The 2026 ETF Themes Are a Macro Bet Disguised as a Product Menu

Bloomberg and J.P. Morgan released their thematic ETF predictions for 2026. AI, infrastructure, and defense. Three categories. One shared assumption. That capital-intensive industries will absorb liquidity without a hitch. That is a structural bet on the global macro environment, not a simple sector preference. The firms present these as product menus. They are actually macro forecasts wearing marketing suits. Institutional asset managers do not predict themes in a vacuum. Their picks reflect an internal view on the rate cycle. AI data centers, power grids, and defense contractors require multi-year, low-cost capital. When two major institutions signal sustained appetite in these sectors, they are implicitly stating that the rate environment will not break these asset classes. Yet neither Bloomberg nor J.P. Morgan explicitly said anything about rates. That omission is not neutral. It is the most important detail of the entire report. These three themes map to one economic logic: an investment-driven growth cycle. AI falls under the umbrella of capital expenditure. Infrastructure is textbook investment. Defense is sovereign spending, but still investment. This is not a consumption story. It is not an export story. It is a bet that total factor productivity gains from AI will justify the capex expansion, and that the state will keep the infrastructure engine running. That is a coherent view. It is also a fragile one. Let me break down the components. AI is the most obvious theme. The AI sector is in a capex super-cycle. Microsoft, Google, and Meta are building out data centers at an unprecedented scale. But the scale of that build-out is the issue. The electricity demand is enormous. Data centers consume massive amounts of power, which creates pressure on the electrical grid. The chip supply chain is strained. This is a supply-chain-heavy thesis. When the market prices AI ETFs, it is pricing the full stack: the chips, the cooling systems, the power infrastructure. The bottleneck is not the model innovation. The bottleneck is the physical world. Land, power, and cooling are the real constraints. Investors are treating AI like a software story, but it is a hardware story. Infrastructure is the second pillar. This is tied to the global fiscal expansion cycle. The US passed the infrastructure bill. Europe has its recovery fund. China is deploying special bonds. These programs are driving concrete demand for industrial metals, equipment, and engineering services. But there is a latent flaw in this thesis. The funding source is government debt. When you see infrastructure ETFs, you are seeing a bet on government balance sheet expansion. If the fiscal environment tightens, this entire sector faces a funding squeeze. Defense is the third pillar. The geopolitical tension is the driving force. NATO nations are pushing toward 2% of GDP on defense. This is a response to a real geopolitical environment. The defense sector has a strong earnings visibility, because the contracts are long-term and government-backed. But the defense thesis is binary. If geopolitical tensions de-escalate, the defense budget boom stalls. This is a tail-risk against the theme. The real issue is what these three themes share. They are all sensitive to interest rates. Capital-intensive industries trade like long-duration bonds. When rates rise, their valuations compress. If inflation rebounds, the rate path tightens, and the whole thesis breaks. The current market is pricing a certain rate path. If that path shifts, the entire strategy needs to be rewritten. I have spent years auditing the accounting of crypto and tech projects. The same discipline applies here. When I look at this ETF theme list, I see a latent risk that is not being priced. It is a supply-side issue. The AI buildout is consuming the same resources as the infrastructure buildout. Copper is the clearest example. Data centers need copper for power and cooling. Infrastructure needs copper for grids. Defense needs copper for ammunition. Three sectors, one commodity, finite supply. The report treats these as distinct sectors, but they are competing for the same inputs. There is also the issue of the power supply. The AI infrastructure buildout is a massive demand shock to the electricity grid. The utility sector needs to invest to keep up. This is a circular problem. The more AI infrastructure we build, the more we need to upgrade the grid, which requires more funding, which competes with the defense budget. The report does not mention this interconnection. That is the structural weakness. What did the bulls get right? The AI revolution is real. The productivity gains are visible. The tech giants are not slowing down. The infrastructure cycle is in motion. The fiscal spending is real. The defense spending is a bipartisan issue in the US. These are solid, sustained themes. They are not a fad. But the key insight is that the market is pricing these themes as if the funding will be endless. That is a fragile assumption. The funding cost is tied to the rate. The rate is tied to inflation. And inflation is tied to the very resources that these themes are consuming. It is a feedback loop. The bulls have not priced the feedback. For the retail investor, the passive inflow into these ETFs is a double-edged sword. The passive buying creates a self-fulfilling prophecy in the short term. But it also creates a risk of systemic overlap. The market is creating a concentrated bet on a few specific macro outcomes. Here is the hard truth: this is not an investment thesis. It is a political platform. These are the sectors that the government is subsidizing and directing capital toward. The ETF is a tool for distributing the fiscal policy. It is not a vote of confidence in the market. It is a vote of confidence in the state. The institutions are capturing the policy signal and packaging it as a product. The implication is that the risk of a sudden policy shift is high. If the government decides to tighten the fiscal budget, the infrastructure theme collapses. If the AI regulation gets strict, the AI theme suffers. If the geopolitical environment shifts, the defense theme reverses. This is not a bottom-up strategy. It is a top-down bet. The market is treating these as secular themes. That is a mistake. These are cyclical themes. They are tied to the policy cycle, the rate cycle, and the geopolitical cycle. The only question is the cycle. The data will reveal the answer. My recommendation is not to avoid these themes. The data is not there to support that. The recommendation is to understand the risk. The risk is not in the asset. The risk is in the macro assumption. The risk is in the rates, the inflation, and the geopolitical. The risk is the feedback loop. The risk is the demand for the same resources. The risk is the same as the crypto leverage: a high yield is a warning, not a welcome. When I look at the historical pattern, I see this. The market tends to over-predict the scale of the project. The history is full of these. The past is prologue. The AI theme is the new internet. The internet was a real thing, but the internet bust was a real thing. The infrastructure is the same as the railroads. The defense is the same as the war economy. The key is the timing. The 2026 is the timing question. The capex is running ahead of the cash flow. The capex is running ahead of the productivity. The thesis is running ahead of the reality. The data will show the gap. A real signal to watch is the corporate earnings. The tech giants will report the capex numbers. If the capex numbers are high and the earnings are not, the AI theme is in trouble. The infrastructure is tied to the government spending. If the government spending is not met, the thesis breaks. The defense is tied to the geopolitical events. If the geopolitical events de-escalate, the defense theme reverses. I need to look at the data. The data is not the prediction. The data is the evidence. The evidence is the only thing that matters. The code does not lie. People do. The institutions are the people. The numbers are the code. This is the final takeaway. The ETF is a product. The product is a promise. The promise is a bet. The bet is a risk. The risk is the macro. The macro is the data. The data is the signal. The signal is the truth. The truth is the risk. The risk is the warning. The warning is the yield. The yield is a warning. Do not be the one who is left with the bag. Audit the promise, not the poster.