Aussie equities: How Xero and NEXTDC are leveraging AI technology

While the US-based ‘Magnificent Seven’ takes the spotlight, several ASX-listed companies are leveraging of AI technology, reports Pendal analyst ELISE McKAY

• Xero is close to testing an AI-based business companion
• NEXTDC is building local AI infrastructure
• Find out about Pendal Focus Australian Share fund

ON A TOUR of the US late last year, Pendal Aussie equities analyst Elise McKay met with dozens of companies – and found generative AI was a topic in almost every meeting.

“There’s strong evidence that over the longer term generative AI will have a big impact across the business landscape,” McKay said at the time.

Aussie equities investors should keep an eye on how companies are investing in AI, as well as which infrastructure suppliers are best placed to benefit in a fast-evolving market.

Below is an update from McKay, outlining how her expectations are playing out after the recent ASX results season. 

Here she focuses on ASX-listed data centres player NEXTDC and software maker Xero – both held by Pendal.

XERO (ASX: XRO)

Generative AI – a machine-learning model that is trained to create new data, content and insights – is being embraced globally as corporates leverage “large language models” and infrastructure available off the shelf. 

Xero – a cloud-based finance and accounting software maker – is a good example.

At its inaugural investor day, Xero announced a generative AI tool called JAX (short for Just Ask Xero).

The tool is designed to help customers with personalised insights and tasks such as raising and sending invoices to customers. 

Xero plans to test the tool with customers later this year. 

Xero first started working on its generative AI strategy about eight months ago and made a key hire in Eitan Sharon as its data and science lead in November.

Within months of Eitan joining, Xero had a functional tool in place, demonstrating the speed with which these products can be built and deployed.   

NEXTDC (ASX: NXT)

With rapid growth in demand for generative AI tools comes demand for infrastructure to run and support training and inference models.

Training refers to the process of building AI models, while inference refers to how those models are deployed in an application or business setting.

Focus is expected to shift from AI training to inference over time – which may advantage some players and disadvantage others.

Australian data centre operator NEXTDC (ASX: NXT) hopes to play a significant role in both markets.

NXT is expected to launch its first Nvidia-based “AI factory” in coming months. 

It’s our expecation the AI factory will use graphics processing units made by Nvidia – the $US 2 trillion-dollar NASDAQ-listed chipmaker that has become the face of the artificial intelligence hardware manufacturing race.

This would be complementary to existing NXT data centre infrastructure which is ideally positioned to hosts inference model technology – which requires proximity to clients to limit latency. 

NXT has recently signed contracts with hyper-scale players in Sydney and Melbourne to support their cloud deployments.

AI deployments would be over and above these contracts, and — if following the US trend — at a size significantly larger than what has recently been signed.

There are now also contracts being tendered in market for sizeable AI deployments for both training and inference needs. These are reportedly significantly larger in size again than recent contracts. 

Demand for onshore AI infrastructure

NXT’s customers are expected to differentiate their needs depending on whether they are contracting for training or inference work. 

But in both cases, Australian clients prefer onshore AI infrastructure technology due to data sovereignty issues and the need for low latency. 

Training models can be located close to wherever access to power is available and priced cheaply, while inference models need to be in metro locations. 

Find out about

Pendal Horizon Sustainable Australian Share Fund

Read more

This is well-suited to NXT’s footprint which is spread geographically across six states and territories of Australia and closely located to CBDs. 

NXT is well positioned to win its fair share with available inventory in all markets and great relationships across the customer base.    

Access to power

Access to power is one of the big issues that continues to face the AI industry

Data centres, crypto currencies and AI consumed 2% of global energy demand in 2022, the International Energy Agency estimates.

The IEA forecasts global demand for energy from these sources to grow at 8% to 23% from 2022 to 2026. 

In the US, data centres accounted for about 4% of energy demand in 2022. They are expected to account for half of US electricity demand growth over the next three years. 

Clearly the AI industry needs resolve this issue to meet demand.

The issue will be attacked from multiple angles including increased compute efficiency (eg Huang’s law), new cooling technologies (eg direct-to-chip cooling, liquid immersion) and innovative power solutions (eg small modular reactors). 

About Elise McKay and Pendal Australian share funds

Elise is an investment analyst and portfolio manager with Pendal’s Australian equities team. Elise previously worked as an investment analyst for US fund manager Cartica where she covered a variety of emerging market companies.

She has also worked in investment banking and corporate finance at JP Morgan and Ernst & Young.

Pendal Horizon Sustainable Australian Share Fund is a concentrated portfolio aligned with the transition to a more sustainable, future economy.

Pendal Focus Australian Share Fund is a high-conviction equity fund with a 16-year track record of strong performance in a range of market conditions. The Fund is rated at the highest level by Lonsec, Morningstar and Zenith.

Pendal is an independent, global investment management business focused on delivering superior investment returns for our clients through active management. 

Contact a Pendal key account manager here