Tech-N-AI Talks logo Tech-N-AI Talks

AI Doom vs AI Boom: Prosumer Investing Guide

AI risk warnings are getting louder. Here is how prosumers can invest in the AI boom without ignoring AI safety and AI doom scenarios.

AI Doom vs. AI Boom: What Prosumers Need to Know Before Investing — illustrative featured image
The Centre for the Governance of AI once asked a room of machine learning researchers a simple question: what probability do you assign to "humanity fails to control advanced AI"? The median answer came back around 5%. A follow-up study put the number higher. Last week, one researcher told CNBC the odds of AI "killing all humans" sit above 10%. That number deserves a beat of silence. Ten percent is not a fringe prediction from a guy with a podcast. It is roughly the same probability meteorologists assign to a coin-flip thunderstorm on a humid July afternoon. You would bring an umbrella. Yet plenty of us are dumping our savings into AI stocks with the casual confidence of someone who has never checked the forecast. This is the tension every prosumer now lives inside. The same technology that might, in the worst case, end civilization is also the best wealth-creation engine of our lifetimes. Both things can be true. The question is what you actually do about it on a Tuesday afternoon when your brokerage app is open. ## The two camps, and why both are partly right ### The doom case The AI safety crowd has gotten sharper over the past two years. Their argument is not that a chatbot will become sentient and hunt you down. It is that we are building systems whose capabilities are scaling faster than our ability to verify what they want, and we are wiring them into critical infrastructure while we do it. Alignment research is underfunded relative to capability research by a wide margin. Nobody has a working theory for how to guarantee that a superintelligent system shares our values. The 10% figure is a subjective probability, not a measurement. But subjective probabilities from domain experts are how we price nuclear risk, pandemic risk, and asteroid risk. Ignoring them because they make us uncomfortable is not a strategy. ### The boom case The other camp points at revenue. Nvidia's data center business went from a rounding error to the majority of its revenue in under three years. Microsoft, Google, and [Amazon](https://www.amazon.com/) are all monetizing AI at scale. Every serious enterprise software company has shipped an AI feature that customers actually pay for. Unlike the crypto cycle, this boom has cash flow behind it. The boom case does not require believing AI is safe. It only requires believing that the next five years of deployment look like the last two. That is a much easier bet to underwrite. ## What the doom scenario actually implies for your portfolio Here is where most coverage gets lazy. Writers either tell you to panic or tell you to buy the dip. Neither is useful. The honest answer is that tail risk changes *how* you invest, not *whether*. A few concrete implications: - **Concentration risk is real risk.** If you hold 60% of your net worth in three AI-adjacent names, you are not diversified. You are making a leveraged bet on one thesis. - **The doom scenario is not a stock-picking problem.** If the worst case happens, your index fund does not save you either. So the rational response is not to rotate into "safe" AI stocks. It is to size your exposure so that a 40% drawdown does not change your life. - **Insurance exists.** Physical gold, short-duration Treasuries, and even a small allocation to non-correlated assets are not paranoia. They are the same logic as a fire extinguisher. - **Time horizon matters more than ticker.** If you are 25 and investing monthly, a crash is a discount. If you are 60 and need the money in five years, AI stocks are a poor home for it regardless of what you think about alignment. ## Where the real money is being made right now Forget the headline names for a second. The most durable AI investment over the past 18 months has not been a chipmaker. It has been the boring layer underneath: power, cooling, and data center real estate. | Layer | Example plays | Why it holds up | |---|---|---| | Compute | Nvidia, AMD, TSMC | Direct beneficiaries, but priced for perfection | | Power and cooling | Vertiv, Eaton, Schneider Electric | Every GPU needs electricity and heat removal | | Hyperscale real estate | Equinix, Digital Realty | Long leases, sticky tenants | | Application layer | Microsoft, Salesforce, Adobe | AI features lift ARPU without new capex | | Picks and shovels for prosumers | Cursor, [Perplexity](https://www.perplexity.ai/), Claude, [ChatGPT](https://chat.openai.com/) Plus | Direct productivity gains you can measure | The bottom row is the one most people ignore. If you are a prosumer, your highest-ROI AI investment is often not a stock. It is a $20 to $40 monthly subscription that saves you five hours a week. That is a return no index fund will match. ## Our take: what we actually recommend We are not financial advisors, and this is not personalized advice. But here is how we think about it. **For the tool layer, we pay for [Claude Pro](https://claude.ai/) and ChatGPT Plus.** Both, not one. They fail in different ways, and the redundancy is worth the cost. Cursor is the best $20 a month a working developer can spend. Perplexity has replaced Google for research-heavy tasks. **For the stock layer, we treat AI as a satellite, not a core.** Core holdings stay in broad index funds. AI exposure sits at 10 to 15% of the equity sleeve, spread across at least four names and two layers of the stack. We would rather own the power and cooling trade than chase another 30% in a chip name that already trades at 40x forward earnings. **For the risk layer, we keep 6 to 12 months of expenses in cash or short Treasuries.** This is not a doom hedge. It is what lets you sleep when the headline is bad and the market is down 20%. **For the safety layer, we follow the research.** Groups like the Centre for the Governance of AI, METR, and the AI Safety Institute publish work that is worth reading even if you never invest a dollar. Understanding the risk is part of understanding the asset. ## The mistake most prosumers make They treat this as a binary. Either AI is the future and you go all in, or AI is a threat and you stay out. Both positions are intellectually lazy and financially expensive. The people who will do well over the next decade are the ones who can hold two ideas at once: this technology might seriously harm us, and it is also the best place to put incremental capital right now. The portfolio that reflects both is not complicated. It is just disciplined. Diversify across layers. Size positions so a bad year does not ruin you. Pay for the tools that make you faster. Read the safety research even when it is uncomfortable. And when someone quotes a 10% chance of catastrophe, do not dismiss it and do not panic. Just check your allocation. ## FAQ ### Is AI doom risk a reason to avoid AI stocks entirely? No. A 10% tail risk is a reason to size positions carefully, not to sit out a multi-year capital cycle. The same logic applies to any concentrated bet. ### What is the single best AI investment for a prosumer? In our view, a paid subscription to a frontier model plus one specialized tool (Cursor, Perplexity, or similar). The productivity return usually beats any single stock pick on a risk-adjusted basis. ### How much of my portfolio should be in AI? Most prosumers we talk to land between 10 and 20% of their equity allocation. Anything above 30% starts to look like a thesis bet rather than a portfolio.

Frequently asked questions

The doom case The AI safety crowd has gotten sharper over the past two years. Their argument is not that a chatbot will become sentient and hunt you down. It is that we are building systems whose cap

No. A 10% tail risk is a reason to size positions carefully, not to sit out a multi-year capital cycle. The same logic applies to any concentrated bet.

What is the single best AI investment for a prosumer?

In our view, a paid subscription to a frontier model plus one specialized tool (Cursor, Perplexity, or similar). The productivity return usually beats any single stock pick on a risk-adjusted basis.

How much of my portfolio should be in AI?

Most prosumers we talk to land between 10 and 20% of their equity allocation. Anything above 30% starts to look like a thesis bet rather than a portfolio.

Related articles