Semiconductors & Infrastructure

SKハイニックスのSolidigmが米国の承認を得て中国に半導体製造装置を送ることを検討し、その見返りとして米国でNANDチップ工場の建設を計画している

• •SKハイニックス傘下のSolidigmは、中国への半導体製造装置の追加輸出について米国の承認を求めている
• •この交渉は、米国国内で新たなNANDチップ工場を建設する検討と関連づけて行われている
• •企業は中国への設備輸出と米国での生産基盤構築のバランスを取ろうとしている

The article examines how market selloffs and rising interest rates are testing the boom in AI data center deals.

The article explores Nvidia's RTX Spark and its potential to run 'everything' on Arm, contrasting it with the historical failure of the original Microsoft Surface PC.

• •Fourteen years ago, Microsoft's original Surface PC, powered by an Nvidia Tegra chip, was a major failure that resulted in a $900 million inventory write-off.
• •The primary reason to care about Nvidia's RTX Spark is that 'everything' is expected to run on the Arm architecture.
• •The article draws a parallel between the past failure of the Surface and the current trajectory of Arm-based computing.

The article reports on a launch event in San Francisco for Nvidia's RTX Spark laptops, focusing on how Windows is being optimized for these new AI PCs.

• •Microsoft and Nvidia held a launch party in San Francisco for the first wave of RTX Spark laptops.
• •The event featured a new wave of AI PCs with a lively hands-on area for attendees.
• •The author attended a quieter Nvidia event the day after the media launch to hear more details.

NVIDIA's RTX Spark anticipates the PC form factor required for the agent AI era, but its novel value proposition makes it difficult to convey from a journalistic perspective.

• •RTX Spark is described as anticipating the PC form factor needed for the agent AI era.
• •The product proposes entirely new values and usage methods, making it hard to explain from a journalist's viewpoint.
• •The title implies high demand, with units selling out immediately despite a price of 1 million yen.

NVIDIA revealed details of the RTX Spark N1X, featuring a TSMC 3nm chiplet design with a 20-core CPU and a 48 SM GPU connected via NVLink C2C.

• •The CPU die, custom-designed by MediaTek, features a 20-core configuration and achieves 300GB/s bandwidth with LPDDR5X-9400.
• •The GPU die contains 48 SMs and 6,144 CUDA cores, utilizing NVLink C2C for connection to the CPU.
• •The device uses a chiplet structure composed of a CPU die and a GPU die manufactured with TSMC's 3nm process.

スタートアップのRicursive Intelligenceは、チップ設計を自動化し加速させることで計算資源の危機を解決しようとするAIを開発している

• •同社は、チップを設計し、各イテレーションから学習して後続の設計を改善するAIを開発している
• •この技術の目標は、チップ設計プロセスの自動化と加速である
• •創業者らは、計算資源の危機を解決するためにこのアプローチを提案している

The article provides a consolidated overview of AWS updates and service changes announced during the week of October 5.

Hitachi, OEDG, and three other companies are conducting a demonstration with the Ministry of Internal Affairs and Communications to shift AI workloads based on renewable energy availability.

• •The 'workload shift' strategy moves AI computation to times and locations where renewable energy power is underutilized.
• •Five companies, including Hitachi and OEDG, are participating in a demonstration project with the Ministry of Internal Affairs and Communications.
• •The project verifies moving data center workloads based on power generation status in Yokohama, Kitakyushu, and along the Joetsu Shinkansen line.

AIエージェントの台頭を背景に、トランザクションデータと分析データの両方を単一のデータベースで処理するCedarDBのエンジンが注目されている

• •従来のデータベースは分析データ用または運用データ用のどちらかに特化して設計されていた
• •データベースエンジンCedarDBは、この両方のデータを同時にサポートするアプローチを採用している
• •エージェントAIの普及に伴い、この統合的なアプローチの重要性が高まっている