KUALA LUMPUR, March 5, 2026, 00:15 (MYT)
- VCI Global’s V Gallant unit launched what it’s billing as Malaysia’s first AI GPU Computing Center using NVIDIA tech, opening its doors March 3.
- The company introduced “Intelli‑X,” an enterprise large language model platform. It intends to offer the product through subscriptions and dedicated clusters.
- VCIG shot higher early in U.S. premarket, but those gains faded as the session went on.
VCI Global Limited fired up what it calls Malaysia’s first NVIDIA-powered AI GPU Computing Center and rolled out its Intelli‑X large language model platform, aiming to ramp up its AI infrastructure offerings. Shares of the Nasdaq-listed firm swung sharply in U.S. trading.
Timing is a key factor for Malaysia, which is now in line to pull in over $40 billion in AI and data centre investment by 2030, research cited by InvestKL shows, as hyperscalers ramp up in Southeast Asia. “High-value tech hiring… and the projected US$40 billion investment in AI data centres” could be a magnet for more regional operations, Phill Brown of Robert Walters said in a report connected to InvestKL. investkl.gov.my
Policy makers want broader adoption, but practical hurdles remain. According to the Ministry of Digital’s Economic Outlook 2026 report, cited by Bernama, AI integration could boost Malaysia’s economy by as much as $115 billion by 2030. The report also points to infrastructure readiness as a sticking point, particularly for small and medium enterprises.
VCI Global’s GPU facility began operations March 3, targeting government, enterprise, and SME clients seeking local compute options and stricter data oversight. GPUs—graphics processing units—are the chips of choice for AI model training and inference, thanks to their parallel processing capabilities.
The company said the centre’s going to handle workloads such as generative AI “inference”—that is, running an AI model after training—plus predictive analytics and workflow automation. Customers, according to the company, will have the option to train and fine-tune models within a “firewall-contained” environment. That setup aims to reduce dependence on offshore cloud providers and help customers manage cross-border data rules.
VCI Global intends to make money off the build by leasing GPU compute, offering dedicated private clusters, and selling Intelli‑X subscriptions. The setup centers on an “enterprise LLM”—a large language model, the kind of software powering chatbots that generate text and reply to user questions.
VCI Global rolled out the launch and announced V Gallant had inked memoranda of understanding with Khalifa Intelligence, UCSI College, and Favoriot. These MoUs aren’t binding—terms can shift before anything turns into a formal contract. “We’re working to create a scalable AI commercialization engine designed to generate recurring revenue,” said Chief AI & Data Officer Chan Wai Mun.
VCIG shot up to $21.80 in early premarket moves at 9:30 a.m. EST, surging 171.48% from its previous close, Public.com data showed. Shares then shifted lower, changing hands at $9.83, still up 22.42% for the day. Investing.com pegged the day’s trading range between $8.20 and $23.40.
VCI Global, in another filing, announced that executive directors Victor Lee and Alex Chua stepped down as of Feb. 28. The company said their exits did not stem from any disagreement. On March 1, the board named Michael Puah as an independent director, confirming he qualifies under Nasdaq’s independence rules.
VCI Global is stepping into a busy field for AI compute, up against cloud giants that already offer GPU power by the truckload. In Malaysia, overseas tech heavyweights—including ByteDance, Google, and Microsoft—have unveiled major commitments to local cloud and AI infrastructure, according to Reuters.
Execution looks tricky here. VCI Global hasn’t said what the centre will cost to build, nor has it shared any numbers on customer commitments or revenue goals. Its announced partnerships? Just MoUs, not actual long-term deals. The company is also up against much bigger cloud players and regional data centre rivals, both of which have the muscle to cut prices and tie up multi-year supply.
