KUALA LUMPUR, March 28 (Bernama) -- Bursa Malaysia closed lower today on profit-taking after a strong three-day rally, and investor sentiment was further shaken in the late afternoon session following news of an earthquake in Myanmar with tremors felt in neighbouring Thailand, said Rakuten Trade Sdn Bhd equity research vice-president Thong Pak Leng. At 5 pm, the FTSE Bursa Malaysia KLCI (FBM KLCI) slid 1.44 per cent or 22.08 points to 1,513.65, its intraday low, from Thursday’s close of 1,535.73. The benchmark index opened 4.16 points lower at 1,531.57 and hit an intraday high of 1,533.52 during the midday session. On the broader market, decliners outpaced gainers 563 to 395, while 408 counters were unchanged, 1,106 untraded, and 133 suspended. Turnover slipped to 2.25 billion units valued at RM2.13 billion from 2.52 billion units worth RM2.41 billion on Thursday.
At NVIDIA's 2025 GTC conference, CEO Jensen Huang introduced the GB300 AI chip — a significant performance boost over its predecessor, the GB200. Expected to start shipping in the latter half of the year, this chip promises stronger performance for AI applications. But Huang didn’t stop there — he also unveiled Vera Rubin, the next-generation AI super chip, which could increase performance by 3.3x compared to the GB300.
Huang's excitement about AI's growth potential was evident, especially with the introduction of Rubin Ultra, which could deliver 14 times the performance of the GB300. He also previewed future NVIDIA chips like Feynman, signaling the company’s ambitions to dominate AI technology and high-performance computing.
Key Highlights from NVIDIA GTC 2025:
AI Chip Advancements:
GB300: 1.5x stronger than last year's GB200; launching later in 2025.
Vera Rubin (2026): 3.3x more powerful than GB300; enhanced memory and bandwidth.
Rubin Ultra: Expected to offer 14x the performance of GB300.
AI Supercomputers:
DGX Spark: A personal AI supercomputer for $3,000 with 10 quadrillion AI calculations per second.
DGX Station: Features the GB300 chip, designed for high-speed computing tasks.
New Software Tools:
NVIDIA Dynamo: An open-source project to speed up AI inference and reduce computational costs.
Llama Nemotron AI Model: Optimized for Agentic AI, outperforming older models in speed and accuracy.
AI in Robotics & Autonomous Driving:
Partnership with GM: Integrating NVIDIA AI into GM's autonomous driving tech.
Isaac GR00T N1: A humanoid robot for tasks like object handling.
Quantum Computing & the First "Quantum Day":
Quantum Day event in 2025, exploring the future of quantum computing with industry leaders.
Market Response:
Despite the impressive tech announcements, NVIDIA’s stock dropped by 3.4% after the event. Analysts, such as Blayne Curtis of Jefferies, expressed concerns over longer product release cycles and lack of new revenue sources. While NVIDIA's hardware-software strength is undeniable, the market remains cautious about its future growth potential.
Conclusion:
NVIDIA’s continued dominance in AI and high-performance computing is clear, but market reactions suggest that investors are eager for more immediate developments and evidence of broader growth opportunities. While the company’s AI advancements are exciting, the long timelines and market expectations may be holding back stock performance for now.
Comments
Post a Comment