SK hynix. However, according to […]

Read full article at https://wccftech.com/samsung-memory-comeback-is-one-to-watch/

Well, the idea that Chinese memory modules are cheaper has been outdated, as a recent DDR5 listing by a domestic vendor shows that RAM prices have almost leveled out worldwide. KingBank's DDR5 Modules Are Now Retailing For Prices Similar to Western Alternatives, As Shortages Leave No Other Way We did see the narrative of Chinese memory emerging as a 'saviour' for gamers pop up a few weeks ago, and it was mainly driven by the fact that manufacturers like CXMT weren't as mainstream compared to the 'Big 3' memory suppliers. However, it didn't take much time for the Chinese RAM […]

Read full article at https://wccftech.com/chinese-memory-prices-have-started-to-catch-up-with-global-markets/

NVIDIA's GB300 NVL72 AI racks have been tested across DeepSeek's latest open source models, and through fine-tuning and optimized inference, the results are indeed promising. NVIDIA's Blackwell Ultra Scores Up to a 1.5x Lead Over GB200 NVL72 In Latency-Sensitive Workloads With GB300, NVIDIA's primary focus has been on delivering optimal long-context performance in order to capitalize on the agentic AI wave. In a recent post, we discussed how Blackwell Ultra delivers a 50x increase in throughput per megawatt compared to Hopper GPUs through its extreme co-design approach. Now, the Large Model Systems Organization (LMSYS) has tested GB300 NVL72 for long-context […]

Read full article at https://wccftech.com/heres-how-nvidia-blackwell-ultra-is-dominating-long-context-deepseek-workloads/