power system.” [29]
making gas-powered electricity generation a more attractive option for private sector actors to meet the energy demands of AI data centers.
Today, the increased global demand for gas turbines is creating supply chain constraints. [30] The issue stems from the ability of the handful of suppliers of the turbines to keep up with demand. Currently, three companies will supply most of the current demand: GE Vernova, Siemens Energy, and Mitsubishi Power. [31] Due to increased demand, these companies have extended their delivery timelines. [32] “Mitsubishi states that turbines ordered today will not be delivered until 2028–2030. Siemens reports a record backlog of € 131 billion (US$148 billion). And GE Vernova has announced new turbines will not be available until late 2028 at the earliest.” [33] Thus, given the increasing demand for power consumption, if the supply for turbines does not increase, utilities facing delays in production might be unable to keep up with consumption. Additionally, costs for turbines available for purchase have increased, therefore, inadvertently costing the ratepayer more money. [34] However, there are alternatives we can use to help as near-term solutions, including: energy efficiency solutions, virtual power plants, grid-enhancing technologies, clean pre-powering, and hybrid “power couples” sited at existing fossil generator points of interconnection. [35] Therefore, despite the short supply of turbines, natural gas is well positioned to meet the current and future demand of AI data centers. Carbon Capture and Sequestration (CCS) for Natural Gas-Powered AI Datacenters
Background
Carbon capture and sequestration (“CCS”) technologies can capture up to 90-95% of CO 2 emissions from large natural gas-fired power plants. [36] The US electrical grid emits 340-420 kg CO 2 e/ MWh on average; but when gas-fired power plants are paired with CCS technology, a gas-powered plant emits 80-120 CO 2 e/MWh on average. [37] While renewables and nuclear emit less CO 2 , gas-powered plants paired with CCS are more dependable and flexible, while also being cost effective at $70-100/MWh compared to $77/MWh for nuclear and $87/MWh for solar. [38] Without CCS technology, gas-powered plants cost $37/MWh on average, but CCS is necessary for gas-powered plants to be environmentally sustainable on a large scale. Even at $70-100/MWh, gas-powered plants paired with CCS technology are fiscally competitive with nuclear ($77/MWh) and solar ($87/MWh). Moreover, scale and regulatory incentives will further reduce costs of gas-powered-CCS plants. [39]
Regulatory Environment Ripe for Investment
Section 45Q Federal Tax Credits
Enacted in 2008, Treasury Regulation § 1.45Q (“Section 45Q”) incentivizes qualified facilities that capture and permanently sequester or utilize carbon emissions, including emissions from natural gas-fired power plants, through federal tax credits. [40] Under Section 45Q, facilities must begin construction before January 1, 2033 to be eligible to claim tax credits for the 12-year period after the project is placed into service. [41] Specifically, for natural gas-fired power plants, the facility must capture at least 18,750 metric tons per year of carbon emissions, with equipment designed to capture at least 75% of baseline emissions, to qualify for Section 45Q tax credits. [42]
Introduction
While natural gas-powered electricity is a key component to meet AI data centers’ unprecedented demand for electricity, carbon emissions are a primary concern. Carbon capture and sequestration is paramount to curbing carbon emissions, but carbon capture and sequestration technologies face challenges as they increase project timelines, raise costs significantly, and are highly regulated by federal and state governments. However, regulatory and tax incentives from the federal government are reducing barriers to entry,
As previously mentioned, enacted in July
8
N at i onal A ssociation of D i v i s i on O rder A nalys t s
Made with FlippingBook interactive PDF creator