可再生能源

The Real Reason Big Tech Companies Are Investing Billions in the Solar Industry

Editor
7 分钟阅读

As of 2025, the global investment in renewable energy by big tech companies has surpassed $20 billion annually, with the solar sector accounting for approximately 60% of this. Amazon Web Services (AWS) operates over 400 solar projects worldwide, while Google is directly investing in large-scale solar farms in Texas and Nevada, aiming to achieve 24/7 carbon-free energy by 2030. Microsoft is also expanding its solar power purchase agreements (PPA) in desert regions to achieve carbon negativity by 2030, and Meta has announced that 100% of the power required for its data center operations will be sourced from renewable energy.

The Real Reason Big Tech Companies Are Investing Billions in the Solar Industry
Photo by DALL-E 3 on OpenAI DALL-E

However, industry experts analyze that these investments in solar energy by big tech are not simply to operate data centers solely on solar power. In reality, the power demand of AI data centers is continuous around the clock, and especially large language model (LLM) training and inference tasks require stable power supply day and night. Considering the intermittency issues of solar power generation, the consensus in the industry is that the reasons big tech companies are praising solar power are far more strategic and complex.

The Most Practical Choice for Responding to ESG Regulations

The first reason big tech companies are focusing on solar power is that it is the fastest and most economical solution for ESG (Environmental, Social, and Governance) regulations and carbon neutrality obligations. Google, Amazon, Microsoft, and Meta have all pledged to achieve carbon neutrality (net zero) between 2030 and 2040, which requires significantly reducing the carbon emissions of the power used for data center operations. According to the International Energy Agency (IEA), as of 2024, the global power consumption of data centers is about 460 TWh, equivalent to the total power consumption of Argentina.

When analyzing carbon emissions by power source, construction speed, capital expenditure (CAPEX), and political resistance, the superiority of solar power is clear. The carbon emissions of solar power generation are nearly zero at 40-50g CO2eq per kWh, and the project construction period is very fast, ranging from 6 to 18 months. In contrast, nuclear power plants have near-zero carbon emissions but take more than 10 years to build, with initial investment costs of $7-12 million per MW, 5-8 times higher than solar’s $1-1.5 million. Natural gas is quick to build but unsuitable for ESG goals due to its high carbon emissions of 490g CO2eq per kWh.

In fact, an analysis of the renewable energy power purchase agreements (PPA) signed by big tech companies shows that solar accounts for more than 55% of the total. Amazon alone signed solar PPAs totaling 8.5 GW globally in 2024, enough to power about 2 million households. Google also announced participation in 5.2 GW of solar projects in the U.S. and Europe in 2024, expecting to reduce CO2 emissions by about 8 million tons annually.

Minimizing Political Risk and Securing Long-term Economic Viability

The second key reason big tech companies prefer solar power is the minimal political risk. If a data center company pursues the construction of a nuclear power plant, it must consider numerous variables, including government approval processes, opposition from civic groups, complex environmental impact assessments, and long-term regulatory risks. In the U.S., the approval process for new nuclear plants alone takes an average of 7-10 years, and delays due to local resident opposition and lawsuits from environmental groups are frequent during construction. In contrast, solar power plants have relatively simple permitting procedures and minimal environmental impact, resulting in little local opposition.

Additionally, the continuously declining levelized cost of electricity (LCOE) for solar power is a significant factor in its long-term economic viability. According to the 2024 report by the International Renewable Energy Agency (IRENA), the LCOE for solar power in the U.S. is the cheapest at 3-4 cents per kWh, compared to wind at 4-6 cents, new nuclear at 12-18 cents, and natural gas at 8-12 cents. Notably, the price of solar modules has dropped by over 85% in the past decade and is expected to decrease by an additional 20-30% by 2030.

The power demand patterns of big tech companies also show significant synergy with solar power. Generally, AI training tasks continue 24/7, but actual service traffic and backend processing tasks tend to concentrate during daytime and evening hours. According to internal data from Google Cloud, the power demand for search and YouTube services is concentrated between 9 AM and 6 PM local time, accounting for about 65% of total daily demand. This largely aligns with the peak solar generation hours, enhancing grid operation efficiency.

Microsoft has developed an AI workload scheduling system to optimize these load patterns, assigning more computational tasks during the solar-rich daytime. This approach reportedly reduces overall power costs by about 15-20%. Amazon also announced that it adjusts some AWS batch processing tasks to align with solar generation patterns, saving approximately $300 million annually in power costs.

However, industry experts emphasize that the ultimate goal of big tech companies is not to operate solely on solar power but to adopt a ‘solar+nuclear’ combination. Their long-term strategy is to reduce costs and carbon emissions with solar during the day and ensure stability with small modular reactors (SMR) for nighttime and base loads. In fact, Microsoft has signed a power supply agreement with fusion startup Helion Energy starting in 2028, and Amazon announced a large data center investment near the Susquehanna nuclear plant in Pennsylvania. Google also secured a priority purchase agreement with California-based SMR developer Kairos Power to acquire nuclear power in line with the commercialization timeline in the early 2030s.

The massive investment by big tech in the solar industry is directly impacting the performance of related companies. U.S.-based solar module manufacturer First Solar reported $870 million in revenue for the third quarter of 2024, a 28% increase from the same period last year, with a significant portion stemming from long-term supply contracts with big tech companies. Inverter specialist Enphase Energy also saw its stock price rise by 45% compared to the beginning of the year due to increased demand for large-capacity inverters for data centers.

The domestic solar industry is also benefiting from this global trend. Hanwha Solutions is expanding its solar module plant in Georgia, USA, with a production capacity of 3.3 GW, and expanding supply contracts with Amazon, Google, and others. The company announced that its solar business revenue is expected to increase by 35% year-on-year in 2024. OCI is also strengthening its position in the North American solar market supply chain by expanding its polysilicon plant in Malaysia.

Ultimately, the reason big tech companies are praising the solar industry is not to operate data centers solely on solar power but as a strategic choice to respond to ESG regulations, minimize political risk, secure long-term economic viability, and ultimately build an optimized power portfolio through a combination with nuclear power. This trend is expected to continue for the next 5-10 years and serve as a key driver for the sustained growth of the solar industry. However, considering the intermittency and storage cost issues of solar power, as well as the 24-hour power demand characteristics of AI workloads, the industry foresees that a combination with nuclear power will be inevitable in the long term.

This article is intended for informational purposes only and is not a solicitation to invest or a recommendation of specific stocks. All investment decisions should be made at the individual’s discretion and responsibility.

#First Solar #Enphase Energy #SolarEdge #Hanwha Solutions #OCI

Editor

Leave a Comment