Investing in Infrastructure
Infrastructure Quarterly: Q4 2024
December 17, 2024 10 Minute Read Time
Introduction
Associated Contact
Head of Infrastructure Research
Associated Contact
Senior Research Analyst
In this Q4 2024 edition of the Infrastructure Quarterly, we provide a year-end review of the economic impacts on infrastructure, unpack the latest investment trends and highlight the unprecedented scale of resources required to meet the rising demand for power and data. Concerns about diminishing returns from Generative AI models resurfaces with each earnings season, so we asked CBRE’s in-house Chatbot, Ellis AI, whether Generative AI is hype or reality.
Market performance
Listed markets outperform
In Q3 2024, the global capital markets were impacted by volatility due to mixed signals from the earnings season and the still-pending outcome of the U.S. presidential election. Inflation dropped significantly by 50 basis points in both Canada and the U.S., gravitating toward the newly established 2% floor for Western developed countries.
Easing inflation and gradual policy rate cuts set the background for outperformance of interest-rate sensitive assets such as listed equities. This past quarter was particularly successful for listed infrastructure which achieved a 30% one-year return (Figure 1), rebounding from a 5.3% return in the previous quarter. Listed infrastructure remained an attractive investment as a defensive, inflation-linked strategy and concluded the quarter on a strong note. Unlisted infrastructure recorded a positive, but more muted return over the last year as valuations integrated the lagging impacts of higher interest rates.
Figure 1: Infrastructure, bonds and equities annualized total returns (%)
In the middle of 2024, some investors were concerned about market overvaluation due to the uncertain long-term adoption of Generative AI. The big tech industry, focusing on four of the leading companies, performed exceptionally well (Figure 2) with Q3 earnings surpassing estimates, a well-diversified revenue stream and early incorporation of AI into their business mix. Nvidia, in particular, capitalized on the upswing with strong market dominance and high demand for its products. As supporting infrastructure for diverse data demands, data centers exhibited a lower but lately more consistent share growth than the big tech industry.
Figure 2: Share performance for major technology and data center companies
Infrastructure fundraising crawls up
In the first three quarters of 2024, fundraising gradually approached the levels seen over the past five years and is slowly gaining momentum (Figure 3). As we approach the end of the year, we anticipate a reasonable finish. Last year, the final quarter ended strong predominantly due to the final close of a few mega funds. The outlook for infrastructure fundraising is positive with nearly $200 billion expected to be raised over the next three quarters according to Infralogic. Among the 105 funds, the largest portion of fundraising is predicted for Q4 2024, with potential for an even larger amount as many targets are yet to be disclosed.
Figure 3: Infrastructure fundraising, value ($ billions)
In recent quarters, investment managers geared their fundraising approach toward North America, followed by Asia Pacific (Figure 4). Under President-elect Trump, the next U.S. Administration will target higher tariffs and protectionist policies. Looking at infrastructure funds in the market, a multi-regional focus remains the top choice targeting $251 billion, or nearly half of all desired capital. This approach enables investors to spread political and regulatory risk and hedge their bets by region.
Figure 4: Funds in the market, regional focus ($ billions)
Deals
Increasing interest in digital infrastructure
Deal values displayed a similar positive trend as fundraising increased at 4% year-over-year (Y-o-Y) (Figure 5). The year so far offered promising financing conditions for infrastructure, particularly with imminent rate cuts. However, year-to-date, M&A activity slowed likely due to the interest rate surge in the previous year and cautious investor sentiment.
Digital infrastructure investment grew by almost 30% compared to the same period last year with large ticket sales such as data center platform AirTrunk in Asia Pacific and Telecom Italia’s fixed telecoms network in Italy. According to Infralogic, the acquisition of AirTrunk commanded over AU $24 billion, which will make it the largest data center platform in the Asia Pacific region, pending approval. In Europe, Telecom Italia’s fixed network spin-off will be the first to separate the network from the service provider. The service consists of 550 telecom operators and captures 84% of the market share. The segment is valued at a maximum of EUR 22 billion (with earn-outs) and represents Italy’s backbone of the digital sector covering a majority of the country’s households and fiber cables.
In emerging markets, the Brazilian government announced an ambitious program to expand the country’s railway lines as greenfield projects. The program, introduced in May 2024, will modernize existing lines and add 890km of new lines totaling approximately R $200 billion. However, this ambitious project faces political challenges and raises questions about its longevity. The future of the project is challenged by the Brazilian electoral calendar which could later impact the schedule for deploying greener transportation.
Figure 5a: Private infrastructure dealmaking, values by deal type ($ billions)
Figure 5b: Private infrastructure dealmaking, market share by sector (%)
Q&A
Disclaimer: the following Q&A segment of this Infrastructure Quarterly was crafted with Generative AI support.
Like most of our readers, we experiment with large language models to test the power of cutting-edge technology. Our ChatBot, Ellis AI, delivered a well-informed and enlightening perspective on the future of its success in less than 15 seconds. Read the unedited answer below:
Is Generative AI hype or reality?
Generative AI is both a reality and a subject of considerable hype, and its perception can depend on various perspectives:
Reality:
- Technological Advancements: Generative AI has made significant strides, enabling the creation of text, images, music, and other media that are increasingly realistic and coherent. Models like GPT-3 and DALL-E have demonstrated tangible capabilities.
- Practical Applications: Businesses are adopting generative AI for various applications, including content creation, customer service (chatbots), design, drug discovery, and data analysis, showcasing its utility across sectors.
- Research and Development: Ongoing research continues to improve the capabilities and efficiency of generative models, leading to advancements in understanding and generating complex data.
- Economic Impact: Many companies report enhanced productivity and innovation thanks to generative AI tools, indicating its potential to contribute to economic growth.
Hype:
- Overpromising Capabilities: There can be exaggerated claims about what generative AI can achieve, leading to unrealistic expectations. Some applications may not be as robust or reliable as marketed.
- Ethical Concerns: The potential for misuse (e.g., deepfakes, misinformation) is often highlighted in discussions, creating a sense of apprehension that may overshadow the positive aspects.
- Job Displacement Fears: There is significant media focus on the potential for job losses due to automation, creating narratives that may not fully consider the new job opportunities that AI could create.
- Market Speculation: The rapid rise in interest has led to speculation in the tech industry, with startups attracting considerable funding basing on generative AI, which may not always be sustainable.
Conclusion:
Generative AI is real and has transformative potential, but it is essential to navigate the hype cautiously. Its capabilities are impressive, but understanding its limitations, ethical implications, and the context of its applications is crucial for a balanced perspective.
Sector insights
Power and utilities
Over nearly two decades, we observed a slow-growing divergence between stagnating U.S. electricity demand and growing real GDP. With the pressing need to power the adoption of Generative AI and complex language models, power demand is expected to soar. According to an October 2024 Wood Mackenzie report, analysts forecast a wide range of power demand growth from U.S. data centers—between 10%-20% until 2030, equivalent to 13-25 gigawatts of additional power capacity. Although other drivers contribute to power demand growth, digital infrastructure remains the primary one (Figure 6).
S&P Global Ratings recently said that the growth in data centers will drive up the electricity sales of North American investor-owned regulated utilities and support the industry’s credit quality. It also opens up questions about regulation and its efficiency and fairness. The new demand from data centers will necessitate additional capital spending and higher customer bills. Assigning the related data center costs to existing residential customers will result in disruptions and regulatory setbacks. The average monthly U.S. electric bill represents 2.2% of the U.S. median income and experts suggest that it should not exceed 2.5% for an efficiently managed industry.
Two other major drivers of demand are the renaissance of domestic manufacturing and the electrification of heating and transport. The next U.S. administration has pro-growth policies and higher tariffs on imported products and might encourage continued investment in local manufacturing of batteries, solar parts and semiconductors. The federal subsidies for electric vehicles (EV) are uncertain but many regions and states—California, the Northeastern and Western states—have local policies promoting EV adoption.
Figure 6: U.S. electricity load growth by customer segment, 2023-2028
Renewables
The United Nations Climate Change Conference (known as COP29) was held in Baku, Azerbaijan in November 2024. Ahead of the conference, the International Energy Agency published its forecast for global renewable capacity. The forecast projected a 70% increase in capacity additions in 2030 compared to the peak level in 2023, yet global renewables will fall short of the previous conference—COP28’s goal of tripling the installed capacity.
In Q3 2024, U.S. solar power benefitted from strong customer demand and rising power purchase agreement (PPA) prices to offset the premium-priced domestic components. Offshore wind is more challenging; however, over 30 gigawatts of offshore wind capacity was reported globally as of October 2024 (Figure 7). The auction process, which is more prevalent in Europe and Asia Pacific than in the United States, allows countries to secure electricity generated by renewable energy at competitive prices. France, leading in offshore wind activity, announced plans for 9 gigawatts of capacity in the coming months. Norway and Ireland began preparations for new offshore wind auctions with both governments supporting development through subsidies or site approval.
Figure 7: Auction capacity (MW) by region
The next U.S. Administration will shift policy priorities away from clean energy to nuclear and fossil fuels. Currently, the utility scale sector—projects larger than 5 MW—dominates the solar market. Capacity increased by 77% Y-o-Y in 2023 compared to the previous year according to the Lawrence Berkeley National Laboratory 2024 report. On a national level, solar only contributed 5.6% of the electricity generation in 2023 but it accounted for more than 50% of all new capacity added to the grid. Deployment is uneven across the U.S., with Texas leading the nation in newly installed capacity at 4.3 gigawatts but lagging on cumulative capacity (Figure 8). Furthermore, eight states in the U.S. observed a significant double-digit capacity additions. The Laboratory forecasts continued growth in annual solar deployment over the next six years.
Figure 8: Annual solar capacity additions by ISOs (GW)
Digital infrastructure
AI’s computing power requirements will drive a significant need for data center development over the next decade. Data centers are facing power transmission bottlenecks and strive to adhere to retiring fossil fuels to comply with climate policies. The supply shortage cannot be addressed with renewables in the short-term due to their intermittency, supply chain challenges and insufficient grid infrastructure.
During the past quarter, technology companies expressed interest in nuclear as a medium-term solution for meeting data center energy demands as on-site generation. As one example, Google signed a deal with Kairos Power to bring 500 MW of power from multiple small modular reactors (SMRs), with the first reactor planned to come online by 2030. Google believes nuclear energy can play a pivotal role in meeting 24/7 data center demand and eventually scale globally to deliver enormous benefits to power grids.
Schneider Electric suggested that SMRs—with capacity of up to 300 MW per unit—could be a solution offering lower capital costs, shorter deployment time and environmental benefits. However, the scalability and mass deployment of this technology remains uncertain.
Nuclear energy is the most abundant source of low-carbon electricity with two times more capacity factor than renewables (Figure 9) and could deliver optimal long-term green energy. Globally, many countries have adopted new reactors or reactivated idle reactors. Japan and France are restarting investment in nuclear power and developing more efficient nuclear fusion technologies. While regulatory requirements for nuclear have evolved, they remain inconsistent across regions. As a point of reference, the U.S. has only bought three reactors in the past twenty years.
Nuclear power also suffers from public safety concerns. Many early stage SMR technologies are vulnerable due to the designs of core power control systems and operational characteristics. A well-regulated and supervised technology would, therefore, be an ideal solution.
Figure 9: Capacity factors by energy source (2020)
Transport
The main theme for transport relates to the scale and urgency of the sector’s decarbonization. The electrification of transport continues at a strong but uneven pace, depending on the degree of government support and the related emissions regulations. S&P Global Ratings estimates that the global light electric vehicle mix will stand at 19%-20% in 2025 led by China but with softening penetration rate in Europe. Following the U.S. elections, the outlook is mixed. S&P predicts product launches in the U.S. at more affordable ranges and momentum due to the Environmental Protection Agency’s recent standards for vehicles from 2027-2032.
The investment in robust supply chains is another factor driving electrification. According to the Clean Investment Monitor, companies announced $89 billion in new investments in clean energy and vehicle technology manufacturing in the U.S. in the past two years. This is more than triple the amount invested over the previous two years with two-thirds of the investment in Q3 2024 geared toward the electric vehicle supply chain.
For other fuel technologies, the prognosis is more unsettled. Hydrogen can substitute fossil fuel in power, manufacturing and most importantly, transportation segments. It can be combusted with minimal associated greenhouse gas emissions. However, a study published by Harvard University in Joule on October 8, 2024 estimated costs for green hydrogen at $500-$1,250 per CO2 ton abated. A significant price drop would depend on the costs for storing and developing hydrogen which needs to decrease substantially. While U.S. subsidies and tax credits could alleviate some of the financial burden, it is dependent on the region. Industries might find it more cost-effective to invest in other carbon reduction strategies rather than facing expensive hydrogen costs.
Figure 10: Share of electric vehicles as % of total sales
Conclusion
Throughout 2024, the infrastructure sector was impacted by various market forces, including interest rate cuts, easing inflation and booming demand from key sectors. Investments in digital infrastructure are expected to grow into next year with continued investments into mainstream and alternative energy sources to meet the rising AI demand. Effectively managing supply and demand factors in an industry ahead of a capital expenditure supercycle will be crucial to capitalize on these opportunities.