Published 30 Jul 2026 · Refreshed 31 Jul 2026

Global Macro — Trends Register

The layer above companies · global structural trends, authored once, that feed every name’s destination analysis

The shared macro layer behind VertexUno’s company research (rulings #33 tailwind-accrual & #34 macro layer). Context, not advice: these trends raise or lower a company’s scenario weights — they are never a recommendation and never a company fact by themselves. Trends are tiered (analyst inference / third-party / fact), dated 31 July 2026, and reviewed monthly. Five trends, all status: seed.

Why this exists

One register, not fifty re-derivations

Instead of every company note re-deriving the same global forces, VertexUno keeps one shared register of the structural trends that shape company destinations. Each company’s analysis cites the relevant trends and states the company-specific transmission — and, per ruling #33, whether the company’s moats capture the tailwind or compete it away.

A trend earns a place here only if it plausibly reshapes two or more portfolio names. Every trend is dated, tiered-sourced, and argues its own reversal (symmetric by design). The flow is two-way: trends flow down into company work, and observations picked up while researching a company flow back up as evidence — so the register is emergent from the portfolio’s own ground truth, not a top-down opinion sheet.

Tailwind capture = a named moat Size of the prize ≠ return on it Every tailwind argues its reversal

1 · AI compute & power build-outstructural · 3–15y

2 · Deglobalisation, reshoring & supply-chain securitystructural · 5–20y

3 · Debt, rates & the fiscal-dominance regimestructural · 3–15y

4 · Energy security → electrification & storagestructural · 5–20y

5 · Global aging & workforce declinestructural · 10–30y

Trend 1 · structural · 3–15 years

AI compute & power build-out

Touches NVDA · Micron · Alibaba · Tencent · Meta · Keppel DC REIT · CapitaLand · BYD/CATL (storage). Seed · updated 31 July 2026.

AI demand is running ahead of the compute and power that serve it. The chain has four sequential bottlenecks — chips → memory → datacenter shells → power/grid — and the binding constraint is migrating down that chain toward power. The value question is not whether AI demand is real; it is who owns the scarce link.

Ground truth (Alibaba, primary — Q4 FY2026 call & 20-F). Compute is fully utilised (“not a single card… idle”); the cost to deploy a new AI server is roughly double a year ago; the three-year cloud/AI capex plan is now “likely to overshoot” against a stated need for ~10× the 2022 datacenter footprint by 2033; cloud growth accelerated to +40%. One company so far — a strong, primary-backed lead, not yet confirmed at a second name.

What it brings about. Scarcity rents accrue to the owner of the current bottleneck (leading-edge logic, then high-bandwidth memory, then firm power + datacenter capacity). Datacenter electricity demand pulls a power & grid super-cycle and turns datacenter real estate into a scarce, contracted asset. Commodity layers — generic cloud resale, merchant capacity in a glut — see the gain competed away. Capital then floods the visible bottleneck, seeding an over-build.

Accrual (ruling #33). The improvement is enormous but accrues only to owners of the scarce link — and that link moves down the chain over time. Memory and merchant capacity are the traps: a real tailwind, competed away at the next down-cycle. Keppel DC REIT / CapitaLand / Keppel are the cleanest way to own the physical bottleneck without taking silicon-cycle risk. Name the current bottleneck moat before crediting AI to any company’s value.

Reversal. An efficiency step-change (better models or silicon-per-watt), an AI-capex digestion pause, or a power/permitting cap. The capital-cycle tell: peak lead times = peak returns, then the glut.

Trend 2 · structural · 5–20 years

Deglobalisation, reshoring & supply-chain security

Touches the five Japanese trading houses (Marubeni, Mitsubishi, Mitsui, Sumitomo, Itochu) · Micron · CATL · BYD · Berkshire · Alibaba. Seed · updated 31 July 2026.

The forty-year single-market assumption is giving way to a security-first organisation of trade — tariffs, export controls, local-content rules, friend-shoring and stockpiling. Supply chains duplicate and regionalise: a China-cost world and a local-content world, rather than one optimised globally. Energy and AI-compute are the two supply chains being secured hardest, so this is their parent trend.

What it brings about. Input costs rise and supply chains duplicate. Subsidised domestic build-outs raise over-capacity risk inside protected walls; arbitrage rents flow to whoever can move physical goods and de-risk across a fragmented world; a critical-minerals scramble begins. Value migrates to domestic-wall incumbents and to fragmentation-arbitrageurs, while pure-play exporters lose access to the walled-off pools.

Accrual (ruling #33). Deglobalisation is not a tailwind — it is a redistribution with a net cost. Subsidised capacity (chips, cells, solar) is the commodity bucket, competed away to the next entrant behind the same wall. The moated captures are intermediary networks with real switching costs (the trading houses) and genuine cost or scale leaders.

Ground truth (Alibaba, primary — Q4 FY2026 call). China’s tech-sovereignty stack is forming from the inside: Alibaba’s in-house T-Head GPUs “achieved production at scale”, now serving >60% of its external compute, with domestic chip-production capacity still the binding constraint under US export curbs. One company so far — a lead, not yet aggregate.

Reversal. Tariff rollback or trade détente; evidence that “reshoring” is mostly transshipment re-routing (China → ASEAN / Mexico) rather than real decoupling; a growth shock that re-privileges lowest cost.

Trend 3 · structural · 3–15 years

Debt, rates & the fiscal-dominance regime

Touches Ping An · AIA · OCBC · MarketAxess · the REITs (CICT, Keppel DC, CapitaLand, Howard Hughes, Hang Lung) · Brookfield · Berkshire / Markel float. Seed · updated 30 July 2026.

A regime change in the price and quantity of money: away from the 2010s’ zero-rates world toward higher-for-longer real rates, elevated debt, larger deficits and a stickier inflation floor, with policy increasingly constrained by the cost of servicing that debt. Reshoring, energy capex and defence spend push the floor up; aging pulls the other way. Net direction is contested — and that two-sidedness is the point.

What it brings about. Float and spread businesses re-rate up (insurers and banks earn real yield again); long-duration leveraged assets re-rate down via higher capitalisation rates and costlier refinancing (REITs, infrastructure); bond issuance and turnover rise — a credit-trading tailwind. Value migrates to owners of cheap, sticky, long-duration liabilities and to inflation-protected, low-leverage real assets.

Accrual (ruling #33). Rates re-price liabilities and assets in opposite directions, so the accrual question is balance-sheet structure, not the rate move. A float rise with poor underwriting is competed away; short-funded, high-gearing names give the higher yield straight back as interest cost.

Reversal / the other side. Core inflation breaking down (aging plus AI disinflation winning), or a recession forcing rates sharply lower — which flips REITs to a tailwind. This is a volatility map, not a directional bet.

Trend 4 · structural · 5–20 years

Energy security → electrification & storage

Touches BYD · CATL · insurers (transition-asset exposure) · oil-linked and internal-combustion makers. Seed · updated 30 July 2026.

Nations reducing dependence on imported oil and fragile supply — sharpened by Middle-East instability, great-power rivalry and a broad self-sufficiency turn — is a structural tailwind for electrification (electric vehicles), domestic battery and storage supply chains, and grid firming. Electrons you can generate at home beat barrels shipped through contested straits.

What it brings about. Higher electric-vehicle and storage adoption to displace imported oil; subsidised national cell manufacturing, and so over-capacity risk inside protected blocs; a bifurcated world market (China-cost versus local-content). The value migrates to whoever holds the durable low-cost or network position inside each bloc.

Accrual (ruling #33). The tailwind is volume-real, but in a bloc flooded with subsidised capacity the gains are competed away to consumers. BYD and CATL keep a slice only via cost-scale leadership, and only partially through the capital cycle. Security-driven demand raises the size of the prize far more reliably than the return on it.

Reversal. Durable cheap oil (demand destruction, US shale); policy rollback of electric-vehicle support; a breakthrough in a rival non-battery pathway. The security motive is more durable than the price motive.

Trend 5 · structural · 10–30 years

Global aging & workforce decline

Touches Ping An · AIA · OCBC · BYD (robotics optionality) · healthcare · autos. Seed · updated 30 July 2026.

Populations are aging and, in several major economies (China, Japan, Korea, much of Europe), the working-age cohort is shrinking. It is one of the most predictable structural forces there is — the people who will be old in 2045 are already alive — which makes it unusually useful for destination analysis.

What it brings about. A shrinking labour supply drives an automation and robotics imperative; healthcare, eldercare, pharmaceutical and insurance demand rise structurally; savings, pension and protection demand shift (a direct insurer driver); and the consumption mix tilts from goods to services. Aged economies become the earliest, largest markets for automation and eldercare — and face the fiscal and growth headwinds that shape rates.

Accrual (ruling #33). Aging is a demand tailwind; whether it is owner value depends on the moat. Robotics gains accrue to platform or cost-scale leaders, not commodity hardware; insurance benefits accrue where underwriting discipline and distribution hold. A premium-volume rise with poor pricing is competed away.

Reversal (partial). Pro-natal policy success; large-scale immigration; a productivity surprise. Near-certain in direction, less so in timing and magnitude per country.

Method

How the layer reaches company value

The macro layer supplies the trend and its second-order chain; the company layer supplies the company-specific incidence and accrual. No double-authoring: if a global fact lives here, the company page cites it rather than re-arguing it.

Table scrolls sideways on a phone.

Discipline	What it means for a company note
Cite, don’t re-derive	Each destination references the relevant trends and states this-company-via-this-mechanism.
Accrual test (#33)	For every tailwind, name the moat that captures it — or model it as consumer surplus, not owner value.
Signal in the aggregate	One company observation is a lead; several converging across different names are evidence for a trend.
Symmetric & dated	Every trend argues its own reversal and carries a date; a stale trend is a defect.

Tiering binds as it does for companies: press and expert commentary are leads, not facts. Macro is context and hypothesis — it moves a company’s scenario weights; it never becomes a company fact by itself. The five trends above are early (status: seed): the reasoning is structural and portfolio-grounded, and is being corroborated as each company is researched.