Two-line summary: We’re running the economy more on credit than on income and production (that buys time, not value). More intermediaries won’t fix low circulation, heavy debt service, and weak productivity; better rails and real production will.

I know, by this title you might think I’ve turned into some kind of communist or fallen for socialist rhetoric. That’s not it. I don’t like labels. I prefer to see myself as neutral, someone who looks at things as they are (beyond ideology or convenience).

The Paycheck World (and why “more finance” is not progress)

We’re living in a paycheck-driven world. Public debt keeps climbing and is projected around 95% of global GDP in 2025 and near 100% by the end of the decade (Rising Global Debt Requires Countries to Put Their Fiscal House in Order (IMF)).

US M2 Velocity (FRED) Source: M2 Velocity (FRED)

If “circulation” sounds abstract, translate it into money velocity. In the U.S., M2 velocity hovered near 1.39 in Q2 2025 (well below late‑1990s levels), which signals more money stock than rotation (M2 Velocity (FRED)). When velocity stays low, policymakers try to keep activity alive by other means (credit channels, subsidies, tax tweaks).

We actually have less real money moving than before. Since the end of the gold standard, money turned into a confidence instrument rather than a mirror of productivity. Central banks learned to simulate stability with policy while growth detached from creation. The old saying that “every company eventually becomes a bank” became a business plan in countries where credit is the oxygen of daily life (Brazil, Argentina). When every company wants the spread of a bank, the economy looks liquid until it isn’t.

When Everyone Becomes a Bank

Money was made to move, not to pile up behind layers of intermediaries. A healthy economy is about circulation. Cross‑border bank credit hit a record USD 34.7 trillion in Q1 2025, with lending to non‑bank financial institutions up ~14% YoY (International Banking Statistics (BIS)). In plain English: the system is wiring around itself to keep activity from stalling.

Reshoring, Liquidity, and the New Geography of Risk

The United States is quietly rebuilding domestic production. Manufacturing construction spending sits near record territory (about USD 223 billion SAAR in Jul 2025) (Total Construction Spending: Manufacturing (FRED)). Trade patterns confirm the shift: Mexico overtook China as the top U.S. goods supplier in 2023 (Mexico overtakes China as top U.S. supplier (AP News)).

Reshoring tightens global liquidity (fewer goods made overseas, higher costs at home, capital re‑routed into plants and equipment). When circulation slows, speculative finance starts to show its cracks.

The Central Bank Playbook (keep money moving, whatever it takes)

The playbook is simple. Encourage new financial rails and “easy fast loans” to force circulation. Seed thousands of intermediaries so activity does not visibly stall. That is why credit to non‑banks is sprinting even with mixed growth (International Banking Statistics (BIS)).

And you can see how that works in Brazil.

Brazil, the Mirror (progress and fragility in the same chart)

PIX changed how money moves in Brazil (efficient, instant, public infra). By June 2025, there were about 159.9 million people registered as PIX users and roughly 174 million total users including businesses (Pix Statistics (Banco Central do Brasil); Pix adoption milestones (Agência Brasil)).

Brazil DSR (BIS) Source: Debt Service Ratio (Brazil) (BIS)

But inclusion is not solvency. The share of Brazilian families with overdue bills hit record territory in Aug–Sep 2025 (~30.4%–30.5%) (Household indebtedness at record levels (CNN Brasil)). Look at the debt‑service ratio (DSR) to see pressure on income: for Brazil’s private non‑financial sector it rose into ~27% in early 2025 (Debt Service Ratio (BIS)). This is consistent with a consumption model leaning on credit rather than wages.

The “anesthesia” shows up in retail credit. Revolving credit‑card interest reached ~451.5% a year in Sep 2025 (Revolving Credit Card Interest at 451.5% (Agência Brasil)). Public finances carry weight too: in Feb 2025, gross public debt stood around 76.2% of GDP (and ~88.7% under IMF methodology) (Brazil Gross Debt Hits 76.2% of GDP (Reuters)). Meanwhile, DREX advances as programmable settlement infrastructure in a wholesale network using tokenized deposits (traceability at the settlement layer, not a “new bank”) (Real Digital / DREX Overview (Banco Central do Brasil)).

AI as the Last Big Bet (hope as policy)

The AI buildout looks like industrial policy by proxy. Big Tech capex surged (Microsoft around USD 30B in a single quarter; Alphabet to ~USD 85B in 2025; Meta at USD 60–72B; Amazon with multi‑billion state projects) (Microsoft to Spend Record $30B This Quarter (Reuters), Alphabet boosts 2025 capex to ~$85B (Reuters), Meta capex guidance 2025 (Reuters), Amazon to invest $10B in NC (Reuters)).

Selected AI‑Driven Capex Sources: Reuters company guidance/articles; mixed basis noted in text.

But capex meets physics. Global electricity demand from data centres could more than double by 2030 to ~945 TWh, with AI‑specific loads set to quadruple (AI to Drive Surging Electricity Demand (IEA)). In Brazil, data‑center expansion also faces grid connection lead times and substation buildouts (if the network does not scale, productivity from AI capex arrives slower).

Gold, Sanity and the Signal Inside the Noise

While the narrative chases the future, central banks bought 1,000+ tonnes of gold per year in 2022–2024, taking official reserves to modern highs (Gold Demand Trends — Central Banks 2023 (World Gold Council); Gold Demand Trends — Central Banks 2024 (World Gold Council); Gold demand up 1% in 2024 (Reuters)). I moved part of my portfolio to gold back in 2023 for that exact reason (not because I am a pessimist, but because gold still means something when confidence does not).

China’s Hybrid Game (strengths and cracks)

China plays long and quiet. It is not purely socialist or capitalist; it is a hybrid that uses market tools and state control side by side (pragmatism beats ideology). It also lives with cracks: local‑government financing strains and property‑sector stress have prompted refinancing packages since 2024 (China tackles hidden local government debt (Reuters); broader context in IMF Blog). The point is that coordination gives time to absorb shocks.

Anticorpos (what the other side gets right)

Fintechs deliver real efficiency and inclusion (PIX is proof). The problem is not the tool; it is when the tool replaces productivity and pretends to be value creation.

Why More Fintechs Won’t Fix This (and What Would)

If the question is whether adding more financial intermediaries solves a slowdown driven by low circulation, heavy debt service, and weak real productivity, the answer is no (the data above already told us why: low money velocity, record public debt, credit growth concentrated in non‑banks, and household pressure shown in DSR/NPL and revolving credit rates). More apps and rails can make payments faster, but they do not change what is moving (mostly debt) nor why it needs to move (to service itself).

What actually helps (and how to measure):

  • Shift capital from intermediation to production (energy, compute, logistics, housing). Track TFP and unit‑cost declines, not just capex headlines (if TFP does not rise, you are dressing debt as growth).
  • Guardrails on expensive short‑term credit (rotativo, BNPL, fee stacking) so “circulation” is not just compounding interest. Watch DSR and NPL trending down while credit still grows.
  • Design CBDC/DREX as public infrastructure (settlement, programmability, fee competition), not as a hidden tax on liquidity. Monitor spreads, settlement latency, and small‑merchant costs.
  • Re‑industrialize with bottleneck maps (power, grid, fabs, skilled labor). If reshoring is real, see it in manufacturing output and exports, not just construction spending.
  • Make inclusion ≠ leverage (PIX is inclusion; solvency needs income). Pair financial access with income productivity so the share of families in arrears falls and velocity rises.

Scenarios (base / bull / bear)

  • Base: velocity inches up, DSR stabilizes, AI delivers limited gains (growth drags).
  • Bull: AI lifts TFP, energy holds, DSR falls while credit still expands (soft landing).
  • Bear: energy constrains data centres, DSR rises, NPL widens (repricing of risk and tighter liquidity).

What Would Change My Mind (clear thresholds)

  • Velocity: M2V +0.10 per quarter for 4 quarters (with real wages rising).
  • Debt service: Household DSR −2 p.p. in 12 months while total credit +5% YoY.
  • Productivity: TFP +1 p.p. (12–18 months) after AI capex, adjusted for energy constraints.

TL;DR

The system is running more on credit than on income and production (that buys time, not value). If AI delivers productivity before debt service and energy constraints bite, the landing improves. If not, adding more financial intermediaries just delays the adjustment.

Final Note (read me)

This is not an economics report or investment recommendation. It is a set of observations from an enthusiast and personal investor (based on public data and what I’ve lived in the market). Do your own research, protect your family, and build things that last.

References