Tag: Data centre

Posted on

AMD misses on data centre revenue – but profit and revenue are up

AMD

AMD shows mixed results in Q4 2024 earnings report

Advanced Micro Devices (AMD) has released its Q4 2024 earnings report, revealing a mixed bag of results that have sparked interest and discussions among investors and analysts.

While the company managed to surpass expectations in certain areas, it fell short in others, notably in data centre revenue.

Key highlights

  • Earnings per share (EPS): AMD reported an adjusted EPS of $1.09, slightly above the expected $1.08. This indicates a stable performance in terms of profitability, showcasing the company’s ability to manage costs effectively.
  • Revenue: The company posted a revenue of $7.66 billion, surpassing analysts’ estimates of $7.53 billion. This positive result highlights AMD’s continued growth and market presence, particularly in its core segments.
  • Data centre sales: Despite the overall positive revenue, AMD’s data centre sales fell short of expectations. Sales in this segment nearly doubled to $3.86 billion, reflecting a 69% increase year-over-year. However, this figure was below the anticipated $4.14 billion, signaling challenges in meeting the high demand and competition in the data center market.
  • Income: AMD reported a net income of $482 million, or 29 cents per share, down from $667 million, or 41 cents per share, in the year-ago period. This decline in net income suggests that the company faced increased expenses or other financial challenges during the quarter.

First quarter guidance

Looking ahead, AMD has provided guidance for the first quarter of 2025. The company expects Q1 sales to be around $7.1 billion, with a gross margin of approximately 54%. This forecast indicates cautious optimism, with AMD aiming to navigate the complexities of the semiconductor industry and maintain steady growth.

CEO’s statement

AMD’s CEO, Lisa Su, reportedly expressed confidence in the company’s future, particularly in the data centre AI market.

She highlighted the significant opportunities and potential for growth in this area, predicting strong double-digit revenue and EPS growth for 2025.

Her statement underscores AMD’s strategic focus on innovation and expanding its market share in high-growth segments.

Future

AMD’s Q4 2024 earnings report presents a nuanced picture of the company’s performance. While it has achieved notable successes in certain areas, challenges remain, particularly in meeting data centre revenue expectations.

As AMD continues to navigate the competitive landscape of the semiconductor industry, its future strategies and market positioning will be closely watched by investors and analysts alike.

Posted on

Amazon goes nuclear, to invest more than $500 million to develop small modular reactors (SMR)

AWS nuclear power

Amazon Web Services (AWS) has announced the signing of an agreement with Dominion Energy, the utility company of Virginia U.S., to explore the development of a small modular nuclear reactor near Dominion’s existing North Anna nuclear power station.

As Amazon’s cloud computing subsidiary, AWS has an ever-growing demand for clean energy, particularly as it expands into generative AI. This agreement aligns with Amazon’s journey towards net-zero carbon emissions.

Amazon joins other major tech companies like Google and Microsoft in turning to nuclear power to meet the increasing energy needs of data centres.

Posted on

Big Tech aiming to raise $100 billion for AI data centres

Fund creation for AI

In a substantial effort to strengthen the infrastructure required for artificial intelligence (AI), BlackRock and Microsoft have unveiled a significant fundraising endeavour.

The initiative, dubbed the Global AI Infrastructure Investment Partnership (GAIIP), seeks to secure $30 billion in private equity capital, with the possibility of leveraging up to $100 billion including debt financing.

The main objective of this initiative is to establish new and larger data centers to accommodate the escalating demand for computing power spurred by advancements in AI. These data centres are vital for meeting the growing computational requirements of AI applications, which necessitate substantial processing power and storage capacity. Additionally, the partnership will focus on investing in the energy infrastructure required to operate these data centres in an environmentally sustainable manner.

BlackRock, the global investment management corporation, contributes its vast network of corporate relationships and private equity expertise. Microsoft, a pioneer in technology and AI, offers the necessary technological expertise and industry leadership. Together, their goal is to establish a strong infrastructure that will bolster AI innovation and contribute to economic expansion.

The investment will be primarily directed towards the United States, with a portion also being allocated to partner countries. This strategic emphasis aims to boost American AI competitiveness and encourage worldwide cooperation. The partnership is designed to support an open architecture and a wide-ranging ecosystem, enabling a variety of partners and companies to leverage the infrastructure.

NVIDIA, a leading force in AI technology, will contribute to GAIIP by providing its expertise in AI data centres and manufacturing facilities. This partnership is anticipated to improve AI supply chains and energy procurement, offering advantages to both consumers and the broader industry.

This collaboration marks a substantial move towards establishing the infrastructure of tomorrow and powering it in an eco-friendly manner.

Posted on

AI power hungry data centres go green in Singapore

AI thirst for power

Singapore is actively addressing the dual challenges of the increasing demand for artificial intelligence (AI) and the pressure on energy resources. The nation recently unveiled a green data centre roadmap to bolster its digital economy aspirations.

Growing AI Demand and Energy Strain

With the increasing demand for digital and AI computation, the requirement for data centre capacity has grown. This heightened demand is exerting pressure on national energy networks, calling for urgent attention.

Goals

The Green Data Centre plan is designed to deliver a minimum of 300 megawatts of extra capacity. Singapore’s strategy includes improving energy efficiency in all data centres, implementing energy-efficient IT equipment, and providing incentives or grants to promote resource efficiency.

Sustainability

Singapore acknowledges the substantial contribution of data centres to its ICT sector’s emissions, accounting for 82% of the sector’s emissions and 7% of the nation’s total electricity consumption. By prioritizing energy efficiency, renewable energy, and sustainability, Singapore strives to maintain its position as a premier digital hub while reducing its environmental impact.

International status

Singapore’s data centers capitalize on the nation’s global standing as a commercial and digital centre. The city-state ranks as the second-largest data center market in Southeast Asia and holds the sixth position in the Asia-Pacific region.

In conclusion, Singapore’s focus on green data centres demonstrates its pledge to sustainable development amidst increasing energy needs driven by AI. With an emphasis on energy efficiency and renewable energy, Singapore seeks to harmonize technological progress with ecological stewardship.

Posted on

Energy hungry data centre power solution

AI data centre

The use of nuclear reactors for data centres is a controversial and complex topic that has both advantages and disadvantages

Nuclear reactors can provide a reliable, stable, and carbon-free source of electricity for power-hungry data centres, which are essential for the operation of various applications, such as artificial intelligence (AI).

Grid overload

Nuclear reactors can also reduce the dependence on the existing grid, which may be vulnerable to blackouts, fluctuations, or cyberattacks. On the other hand, nuclear reactors require a high initial investment, as well as strict safety and regulatory standards. Nuclear reactors also pose potential risks of radiation, waste disposal, and proliferation. Moreover, nuclear reactors may not be suitable for all locations, as they may face public opposition, environmental concerns, or geopolitical issues.

Small Modular Reactor (SMR)

One of the possible solutions to these challenges is to use small modular reactors (SMRs), which are advanced reactors with about a third of the power generation of a traditional, large nuclear plant. SMRs are designed to be more flexible, scalable, and cost-effective than conventional reactors, as they can be built off-site and transported to the desired location. SMRs can also be integrated with renewable energy sources, such as solar or wind, to create a hybrid system that can balance the power demand and supply.

However, the technology of SMRs is still in its early stages of development and deployment, and there are currently no data centres in the world that use built-in nuclear reactors. Therefore, it remains to be seen whether nuclear reactors will become a common or viable option for future data centres. The decision to use nuclear reactors for data centres should be based on a careful evaluation of the benefits and risks, as well as the alternatives and trade-offs, of each specific case.

It has been calculated that a ‘norma’ data centre (whatever that is), needs 32 megawatts of power flowing into the building. For an AI data centre, it’s closer to 80 megawatts.

AI systems are using all this extra electricity simply because they are doing so much more processing than standard computing. They are chewing through far more data.

As AI continues to develop, so too will the power requirement needed to run these monsters.