Thijme Schriek @DrWily: "- 𝗔𝗜 𝗜𝘀 𝗚𝗿𝗼𝘄𝗶𝗻𝗴, 𝗕𝘂𝘁..."

Netherlands

- 𝗔𝗜 𝗜𝘀 𝗚𝗿𝗼𝘄𝗶𝗻𝗴, 𝗕𝘂𝘁 𝗖𝗮𝗻 𝘁𝗵𝗲 𝗣𝗼𝘄𝗲𝗿 𝗚𝗿𝗶𝗱 𝗞𝗲𝗲𝗽 𝗨𝗽? - The growth of AI has been exponential, with every major tech company expanding its capacity as quickly as possible. Data centers, new models, integrations, and more are growing rapidly, and the primary resource is electricity. Just like the Bitcoin network, AI all around the world is consuming electricity and hardware at an alarming rate. New data suggests that places like America may struggle to keep up. So, let's take a look at it all! - 𝗧𝗵𝗲 𝗨𝗹𝘁𝗶𝗺𝗮𝘁𝗲 𝗕𝗼𝘁𝘁𝗹𝗲𝗻𝗲𝗰𝗸 - Now, AI is rapidly advancing, and this beast is insatiable, consuming electricity and chips at an alarming rate. However, can the US even sustain this? It's great to be at the forefront of this technology, but if you can't feed it, then at some point it will crash into a brick wall. This could stall growth and even allow other nations, such as China, to catch up or even surpass the US in AI. There are multiple significant bottlenecks that AI could face. For example, big tech companies could slow down spending because they run out of cash reserves, or real human data becomes more scarce, which will affect how fast you can train a model. Now there is a high chance that these will affect the growth of AI, but the bigger problem is going to be electricity. According to Goldman Sachs, the demand for electrical power to support AI will grow at a rate of 121% annually. From Q4 2023 to Q4 2024, this growth accelerated to 227%. Meanwhile, BloombergNEF expects global electricity demand from data centers to rise to 1,200 terawatt-hours by 2035 and 3,700 terawatt-hours by 2050. Clearly, the US needs to bring a lot more capacity online; even the US Department of Energy acknowledges this, stating that the US could face issues within 5 years. So, they want to work with major industry players to overcome this challenge, and Trump has even signed executive orders to address this issue. However, I think they won't be able to keep up. If you visit Our World In Data, you can see that the US produces 4,387 TWH of electricity, and only 290 TWH was added between 2014 and 2024. This would be nowhere near enough to support AI, let alone other innovations and industries, such as electric cars and large-scale manufacturing. Now, according to Goldman Sachs, AI can utilize stranded energy and pause models to work around this bottleneck, as the technology and users can adapt to this fairly well. However, it would still slow down AI's growth, and in the end, users always choose the fastest and most powerful provider. So, yes, this could help, but it won't really solve the issue and leads to various inefficiencies. But there is one other country that is in a much better position. - 𝗖𝗵𝗶𝗻𝗮 𝗜𝘀 𝗔𝗵𝗲𝗮𝗱 - Now, China has its problems. While their AI models are competitive, they lack the chips, and so far, they haven't been able to match Nvidia. However, Huawei does want to surpass Nvidia in chip technology, but it remains to be seen if it will actually succeed in this endeavor. However, they do have one MASSIVE advantage in one area: China produces more than enough electricity to meet the demand of AI. If you look at Our World In Data, you can see that they produce over 10,000 TWh of electricity and add an impressive 4,278 TWh from 2014 to 2024. This is almost as much as the US produced in 2024. Clearly, when it comes to this, they are ahead, and it could lead to China winning the AI arms race. Clearly, there is more than enough to worry about when it comes to AI, especially now that markets are at historic highs and incredibly overvalued. This is precisely why I hold some of our funds in cash and own hedges, allowing us to weather the storm while buying any dips. But only time will tell what happens to AI and this race between China and the US. Sources: fortune.com/2025/08/14/data-centers-china-grid-us-infrastructure/ www.goldmansachs.com/what-we-do/goldman-sachs-global-institute/articles/smart-demand-management-can-forestall-the-ai-energy-crisis www.energy.gov/sites/default/files/2025-09/55587_Speed_to_Power_FINAL.pdf about.bnef.com/insights/commodities/power-for-ai-easier-said-than-built/ www.energy.gov/articles/department-energy-releases-report-evaluating-us-grid-reliability-and-security ourworldindata.org/electricity-mix If you liked this short post, FOLLOW my profile for more; if you agree with my investing philosophy, COPY my portfolio to join the ride! My Performance speaks for itself! ✅ +40% in 2024! ✅ Well-diversified portfolio! ✅ Long-term mindset! ✅ Low-risk score of only 4! Be sure to copy my portfolio! $NVDA (NVIDIA Corporation) $TSM (Taiwan Semiconductor Manufacturing Co Ltd - ADR) $META (Meta Platforms Inc) $GOOG (Alphabet) $MSFT (Microsoft)