Leading global investment bank Goldman Sachs says the growing interest in generative artificial intelligence (AI) has resulted in an “arms race” to develop the technology.
This will require vast numbers of high-density data centres as well as the electricity to power them.
New data provided by Goldmans forecasts global power demand from data centres to increase by 50% by 2027 and by as much as 165% by the end of the decade.
James Schneider, a senior equity research analyst covering US telecom, digital infrastructure and IT services, said the balance between data centre supply and demand will tighten in the coming years.
“The occupancy rate for this infrastructure is projected to increase from around 85% in 2023 to a potential peak of more than 95% in late 2026,” Mr Schneider said.
“That will likely be followed by a moderation starting in 2027, as more data centres come online and AI-driven demand growth slows.”
“These trends will open up new opportunities for investors, particularly in hyperscale cloud providers, power utilities, asset managers and data centre operators,” he added.
Meanwhile, Australia’s Kalina Power (ASX: KPO) has expanded its secured land portfolio for the development of its projects to power AI data centre activity in Alberta, Canada, considered one of the world’s top data centre destinations.
The portfolio, which is being developed by the company’s 100%-owned Canadian subsidiary, Kalina Distributed Power (KDP), comprises multiple natural gas-fired combined cycle power plants incorporating carbon capture and sequestration (Power-CCS).
The company secured site control at two new locations in the Crossfield and Clairmont areas of Alberta by way of options to purchase until the end of 2027 and 2028.
The 180-acre Clairmont site has the capacity for one 170-megawatt Power-CCS plant, while the 320-acre Crossfield site has the capacity for at least three.
KDP has now assembled five secured projects representing a total of 785 acres, with the capacity to deploy at least 1,700MW (1.7GW) of Power-CCS plants.
Elsewhere, the S&P Global Commodity Insights Bitcoin Energy Consumption Index has edged lower after reaching an all-time high of 1,312.07 MW-hours per bitcoin mined in early February.
This index reflects the energy consumption per bitcoin mined using a standard graphics card.
S&P said higher bitcoin energy consumption is indicative of growing on-chain activity as more bitcoin miners join the network, contributing to a rise in the “hashrate”.
The hashrate represents the number of computations that mining hardware can perform per second while solving cryptographic puzzles and is generally perceived to enhance the security of the network, given the substantial amount of electricity consumed.
免責聲明:投資有風險,本文並非投資建議,以上內容不應被視為任何金融產品的購買或出售要約、建議或邀請,作者或其他用戶的任何相關討論、評論或帖子也不應被視為此類內容。本文僅供一般參考,不考慮您的個人投資目標、財務狀況或需求。TTM對信息的準確性和完整性不承擔任何責任或保證,投資者應自行研究並在投資前尋求專業建議。