新旗艦款晶片B200
Nvidia CEO Jensen Huang took the stage at his company’s annual developer conference on Monday, March 18, in Silicon Valley to unveil new products designed to keep Nvidia at the forefront of the artificial intelligence industry.
3月18日星期一，輝達公司執行長黃仁勳在公司於矽谷舉行的年度開發者大會上發表將輝達置於人工智慧產業前沿的新產品。
take the stage 上台unveil(v.)(首次)展示;介紹;推出;將…公諸於眾=reveal=present=disclose=make publicat the forefront of sth.為某事的重要領導者artificial intelligence人工智慧=AI
Among these soon-to-be-released products touted by Huang were Nvidia’s new flagship chip — the B200 — as well as software for AI business solutions.
黃仁勳所宣傳即將推出的產品包括輝達的新旗艦款晶片——B200——以及提供AI商務解決方案的軟體。
release(v.)推出=roll outtout(v.)宣傳=promoteflagship(n.)旗艦(產品)flagship store 旗艦店
The B200, also called the “Blackwell” chip, is composed of “two squares of silicon the size of the company’s previous offering” bound “together into a single component,” according to a description in a Reuters’ report.
根據路透社報導中的描述，B200也被稱為「Blackwell」晶片，由「兩個與輝達先前提供的產品大小相當的方形矽晶片」組成，結合「在一起變成一個組件」。
be composed of…由～所組成=be made up ofoffering(n.)(某公司推出的)產品bind-bound-bound綁;結合component(n.)組成部分;成分;零件=component part
While neither the specific details of the B200’s performance nor its price were disclosed by Huang, Reuters noted that the Blackwell chip is purportedly “30 times speedier at some tasks than its predecessor.”
雖然黃仁勳並未透露B200的性能和價格的具體細節，但路透社指出，Blackwell晶片據稱「在某些任務上的處理速度比舊款快了30倍」。
neither…nor…既不…也不disclose(v.)披露、揭露=reveal=unveilpurportedly(adv.)據稱=reportedly=allegedlyspeedy(adj.)快速的predecessor(n.)前任;前輩、原有事物;前身The latest Ferrari is not only faster than its predecessors but also more comfortable.最新款的法拉利賽車不僅比舊款更快，而且更舒適。precede(v.)(指時間或空間上)處在…之前，先於in the preceding chapter在前一章節中=in the previous chaptersucceed(v.)隨後出現；接著發生in the succeeding chapters在接續的章節中=in the following chapterssuccessor(n.) 繼任者;接替者、繼承人;接替的事物This range of computers is very fast, but their successors will be even faster.這個系列的電腦速度很快，但它的換代產品速度會更快。
Despite the lack of technical specifications, CNBC has reported that these new chips are set to “ship later this year.”
儘管缺乏技術規格，CNBC報導這些新晶片預計「今年晚些時候出貨」。
specification(n.)規格=specbe set to VR預計=be scheduled to VR
At the conference, Huang also introduced his company’s new NIM — Nvidia Inference Microservice — software.
在大會上，黃仁勳還介紹了公司的新NIM（輝達推論微服務）軟體。
NIM=Nvidia Inference Microservice輝達推論微服務
This new product is designed to be licensed by companies that already own Nvidia graphics processing units as a means of optimizing Nvidia’s technologies for inference, which is “the process of running AI software.”
這款新產品旨在向已擁有輝達圖形處理器的公司授權，作為優化輝達舊推理技術的手段，推理即指「運行AI軟體的過程」。
license(v.)授權使用graphics processing units圖形處理器(GPU)as a means of~作為…的手段;方法optimize(v.)最佳化、優化
According to CNBC, NIM may even give businesses the ability to run their own AI models.
據CNBC報導，NIM甚至可能讓企業能夠運行自己的AI模型。
run(v.)運行、運作=operate
To further its reach in the AI industry, Nvidia has also announced the development of a complete server system which echoes Huang’s bigger-is-better GPU philosophy.
為了進一步擴大在AI產業中的影響力，輝達也宣布開發一套完整的伺服器系統，這呼應了黃仁勳「越大越好」的GPU理念。
further(v.)增進echo(v.)呼應
Called the GB200 NVLink 2, this system will combine “72 Blackwell GPUs and other Nvidia parts designed to train AI models.”
這個名為GB200 NVLink 2的系統將結合「72個Blackwell GPU和其他旨在訓練AI模型的輝達部件」。
parts=component parts零件、部件
Many industry insiders are already predicting huge possibilities for this new AI-centered server.
許多業內人士已經預測，這款以人工智慧為中心的新伺服器具有強大的可能性。
AI-centered=centered around or on AI人工智慧為中心的
Despite the announcement of these new products and Nvidia’s huge recent market gains, many analysts believe that Nvidia’s current 80% share of the AI chip market may actually slip “several percentage points in 2024” due to increased competition and current Nvidia customers making moves to develop their own chips.
儘管輝達發布了這些新產品以及近期在市場上取得巨大的收益，但許多分析師認為，由於競爭加劇以及輝達目前的客戶開始自行開發晶片，輝達目前在AI晶片市場上80%的份額實際上可能會「在2024年」下滑「幾個百分點」。
share(n.)份額slip(v.)下降;退步=decline=fall=droppercentage points百分點