Preface:

The rise of AI infrastructure indicates that the future AI ecosystem will no longer be limited to software and algorithm innovation, but will extend to all aspects of hardware, energy, and infrastructure.

In today's rapidly developing era, the evolution of AI has gone beyond simple algorithm and model optimization. It is about how to comprehensively build and maintain a broad and solid infrastructure system that can support its efficient operation.

Author | Fang Wensan

Image source | Network

A 700 billion largest AI fund may take shape

Recently, BlackRock, a leading global asset management company, partnered with Microsoft and joined forces with Global Infrastructure Partners (GIP), which specializes in infrastructure investments, and MGX, an AI investment company under the Abu Dhabi government, to jointly announce the establishment of the Global AI Infrastructure Investment Partnership (GAIIP), a major investment initiative.

The fund aims to address the infrastructure and energy demands posed by the rapid development of the AI industry.

The GAIIP fund focuses its investments on two key areas: new and expanded data center projects, and the supporting energy infrastructure required to power these data centers.

Moreover, the GAIIP fund embraces an open and collaborative investment philosophy, committed to promoting open architecture and a broad ecosystem.

This strategy ensures that various partners and companies have comprehensive and in-depth access and collaboration opportunities on a non-exclusive basis.

Initially, the fund will seek to raise up to $30 billion in private equity capital and plans to increase the total investment potential to $100 billion through debt financing and other means.

According to the official announcement, the GAIIP fund will adhere to the principle of open collaboration, support open architecture, and build a broad ecosystem to ensure that various partners and companies have non-exclusive and comprehensive access.

As a seasoned expert in infrastructure investment, GIP has extensive project management and operational experience, which undoubtedly lays a solid foundation for the stable operation and long-term development of the GAIIP fund.

NVIDIA, as one of the fund's key supporters, will leverage its deep expertise and experience in AI data centers to provide strong technical support and guarantees for the fund.

Microsoft stated that the fund will leverage its technological advantages in cloud computing to provide non-exclusive comprehensive access to numerous partners, accelerating AI technology research and development, and jointly promoting AI technological progress and development.

MGX, a technology investment pioneer from the Middle East, is dedicated to accelerating the research and application of AI and advanced technologies by building a leading partner network in the UAE and globally.

AI's development requires more powerful energy support

It is well known that AI models, especially those used for deep learning and processing large-scale data, are highly dependent on supercomputing power.

Therefore, AI companies often build powerful computing platforms by connecting thousands of AI chips.

However, this significant computing power is often accompanied by notable energy consumption issues.

It is estimated that the electricity consumed during the training process of some large language models is comparable to the annual electricity consumption of thousands of ordinary households.

This highlights the importance of ensuring stable and adequate power supply for AI technology research and development.

The International Energy Agency predicts that global data center electricity consumption may exceed 1,000 terawatt-hours by 2026, doubling from 2022 levels.

Taking the United States as an example, the U.S. Department of Energy predicts that electricity demand will increase by 15% to 20% over the next decade.

According to a report by a nonprofit electricity research institute, by 2030, data centers will consume electricity annually accounting for 9% of total electricity generation in the United States.

In response to the surge in energy demand driven by the AI industry, U.S. energy companies are actively responding, with some even planning to postpone the retirement of coal and natural gas power plants and accelerate the construction of clean energy facilities such as solar and wind power.

The energy supply issues arising from the AI industry have attracted significant government attention.

On September 12, tech industry executives including NVIDIA CEO Jensen Huang and OpenAI CEO Sam Altman gathered at the White House with senior Biden administration officials and industry leaders to discuss strategies for addressing the large-scale infrastructure demands of AI projects.

After the meeting, the White House announced the formation of an interagency working group aimed at promoting the development of U.S. data centers and accelerating the approval process for related facilities.

Simultaneously, the U.S. Department of Energy plans to establish an AI data center team to support the development of this field through various projects.

Industry experts point out that the electricity required for AI computing power far exceeds that of previous technological innovations, posing a significant challenge to existing energy infrastructure.

BlackRock CEO Larry Fink clearly stated that leveraging private capital to accelerate the construction of data centers and power facilities is crucial for unlocking long-term investment potential and activating trillion-dollar market opportunities.

He further elaborated that data centers, as the cornerstone of the digital economy, will not only promote sustained economic expansion and job growth through investments but also deeply drive the innovation and development of AI technology.

To this end, the GAIIP fund emerged, poised to provide necessary power infrastructure support for the AI sector.

As a standout among its peers, the GAIIP fund further underscores Wall Street's continued investment interest and confidence in the AI sector.

Marking a significant strategic deployment for AI infrastructure

GIP, one of the initiators, has been a prominent investment institution in the infrastructure sector since its establishment in 2006, with investments spanning key areas such as transportation, energy, digitization, sewage and waste treatment.

To date, GIP's assets under management have exceeded $100 billion, with its latest flagship fund successfully launched in 2019, exceeding $22 billion in size.

In addition to equity investments, GIP actively expands its credit strategy business, with a portfolio spanning over 40 invested enterprises generating combined annual revenues exceeding $75 billion.

Another key initiator, GP, is MGX, a technology investment institution co-founded by Microsoft and members of the UAE royal family.

MGX was officially established on January 22, 2024, with an investment strategy focused on AI industry infrastructure, particularly on two core tracks: semiconductors (covering the design and manufacturing of logic and memory chips) and core AI technologies and applications (including AI models, software, data, life sciences, and robotics).

Sheikh Tahnoun bin Zayed Al Nahyan serves as Chairman of MGX's Board of Directors and concurrently holds the position of Deputy Crown Prince of Abu Dhabi.

In January of this year, BlackRock announced a significant acquisition plan to acquire GIP in its entirety for a total consideration of $3 billion in cash and 12 million shares of BlackRock common stock, with the transaction expected to be completed on October 1.

Upon completion of the acquisition, GIP's four founding partners, including Bayo Ogunlesi, will join BlackRock's Infrastructure Investments division in management roles.

This acquisition not only marks the expansion of BlackRock's infrastructure investment team but also increases the division's assets under management to $150 billion.

BlackRock has high expectations for this division, aiming to provide market-leading, comprehensive infrastructure expertise across equity, debt, and large-scale solutions for transforming markets globally.

Indeed, [infrastructure] has emerged as an increasingly prominent theme in the AI sector, attracting significant attention and participation in investment deals.

For instance, Microsoft announced three major investments in April this year, targeting cloud computing and AI infrastructure in Japan, Malaysia, and Indonesia, totaling billions of dollars.

Furthermore, renowned angel investor Daniel Gross is also an active participant in this field, having not only invested in former OpenAI Chief Scientist Ilya Sutskever but also joining the relevant company as a partner.

Conclusion:

A recent forecast report by Barclays Bank indicates that global capital expenditures in the AI sector will cumulatively reach a staggering $167 billion from 2023 to 2026.

Behind the thriving AI technology lies a double-edged sword phenomenon that cannot be ignored.

This technology not only brings unprecedented convenience and vast development opportunities but also poses a series of potential risks and challenges.

As institutions actively leverage AI technology to drive their rapid development, they must also carefully consider and formulate scientific and reasonable risk management and regulatory strategies to ensure the robustness and sustainability of technological applications.

Some references: Global Hotspot: "Microsoft and BlackRock establish a $30 billion fund dedicated to building supercomputing centers, will NVIDIA take off again?", "BlackRock teams up with Microsoft to launch a $30 billion AI investment fund", AskCore: "Microsoft, BlackRock, and others launch AI fund, with support from NVIDIA", ChinaVenture: "$700 billion, the largest AI fund in history is coming", Zheshang Innovation and Entrepreneurship: "$100 billion, the world's largest AI investment fund is born", Venture Capital Review: "Microsoft partners with BlackRock to establish the world's largest AI investment fund", Zhilingyun Technology: "Microsoft and BlackRock jointly establish the world's largest AI investment fund"