The New 'Oil' in the AI Era — 'Electricity'! A Must-Have Opportunity for Investors in Energy Transition

This article focuses on the practice of trading options. This column focuses on the investment landscape, an opportunity to understand the outlook and teach parents how to use options to handle the opportunities of others. For those who are interested in this feeling, welcomeClick hereJoin Learning and receive a reminder when the next column is updated.

In the era of AI, is a place of rapid development in the city becoming the cornerstone of a new life? Goldman Sachs said that electricity has become the key bottleneck for AI's development as an asset.

Why? Because the explosion of demand for power is unprecedented in the context of the rapid development of AI, the supply side is facing problems, not only with the old and stable problems of the power grid system, but also in the tension and lack of power supply levels.

Why is the power bottleneck constraining the development of AI?

Let's see: How explosive is the demand?

The main thing to understand is that AI data centers have increased in power density over the past time, not only has the shape of the data changed, but the standards of all power supply systems have changed, and the tools are looking at some Institutions that measure the data.

OpenAI's ChatGPT consumes more than 0.5 million daily electricity, equivalent to 0.017 million households per day; the 0.1 million NVIDIA H100 chips required to train the GPT-6 model. Theoretically, the peak load ratio could be reduced by up to 20% if deployed centrally in a local area grid, and would actually need to be reduced by missing peak power or self-built storage. Impact.

By 2027, the power consumption of an AI Server Chassis is expected to be 50x that of the same cloud computing equipment 5 years ago. This drives the data center from a cloud data center 1.0 with a power of only 5-15 kW to a single-rack factory with a power of over 500 kWh.

By 2030, global demand for data center power is expected to soar by 160%. Between 2025 and 2026, only US domestic AI data centers are expected to add 6-13GW of electricity consumption per year, driving the peak load growth rate of electricity to four times the historical average, equivalent to the peak electricity demand of 10-20 million households annually.

So: How big is the supply problem?

One aspect comes from the aging of the power grid. Globally, especially in the United States, grid facilities are generally aging (average use 40 years) and are unable to withstand the computational demands of high-density explosive growth, and many devices are now nearing the end of their design life.

Risks in some regions have begun to ease, as PJM, one of the largest regional power supply markets in the United States, suffers from a major energy bottleneck in 2022, while 9 of the 13 regional power markets in the United States have reached global tightening status this summer, with institutions predicting the US by 2030, The electricity market in all regions of Guangxi will reach a state of frontier tightening.

On the other hand, the speed of self-power supply is not higher than that. Traditional energy sources such as Henry Hub Natural Gas plants have a construction week of 5-7 years, nuclear power stations require more than 10 years, and grid upgrades are long lasting, and short-term internal problems can relieve electrical pressure. Conservation technologies (such as Electrochemical Storage, Fuel Batteries, etc.) can relieve the pressure on the valley, but there is a need to promote integrated regional policies.

So: How serious is the problem?

According to the WTA Securities, it is expected that by the end of 2026, the American Electric Power System will experience a gap of 18-27GW in high-frequency generation capacity, which is similar to the total power currently available at 20 multi-site large test stations at the same time.

To make up for this gap, power grids need energy storage systems that not only provide the appropriate instantaneous power, but also provide continuous power for hours. It is estimated that the total energy storage capacity needed to be added in the next two years will reach 110-205GWh. Compared to 37GWh installed in 2024, this equates to 2-5 of the current size US energy storage market in two years.

The current issue of power shortages has raised the operating costs of AI enterprises, Technology companies are forced to sign up to High Stock Price Purchase Agreements, and the capacity price of PJM power market has reached a new decade high, all of these reasons. On top of the snow, the rise in electricity prices has led to strong restrictions, and the new AI data center project has been suspended.

What are the solutions behind the problem?

In the face of a severe power gap, the market is developing a consensus of integrated, multi-channel solutions.

In the short term (1-2 years), energy storage, especially electrochemical energy, is a key means of alleviating short-term power stress with the advantage of 1-1.5 years of rapid deployment.

So it goes like this: the current grid is not very useful during daylight hours, and it is scarce during night lows. When it is possible to charge batteries (energy storage systems) during the night and support AI connection on the antenna. This model can increase the energy efficiency of the US grid without the need to build new power plants. Tesla's current Megapack energy storage system is a case in point.

Under this logic, energy storage system integrators e.g. $Fluence Energy(FLNC.US)$ , Solid Oxide Fuel Battery (SOFC) Industry $Bloom Energy(BE.US)$ , Battery manufacturers such as $CATL(03750.HK)$ There are opportunities waiting for stocks.

In the medium term (3-5 years), energy storage is portable, but cannot create electricity, so when storage relieves peak pressure, adding a fundamentally new power supply is a must in the medium term.

The cost of generating electricity from solar and wind power sources such as renewable energy is relatively low, construction is relatively fast, and rice is not an unstable feature, and therefore requires this stable and adjustable source of energy to be replenished with pressurized stone.

Then the world's leading power technology companies such as $GE Vernova(GEV.US)$ , clean energy developers such as $NextEra Energy(NEE.US)$ , you will directly benefit from the installation of machines, while Henry Hub Natural Gas and power generation business such as $Vistra Energy(VST.US)$ , their valuation is also likely to be revalued.

In the long run (over 5 years), AI needs a perfect energy source that is both zero-carbon and stable, energy-dense, and extremely high. Nuclear power, especially the safer and more flexible Small Modular Reactor (SMR), provides massive zero-carbon electricity for 7x24 hours, perfectly matching the needs of data centers.

Companies that benefit then, including some large nuclear operators such as $Constellation Energy(CEG.US)$ , SMR Technology Company e.g. $NuScale Power(SMR.US)$ 、 $Oklo Inc(OKLO.US)$ etc. A number of companies in the entire Nuclear Industry Chain, in the middle and downstream, may have diving opportunities, with the symbol of the Industry Chain sign in direct view.

Other than that, the grid upgrade is an ongoing project, and this race takes an entire energy conversion cycle, because the outgoing electricity needs to be able to be delivered. Grid upgrades include transformers, transmission lines, and smart grid systems, where Copper is the key material that cannot or will not be missing.

Goldman Sachs predicts that grid and power infrastructure construction will drive global copper demand growth of around 60% by the end of 2030. This increase is due to the current increase in the consumption of one US based on global demand.

The global copper supply is in place for long-term bottling. Major producing regions, such as Chile and Peru, are declining, with fewer new mining developments and a 17 year development cycle. In the near term, Indonesia's Grasberg and El Teniente, Chile, caused a decline in global copper production by around 6% in 2025. Under the supply balance, Goldman Sachs cut the forecast for balanced global Copper supply in 2025 from just over 0.105 million tonnes to just under 0.0555 million tonnes.

For the upcoming price, Goldman Sachs expects Copper prices to reach $10750/ton in 2027 and to hit $11250/ton in 2028. The cutting requirements for grid upgrades are turning Copper into the new petroleum of the AI era. Related Beneficial Companies Examples: Global Minerals $Freeport-McMoRan(FCX.US)$ 、 $Southern Copper(SCCO.US)$ , can be attributed to the copper ETF symbol $Global X Copper Miners ETF(COPX.US)$ etc.

Dive into opportunities with options

Following a comprehensive investment framework for AI power, let's look at how to make a toolbox. By eliminating Direct Buy Stock Stocks, you can use this tool to increase your investment life while increasing profits and risk taking risks.

The following two symbols detail the useful options strategy strategy of the tool.

1、 $GE Vernova(GEV.US)$ : Bull Call Spread: Bull Call Spread Strategy

GEV is an energy company that is independent from General Electric. Business includes grid equipment, renewable energy and nuclear power, and its share price will be strong in 2025. For investors who are hoping to control costs with a bullish outlook, the Bull Call Spread is a better price spread strategy.

• Strategy Note: This is a hot and before-seen options strategy, created by a Call structure with different pricing options through simultaneous Buy and Buy and Sell at the same dates. The core of the strategy is to supplement the cost of Buy Options by using the income from the options, thereby reducing the risk of a whole investment and maximizing losses.

The IVs are not higher than the current IVs (the forward time interval that is not in the price), but above 89% of the historical percentage point, because IV is priced at the options price, but the IVs are higher than the buy option price, and we can expect that the same IVs will be reconsidered at the time of the reversal.

• Tool operation: Option due date can be selected 3-6 months after the expected replenishment time, preventing the time value of the stop options from eroding quickly. Buy Call's executive price lets you choose the price you like best, make Call's executive price look different, and see if you expect more money to be gained.

If this is the example in the figure, then the loss on the date (not counting manual costs) is: The maximum loss is the Net Cost at the time of Buy, which is approximately $247. The maximum profit is $670 when the share price rises, two options are subordinated, and income = 20*100-247=$1753. At $652.47, the share price reached a close balance.

The variation of the options price will be more varied before the date. You can use the option price calculator OptionPricecalCalculator, based on your forecasts for different dates and different IVs, and calculate options prices based on your forecasts for different dates and different IVs, so Measure the weight.

2、 $Freeport-McMoRan(FCX.US)$ : Cash-secured putCash reserve redemption option

FCX is one of the world's largest listed copper producers and will directly benefit from the long-term upward trend in copper prices. When investors are bullish on the long-term prospects of Copper Copper, hoping for a lower cost warehouse construction, they can look at the long-term outlook of Cash-Secured Put for the Cash Secured Put on the Reserve Term Option Put when hoping for a lower cost warehouse.

• Strategy Note: You can instantly earn a royalty income by taking out a put option (Put). If the stock price is higher than the exercise price, the options void, the equity is your profit. If the share price is below the operating price, you will acquire the right to buy Stock Stocks at the exercise price, regardless of the receipt of the rights, your actual stock holding positions is Share Price - Equity Fee. Currently, IVrelative to the respective highs, the opportunity to sell.

• Tool operation: You can choose a date of 1-2 months, and a shorter point in time can reduce the time value of options faster, so that the interest money can be deposited more quickly. If you have the intention to trade, the Executive Price can choose your Target Price. If you do not wish to trade, the Executive Price chooses a price that you know is unlikely to fail. It is best to use this strategy when it comes to deliberate delivery.

Needless to say, this strategy requires a certain amount of money, and more tools are needed, which can be seen in the following path.

• Delinquency (not counting recurring charges): There are usually two cases when the date is due. One is that the stock price is higher than the management price, the options are empty, so you earn the option money, the result in the example above is about $190. Another is that the shares are priced at the low exercise price, you have to buy Stock Stocks at the exercise price, Stocks need to be bought at the exercise price, and the actual cost of Buy is the option price - royalties, so the figure in the above example is 40-1.9=$38.1.

Finally, remind you to control the position well, how many Puts you can sell, not how many putts you can receive, so you need to set the number of Sell Puts according to how many you can receive, leaving some space for protection, otherwise you may be forced to close the position without expiration due to triggering account wind control.。

At the same time, it is not the strategy of GEV's uplift, that is, FCX's forwarding strategy, it is necessary to keep the company and the client's fundamentals in mind, in order to prevent further losses from expanding.

Okay, let's talk about it here today. welcomingClick hereJoin Learning and receive a reminder when the next column is updated. If you have specific content suggestions, you are also very welcome!

Finally, fellow investors who want to start investing in Hong Kong stock options in the current market, not firstClick hereGrab a new gift bag with options up to HK$2188 in value! But if you are new to options, you can practice handling these top opportunities by starting with Papertrade!