The Zhitong Finance App learned that Nvidia (NVDA.US) CEO Huang Renxun recently revealed that he had asked the new cloud vendor Nscale headquartered in the UK to “work around the clock” to overcome the challenges encountered during the launch of some of its data centers.
According to reports, Nscale is an artificial intelligence (AI) cloud computing provider supported by Nvidia. Recently, some of its data center sites have been facing problems such as construction delays, equipment failure, and power supply bottlenecks. At the same time, its IPO valuation target of 30 billion US dollars has also been questioned by investors.
“Come on, guys, do a good job, work around the clock,” Wong In-hoon recalled his advice to the company at a press conference on Thursday. “That's all I said.”
Nscale did not immediately respond to a request for comment.
Now that the popularity of AI tracks is cooling down, Hwang In-hoon's statement is aimed at reviving market confidence. These new AI cloud vendors, funded by Nvidia, equipped with Nvidia chips, and specially built AI computing power data centers, have had major setbacks in recent weeks. Australian computing power company Firmus Grid, which also endorses Nvidia, has put on hold its planned $30 billion IPO to be listed in Australia; however, after South Korean manufacturer Elice landed in the capital market, its stock price performance was poor.
However, Hwang In-hoon believes that AI new cloud vendors will eventually overcome difficulties because demand for AI computing power is still strong.
“I'm very confident they'll find a solution, and their solution started with the fact that they have excellent customers and are in great demand; they just need working capital,” Wong In-hoon said at a press conference during the GTC conference in Berlin.
Hwang In-hoon said that new cloud vendors are building some of the world's largest and most complex computing power systems, while also dealing with the challenges of engineering construction, supply chain, and delivery. In the past, it could take an infrastructure company three years to build a data center. “But now, if you can't get it up and running within six months, everyone is going to attack.”
Over the past week, Hwang In-hoon has been touring Europe, signing contracts with local companies, and lobbying European business leaders to demonstrate the safety and sustainability of AI. He remains optimistic about the future of AI, but the current market sentiment has changed: as financing costs rise, the market is increasingly worried about whether the huge amount of money pouring into AI infrastructure will pay off.
In the past two weeks, individual stocks in the AI sector have been hit hard, and the much-publicized Philadelphia Semiconductor Index experienced the longest continuous decline since 2008.
Huang Renxun and his executive team held GTC conferences around the world. On the one hand, they introduced the latest AI developments to the outside world, and on the other hand, they worked hard to secure business for the company's expanding product line. In Berlin, the company unveiled new data center equipment, software, models, and announced partnerships with leading European companies.
Notable new devices include the DSX Pod, which is equivalent to a tiny data center and can be used by companies that want to get their hardware up and running faster. With a single container, customers get the server hardware, wiring, and plumbing they need — all installed and tested. Nvidia says it's just plugged in and connected to other similar units.