How I use AI for investing

AI can hallucinate (make up stuff) and lie, however it can be very useful in getting ideas on investing.

Many analysts reduce the current global economic cycle to four big forces:

• AI & Digital Infrastructure
• Energy & Power Systems
• Geopolitics & Defense
• Demographics & Societal Change

Current Thematic Trends

For instance, we can use it to investigate themes and trends (thematic investing). The following categories were all returned by ChatGPT:

AI and data centres - chips, robotics, AI, self-drive, chat/support, HR

Energy/Power - renewables, SMRs, grid infrastructure, SMRs, battery storage, natural gas generators, solar, wind, hydrogen, carbon capture

Minerals and Commodities - Lithium, Copper, Nickel, Rare earth elements, Uranium, Silver

Defence - AI warfare systems, Cybersecurity, Drones, missile defense, space defense

Semi-Conductor supply chains - Foundry expansion (US, Taiwan, Korea), Advanced packaging, AI accelerators, chip sovereignty

Robotics and Automation - Manufacturing, Warehousing, agriculture, construction, logistics

Data Infrastructure - Cloud computing, fiber networks, data centres, edge computing, 5G / future 6G networks

Healthcare and longevity - biotech and gene editing, AI drug discovery, longevity research, obesity / diabetes treatments, digital health

Reshoring - geopolitical issues, supply-chain reshoring, industrial policy, trade blocs, national tech ecosystems. e.g. Moving away from dependency on the USA.

Infrastructure - transportation, electrification, ports and shipping, water systems, construction technology

Precious metals and inflation hedges - gold, silver, commodity producers, real assets

Computing Technologies - quantum computing, photonic computing, neuromorphic chips, spatial computing (AR/VR), AI hardware accelerators

Investing in ETFs

I can ask ChatGPT to suggest ETFs for any of these categories, however this is not always a good idea.

I tend to go for 'picks and shovels' ETFs rather than ETFs that include large companies which may compete with each other.

For instance, I think copper will be in high demand for use in grid infrastructure, motors, generators, cooling in data centres. The biggest demand is in grid infrastructure (by about x7). A global deficit is expected in 2026/27 and at least 80 new copper mines will be required by 2030. Some of the demand will be supplied by recycling of old telephone wires, old BEV motors, etc. So I think that rather than picking companies that actually build grid infrastructure equipment, make copper wires or install grid infrastructure, it will be better to invest in copper mining companies and perhaps copper recycling companies.

So this now becomes a matter of choosing which companies to invest in and as well as looking at their performance figures, I need to consider where their mines are located. A copper mine in Canada is going to be more stable than one in a small country in Africa, for instance. Currently, I have invested in Fresnillo and Glencore as well as the COPX ETF (which is cheap at the moment).

The alternative might be to go for large grid infrastructure companies, but that would mean trying to pick winners (e.g. I have picked GE Vernova GEV or Siemens ENR), but when picking companies, the key is to buy early because many other investors (esp. large investment companies) will have inflated the prices already. Also, these companies will have many forward contracts and they are at full capacity, so there will be little scope to exceed their estimated profits. I invested in GEV 4 months ago (up +40%) and Siemens ENR in early Jan 2026 (up +20%).

Other ChatGPT tactics - AI

AI and datacentres are another thematic choice. I could invest in the obvious candidates like Palantir and nVidia, but these currently have high P/Es  (200+ and 37). Intel and AMD are also in the frame, but both seem to be struggling recently. Rather than backing these two favourites with poor odds (now), I prefer to look at the 'Picks and Shovels' angle again.

What is it that these data centres will need besides copper wires. The answer is that they will all need disk storage and memory. These are in high demand and only a handful of companies make disks and memory. ChatGPT will tell us which ones. Of particular interest to me are the non-USA companies (e.g. the Korean companies Samsung and Hynix) as well as the USA companies of Western Digital, SanDisk, Micron and Seagate. I have bought all of these and they are doing well.

Due to the sad state of affairs in the USA and Trump's tariffs, I have also invested in a few Korea ETFs (CSKR and IKOR - top holdings Samsung, Hynix, Hyundai).

Similarly, I have invested in gold and silver mining ETFs as well as a few individual miners. The demand for silver should increase (esp. for solar panels) and gold is being bought as a stable reserve currency. I don't see demand falling this year.

Another area is Defence. There are many 'picks and shovel' companies but they are all reasonable small and disparate. Individual companies include Leonardo, Thales, Rolls Royce. I hold the ETFs WDEP and DFNG. As defence involves destroying buildings and infrastructure, large plant equipment will be require for rebuilding. ChatGPT can help in identifying companies which may be currently undervalued (shown in red).

• Construction & Engineering Strabag, Webuild, Vinci, Bechtel, Fluor Corp
• Heavy Machinery Caterpillar, HD Hyundai, Komatsu
• Building Materials Holcim, Heidelberg Materials, Wienerberger, Cementir

I have invested in CAT so far and also Siemens. 

Summary

Using ChatGPT, you can not only drill down to find ETFs and companies, but also find the 'picks and shovels' that each industry requires.

The recent war with Iran will affect the price of oil, LNG, aluminium and fertilizer. China, S Korea, Japan and Thailand are most affected by the oil shortage. But hopefully, the Iran situation is short term and it is too volatile for me to invest in. 

After a war will come rebuilding, so for the long term, investing in companies that clear and rebuild may be a safer bet.

Factors to consider

Before investing in a particular company or ETF, consider:

• Market size
• Longevity
• Risk factors - how could it fail?
• Value - is it good value to buy now or have I missed the bus?
• Government/geo-political influence - governments can be fickle
• Alternative technologies - e.g. Sodium Ion batteries do not require Lithium or Cobalt but rare earths cannot be substituted.
• Growth potential
• Performance indicators
• Picks and shovels? Everyone will need to buy picks and shovels in a gold rush - or are you betting on one mine out of many?