Laws that stand firm and independent are essential for artificial intelligence (AI) to thrive, as is the need for the data centres that power the technology to minimise their impact on society.
Amidst the volatile geopolitical climate, rules of law — including cyber laws — currently are being challenged.
These should stand firm and remain independent, above any political interests. “When we don’t have that, we don’t have any basis to even start [thinking] about how AI should be used responsibly,” said Kelly Forbes, president and executive director of AI Asia Pacific Institute (AIAPI).
“That is the larger challenge we’re seeing right now,” Forbes said in a video call with FutureCIO.
Typically, key global powers would navigate and work through issues as a market evolves. The rest of the world then would observe and monitor how things unfold, and work through how these would impact their respective region, she explained.
There currently are two major powers in AI — China and the US — but with the turmoil of the ongoing US-Iran war, amongst ongoing global tensions, the path remains unclear how AI should be adopted, safely.
This is anticipated to remain a challenge the world faces moving forward, Forbes said.
Founded in 2019, AIAPI is a non-profit organisation that aims to drive the responsible development and adoption of AI, including GenAI, across Asia-Pacific. Its focus includes advocating equal distribution and access to the technology.
Asked how Asia-Pacific fits in an AI landscape currently led primarily by a handful of US tech giants, Forbes noted that this was a particularly timely question, as the region traditionally is highly reliant on the West.
Asean has “really woken up” in the past year and focused on building its independence, because it realises that the current global powers are not fully reliable or predictable right now, she said.
Asean state members recognise the need to build regionally and foster more cooperation and stability within this region, so it can be stronger collectively, she added, adding that Asean is expected to become the fourth largest global economy by 2030.
Potential for Asean to rise with AI
Asean, or Association of Southeast Asian Nations, comprises 10 markets, including Singapore, Indonesia, Malaysia, and the Philippines. By 2030, the region’s digital economy is forecast to more than double to US$560 billion.
Coupled with AI, the region’s growth has become a significant opportunity, Forbes said.
However, this needs to be managed alongside growing geopolitical tensions and the lack of accessible infrastructures, such as data centres and reliable power grids, within some Asean markets, she said.
Compute power, amongst others, is needed to compete in this era of sovereign AI, she noted.
Having the necessary infrastructure for AI, as well as ensuring access and independence, becomes more important in discussions about how Asean should come together as a region, she added.
This is a growing concern since Asean governments do not want a “copy-and-paste” version of Silicon Valley. “They want something that reflects the region’s values,” Forbes said.
The region then have to think about how it should go about achieving this, she said, pointing to Singapore-led initiative SEA-LION as an example of such initiatives.
SEA-LION, or Southeast Asian Languages in One Network, aims to create large language models (LLMs) for populations that currently are under-represented in mainstream LLMs. It is trained on data in Southeast Asian languages, including Thai, Bahasa Indonesia, and Vietnamese, with the aim of ensuring the foundational model can better understand cultural context and nuances specific to the region.
There also has been growing need for data centres in the region to support increasing demand for AI.
Equinix is amongst market players stepping up to do so. It operates more than 280 data centres worldwide, including 64 across 18 Asia-Pacific markets.
In Singapore, which is its biggest Asia-Pacific market by revenue, it runs five data centres, with a sixth currently in construction, according to Tanya Ahuja, Equinix’s Asia-Pacific vice president of corporate development.
The tech vendor is seeing strong demand across the region, where it also has customers in Malaysia, Indonesia, and the Philippines. It is building its first data centre in Thailand.
Having spent the last several years on diversifying its footprint in Asia-Pacific beyond major hubs such as Singapore and Sydney, Equinix now is looking to scale and better serve local markets across the region, Ahuja told FutureCIO.
It is more challenging now, particularly due to access to power, she said, adding that the vendor’s priorities have shifted when it begins work on building new sites.
Previously, the main focus was on identifying and acquiring more land on which to build data centres, before it figured out how to pull in the power source.
Now, it searches first for sites that have power, then look to acquire the land needed, she said.
AI push for power highlights need to minimise impact
Mounting appetite for power as AI adoption climbs has put spotlight on the potential impact on economies and, in particular, the environment.
There are rising concerns that the influx of data centres will strain power grids and push up energy costs.
Some 75% of consumers are worried that AI data centres may lead to higher household power bills in their neighbourhood, according to a survey by SambaNova, which polled 2,525 respondents in the US and UK.
Another 75% know that AI data centres are associated with significant energy consumption, the survey found.
Some 71% believe AI data centres will strain their country’s power grid, while 83% think AI vendors should prioritise energy efficiency, even if it means slower rollout of new AI capabilities.
However, just 13% of organisations are monitoring their AI systems’ power consumption, SambaNova noted, citing its 2024 AI Leadership Survey. This is despite 49.8% of business leaders noting they are concerned about AI’s energy and efficiency challenges.
Ahuja acknowledged such concerns and underscored the importance of building facilities that are sustainable and energy efficient.
Some governments, too, already have laid out guidelines datacentre operators have to adhere to, so their sites are run efficiently, she said.
For example, Singapore in February announced plans to introduce legislation to ensure data centres and cloud platforms are managed sustainably.
The Asian nation currently hosts more than 1.4 gigawatts of datacentre capacity and expects demand to grow alongside the adoption of AI.
Its government wants data centres to tap green energy sources and take steps to improve their energy efficiency, including transitioning to more energy-efficient IT equipment.
Its new legislation will impose PUE (power usage effectiveness) mandates, amongst other requirements, on data centres.
Ahuja said Equinix already prioritises sustainability and looks at four key areas to minimise its data centres’ effects on the environment.
These encompass using renewable energy — it has 96% renewable energy coverage across its global portfolio — and prioritising sustainable elements in the design and construction of its data centres. These include tapping liquid cooling systems and exploring new green technology that can be incorporated into its systems, she said.
Equinix also assesses the viability of new sites based on their impact on the local community, the stress level on the local power grid, and whether renewable energy is available, she added.
Increasingly, companies have to think about these key areas to ensure their data centres have the least impact on the environment, she said.
“[Enterprise customers] now need to process more data in more places and transfer growing datasets quickly and securely, all while meeting compliance requirements and pursuing sustainability goals,” Shane Paladin, Equinix’s chief customer and revenue officer, wrote in a post.
And with AI still evolving, organisations want the flexibility to be able to scale in the longer term and support future requirements, such as inferencing on the edge, Ahuja said, when asked about challenges businesses faced today.
While it is still early days in terms of enabling inferencing at scale, businesses recognise that AI will continue to play an important role and they will have to move as the technology advances, she said.
When that happens, at some point, they will need more capacity as their AI deployments widen. Hence, they want the ability to do so when this becomes necessary, she added.
Governments also need to rise to occasion
Governments, too, should step up and establish a better understanding of AI, so they can better identify ways to govern the technology effectively, Forbes said.
There are worries that some tech vendors already have more power than governments, she said, noting that there are signs this already may be happening.
“If that’s the case, we’ll be living in a very different society looking into the future, because we’ve always relied on governments [to run nations and look after their population],” she said.
As it is, there are instance of companies lobbying how laws should be designed and of the impact they have on how such decisions are made, Forbes said.
Governments, though, are in a difficult position, where they have to regulate companies while ensuring laws do not curb the pace of AI innovation, she said.
They will need to find the right balance and to do so quickly, as the technology is advancing at an accelerated pace, she noted.
Both Ahuja and Forbes spoke with FutureCIO in the lead up to ATxE Summit 2026, in Singapore later this month, during which both female executives are scheduled as speakers.
