Intel is seeing demand globally for its newest artificial intelligence (AI) chip, Gaudi 3, and will sell them and bring them wherever they legally can, says Alexis Crowell, Intel’s Chief Technology Officer for Asia Pacific & Japan at Intel’s AI Summit in Singapore on May 9.
Crowell’s comment comes a day after the US revoked Intel and Qualcomm’s licenses to sell its chips to Chinese telecommunications giant Huawei, the latest development in the US-China semiconductor trade war.
Following that, Intel reported that it expects its second quarter revenue to fall below the midpoint of previous projections of US$12.5 billion ($16.94 billion) to US$13.5 billion, which helped send its shares down by 31% in April, their worst month in more than 20 years as reported by Bloomberg.
Intel’s Gaudi 3, which was launched in early April, boasts that it is over twice as power-efficient as and can run AI models one-and-a-half times faster than Nvidia’s H100 graphics processing unit (GPU) chip.
Intel has been reported to expand its strategic focus in Asia Pacific, Japan and India, like many of its peers in the semiconductor industry hoping to de-risk its supply chains.
At Intel’s AI Summit in Singapore, a report commissioned by Intel identified Singapore as a leader in AI maturity among eight economies in Asia Pacific. The report was conducted by global market intelligence firm IDC.
See also: South Korea eyes US$10 bil in support for chipmakers in 2025
IDC found that Singapore's strategic position as a regional hub, coupled with its burgeoning AI startup ecosystem, reputable academic institutions, and substantial government backing, helped propel the nation to the forefront of AI competitiveness and innovation in the region.
This puts it ahead of other countries in the Asia Pacific such as South Korea, Japan and Australia.
In Singapore, the banking, financial services and insurance sectors were identified as key sectors driving AI adoption and spending in the country, alongside the retail industry with the rise of omnichannel shopping.
See also: Nvidia forecast fails to meet loftiest estimates for AI star
Yet on a panel at Intel’s AI Summit, IDC’s VP in data analytics, AI, sustainability, and industry research Dr Chris Marshall, said that Singapore’s challenge in scaling AI is its small and ageing population.
Likewise, Singtel’s head of digital enterprise platforms of Digital InfraCo, Manoj Prasanna Kumar, says that the lack of physical space to house data centres will be a constraint for the city-state.
At present, Singapore is heavy on inferencing, in which banks and financial institutions use existing large language models (LLMs) like Llama and fine tune those data sets, says Kumar. But not a lot of foundational models, which takes a few hundred to a thousand GPUs to train, are being built.
Nevertheless, IDC forecasted that Singapore's AI expenditure will grow at a compound annual growth rate (CAGR) of 25.3% from US$1.7 billion in 2023 to hit the US$4.2 billion mark by 2027.
On a regional level, IDC forecasts that AI spending in Asia Pacific (excluding China) will grow at a CAGR of 28.9% from 2023 to reach US$90.7 billion by 2027.
