Every quarter, the Bureau of Economic Analysis publishes a single number that the financial media treats like a final exam score for the American economy. Cable news runs it in bold. Strategists update their slide decks. And most of it is useless to anyone actually deploying capital. At Nestpoint, we find the real value in GDP data isn't the headline — it's what happens when you crack it open and look at the individual line items doing the heavy lifting. This quarter, we did exactly that, and what we found is a 2% growth print that is less a picture of broad-based economic strength than a story about three very specific drivers: healthcare consumption, federal deficit spending, and AI infrastructure capex carrying an entire economy on their backs. For allocators trying to understand where the cycle actually stands, the composition of that 2% matters far more than the number itself.
The Consumer
Consumer spending added 1.08 percentage points to the headline. Goods contributed essentially nothing. Durables were flat at 0. Non-durables nudged into negative territory at 0.03. Services did all the work, contributing 1.11.
Inside services, household consumption accounted for 0.93. Of that, healthcare delivered 0.51 and financial services and insurance added another 0.26. Roughly three quarters of household consumption growth came from medical care and the cost of insuring things.
Healthcare alone accounted for about 25.5% of total GDP growth in the quarter. This continues a pattern that has been building for years. Strip healthcare out of recent payroll reports and the labor market looks notably less robust. The driver here is straightforward and largely non-cyclical. The population is aging, and an aging population consumes more medical services regardless of how confident anyone feels about their job. That consumption shows up in the data as growth, which is accurate in the same way that paying more for the same insurance policy is technically a "purchase." It is real economic activity. Whether it tells you anything about the underlying vitality of consumer demand is another question.
The same logic applies to insurance itself. Premiums on homes, autos, and health policies have climbed sharply, and those higher prices flow through the national accounts as larger consumer expenditures. Most households are paying more for the same coverage they had a year ago, and that extra payment counts as growth.
Investment
Investment was the largest contributor to the quarter, adding 1.48 percentage points. Inventories added 0.04. Fixed investment did the heavy lifting at 1.08.
The internals tell a more pointed story. Residential investment subtracted 0.31. Higher rates continue to do exactly what higher rates are designed to do, which is suppress interest-sensitive activity. Non-residential investment added 1.39, and within that, structures actually fell by 0.19. The remainder showed up in equipment and intellectual property, dominated by IT.
This is the AI capital expenditure cycle, and it comes with a useful nuance that most macro commentary misses. The binding constraint right now is electricity. Building a new data center requires a power purchase agreement, and those agreements can take roughly five years to arrange. To move faster, hyperscalers and cloud providers have been buying or repurposing existing facilities, including some originally built for cryptocurrency mining, where the power infrastructure is already in place. That explains why so much of the spending is landing in compute hardware and intellectual property rather than in new construction. It also explains why the energy infrastructure supporting this buildout has become an investment thesis in its own right. The data center does not work without the power, and the power does not appear on command. Every facility conversion and every new site requires generation capacity, grid interconnection, and in many cases on-site energy assets. For investors who understand the physical layer beneath the compute layer, that dependency chain is where some of the most durable opportunities in this cycle sit.
A meaningful portion of investment growth is tied to one capital expenditure cycle, executed by a small number of very large buyers, and gated by a physical input that does not scale on a quarterly timeline. AI capex is real and material, and the electricity constraint means it will show up in the data unevenly, with quarters of outsized contribution followed by quarters where permitting and power agreements create natural pauses. That rhythm should not be mistaken for the end of the cycle. The underlying demand for compute infrastructure continues to outstrip supply, and the companies making these investments are not building for a single product cycle. As the buildout matures, the opportunity set will evolve from the initial land-grab into optimization, efficiency, and the next layer of physical infrastructure, including robotics and automation systems that depend on the same compute and energy foundations being laid right now. For allocators, the concentration means the GDP line item will be noisy, but the secular trajectory of the buildout remains intact.
Government
Government spending added 0.73, with federal contributing 0.56 and state and local 0.17. Of the 0.73, roughly 0.43 reflects current operating expenses rather than investment in productive capacity. The portion attributable to growth-oriented investment was 0.04.
In aggregate, government spending accounted for 36.5% of GDP growth in the quarter. Combined with healthcare, two categories produced 72% of the growth. Add the AI capex cycle, and three categories effectively produced the entire 2% print.
There is a longer pattern here that deserves to be named. If you adjust GDP since 2019 to reflect a balanced federal budget, you find that the federal deficit has financed a substantial share of reported growth for six consecutive years. Deficit spending shows up as growth because deficit spending is, by accounting definition, growth. Whether deficit-funded growth deserves the same valuation multiple as organic, productivity-driven growth is a question every investor in long-duration assets should be asking, and one that the equity market has so far been content to leave unanswered.
Trade
Net exports subtracted 1.3 percentage points. Exports added 1.32, while imports subtracted 2.62. The composition leaned toward goods on both sides, with services close to balanced. The growth-rate differential between imports and exports produced the drag, and it was a meaningful one for the headline.
What the Composition Actually Says
Stack the pieces together and the 2% print rests on a narrow base. Healthcare spending tied to demographics. Federal spending tied to a deficit that is unlikely to shrink anytime soon. A capital expenditure cycle in AI infrastructure that is concentrated in a handful of large buyers and constrained by the availability of electricity. Outside those three, contributions are modest at best, and several categories are outright negative.
Two implications stand out for allocators.
First, the gap between the headline and the felt experience of the economy has a straightforward explanation in the accounting. When growth is concentrated in healthcare consumption, federal expenditures, and AI infrastructure, most households and most businesses do not directly participate in any of it. Their experience of "the economy" is shaped by the categories that aren't carrying the headline, and those categories are flat or contracting. The headline says growth. The lived experience says stagnation. Both can be correct at the same time, which is part of why this market environment feels so disorienting to so many people.
Second, the composition raises a structural question about cyclical analysis itself. An aging population will keep spending on medical care through any cycle. A government running persistent deficits will keep adding to nominal output. AI infrastructure spending, unlike the other two, is building productive capacity that has the potential to reshape the cost structure of entire industries over the next decade, and it is pulling energy and industrial investment forward with it. Together, these factors put a floor under reported GDP that may be higher than the underlying broad-based activity would otherwise produce. Recessions defined by two consecutive quarters of negative GDP may become harder to generate, which sounds reassuring until you consider that two of the three pillars are demographic inertia and deficit finance rather than productive expansion. The third pillar, the infrastructure buildout, is the one actually creating new capacity. For investors, the critical question is not whether GDP stays positive but whether the productive investment embedded in the number is large enough, and durable enough, to eventually pull the rest of the economy forward.
The headline reads as growth. The components read as concentration. For allocators, the concentration itself is the map. It tells you where the capital is flowing, where the physical constraints are binding, and where the next phase of opportunity is likely to emerge. Reading the composition is not just an exercise in skepticism about the headline. It is the starting point for figuring out where to deploy.
Analysis informed in part by recent commentary from Mark Meldrum.
