Griffin on Tech: Clouds, clout, and iPhone hot air
More hype on data centres arrived this week, courtesy of NZTech’s new data centre report, which crunches enough numbers to power a hyperscale server farm.
The data centre sector, we’re told, consists of 56 data centres scattered around the country, and already underpins $16.5 billion annually in “ICT GDP”, with indirect digital activity reportedly adding an additional $76.5 billion in “knowledge-intensive services”.
By 2030, the workforce could double as 20 new centres break ground, promising thousands of construction jobs and many billions more in economic activity.
If the numbers make your eyes glaze over, blame AWS after last week’s disastrous launch of its local data centre region produced heavy amounts of scepticism and puzzlement over the multi-billion dollar investment and GDP figures it reheated after its initial announcement in 2021. Listen to my podcast interview with cloud expert Ben Kepes for more on that sorry saga.
Certainly, data centres are important to the digital economy. But questions are increasingly being asked around the world about the implications of the data centre boom for energy generation and prices, as well as water usage.
The pivot to AI-centric data centres is leading to much higher demand from the sector for energy, and the hyperscalers will take as much renewable energy as they can get to hit their carbon reduction goals. The NZTech report suggests that data centres currently only account for 0.6% of New Zealand’s electricity usage and could rise to 2% by 2030. The government puts the figure as high as 7% under a high-growth scenario.
So to avoid data centre operators crowding out other industry players and households hungry for electricity, we need a significant increase in renewable energy generation to come online - and fast. How are we doing on that front?
As NZTech points out: According to the Ministry of Business, Innovation and Employment (MBIE), total national electricity demand is expected to reach around 50,503 GWh by 2035. To meet this growing need, 30 new electricity generation projects are currently committed and underway, set to add…approximately 2,928 GWh of additional annual electricity supply.”
It adds that a further 102 projects are being “actively pursued” that would add around 48,683 GWh of additional electricity supply. But the reality is that it is highly ambitious to expect the bulk of them to be consented and constructed in place by 2035. We just can’t get more renewables online fast enough and we are already witnessing the impacts of constrained electricity supply and high prices.
Source: NZTech
As Deloitte summed up last year: “The pace of renewable energy growth must match the rising demand from datacentres, AI, and nationwide electrification efforts. With substantial reconsenting of renewable energy generation assets over the next 10 years and complex compliance landscape for new renewable energy assets, the pathway to achieve this goal is far from secure.”
Then there’s the issue of water use, with AI data centres employing water cooling to pack in denser computing capacity. New Zealand could, in theory, become a data centre hub for the world, running workloads on renewable energy and sending the data offshore. But we’d really have to up our game on energy infrastructure to support that.
NZTech’s suggestion of a national working group to collectively plan data centre capacity and energy usage is a good one. One-off energy supply deals between data centre operators and energy generators may encourage more investment in renewables projects, but it needs careful coordination to ensure the rest of NZ Inc. has the energy it needs to stay viable.
The other issue that reared its head last week, with the scrutiny of AWS, is the extent to which revenues generated from cloud workloads run here by international hyperscalers are taxed locally. Typically, big tech companies use transfer payments and royalty charges to pay their offshore divisions large sums of money for software and services. That’s why they pay very little tax in New Zealand. Will that change as they increasingly use local hardware to deliver services? There are still questions to be answered around the real value this burgeoning sector of tech represents for Aotearoa.
Sweet AI dreams: NZ schools ready to plug in
With everyone talking about cloud jobs and digital futures, the Government is planning to add artificial intelligence training to the high school curriculum.
“Students will also be able to learn about and use generative artificial intelligence in a range of subjects, with the Government investigating creating a specialist subject on AI for Year 13 students in the future,” the Herald reported.
It’s all part of the replacement of NCEA, which will also seek to introduce new subjects like “civics, politics and philosophy” and “media, journalism and communications”.
Education Minister Erica Stanford’s curriculum revamp includes a new Year 13 subject focused on generative AI, digital systems, and machine learning ethics. It mirrors moves abroad, in the US, China, Singapore and Australia, where the consensus is that AI literacy needs to start much sooner, before students’ curiosity is replaced by chatbots and their handwriting by autocorrect.
The devil will be in the detail here - as one teacher pointed out on NZ this morning, there’s a big difference in teaching kids to use AI as opposed to giving them an understanding of how AI is built, which involves a lot of complicated maths. Does it make sense to offer AI as a class, or build it into multiple parts of the curriculum? That’s for the pedagogues to sort out, but signalling dedicated room for AI in classroom studies moving forward is a positive development that could boost our STEM efforts.
Apple’s iPhone Air: Lighter on innovation, heavier on wallet
Finally, the tech world’s annual parade of incremental upgrades took centre stage as Apple launched the iPhone Air, and, true to Apple form, it’s only Air in the way that its price tag makes your bank account feel lighter. The super-thin device goes on sale for $2,149 on September 19.
At this point, the real innovation is figuring out which kidney to sell to upgrade.
Apple’s other devices received mere tweaks, and there was hardly a mention of Apple Intelligence, the company’s stalled AI efforts. It felt to me, watching the unveiling of the new devices, like the end of an era, the last gasp of the smartphone before it is swept aside and replaced by something else. Will it be smart glasses, or some kind of AI-powered wearable? Who knows, but if the response to Apple’s underwhelming new line-up is anything to go by, we are ready for something new and the design gurus in Cupertino are now in an intense race with OpenAI, Meta and the Android phone makers to deliver it.
