The State of RWAs — Where the Research Disagrees

Everyone agrees RWAs are coming. Nobody agrees how big.

A single-page editorial synthesis that cross-reads 61 research reports on real-world assets and tokenization — global banks, asset managers, the Big Four, crypto-native desks and the official sector — all published between 29 Mar and 27 Jun 2026. The finding: the desks converge almost completely on the direction of travel and contradict each other on nearly everything you can put a number on.

The headline tension: $400B → $14T for the same 2030 horizon. A 35× spread. And an even wider ~36× spread on what the market is worth today.

No build step, no dependencies, no framework. One file — index.html.

---

The argument in three lines

1. The facts of today are settled. ~0.01% penetration of a $300T+ addressable base, ~$15B of tokenized Treasuries anchoring the market, institutional issuers now leading growth. Nobody disputes this.
2. Every forward number is contested. 35× on 2030 size, 36× on today's size, and a straight split on whether stablecoins are the settlement rail (crypto desks) or a flawed form of money to be superseded (BIS).
3. Most of the disagreement is definitional, not factual. Three choices explain the bulk of it — what counts as an RWA (free-float vs. represented vs. full settlement layer), which chains you count (public-only undercounts ~10×), and whether stablecoins are in.

---

What's on the page

#	Section	What it does
01	Hero	The $400B → $14T tension, the source base by category
02	Consensus	The five numbers every desk signs up to
03	The matrix	18 desks × 6 questions on one grid
04	Discrepancy 01 — the 2030 number	Interactive forecast chart + study table
05	Discrepancy 02 — the denominator	Why "today's" market size ranges $19.3B → $695B
06	Discrepancy 03 — stablecoins	Rail, coexistence layer, or flawed money
07	Discrepancy 04 — the lead asset	Treasuries lead now; after that, pick a fight
08	Discrepancy 05 — whose chain	Public trackers vs. the permissioned layer nobody sees
09	Discrepancy 06 — the constraint	Six desks, six different bottlenecks
10	Takeaway	How to read any RWA report without being fooled by scope

~12 minute read.

The chart (section 04)

Two toggles — 2030 forecast / market today and linear / log — over the same six-to-seven study dataset. The encoding is deliberate and switches with the scale:

• Linear → marks are bars. Length encodes magnitude from a real zero.
• Log → marks are dots. There is no zero on a log axis, so bar length would be meaningless; only position is read.
• Colour encodes the definition used (free-float · public-chain · securities · broad · all-in incl. stablecoins), not the desk.
• Hollow dot = the same desk's second sourced figure under a wider scope (21Shares $31B → $350B; Standard Chartered $4T '28 → $2.7T DeFi-active '30).
• Band = a published scenario range (Binance $320B–$4.8T around a $1.6T base).
• pt. est. = the source publishes a point estimate with no range.
• On linear with two figures, the solid bar stops at the lower number and extends in a tint — the solid block never overstates the most conservative sourced figure.

Log tick density adapts to plot width (narrow viewports drop to one tick per decade so labels don't overprint), so the chart re-renders on resize.

The matrix (section 03)

18 desks down, six questions across, grouped by desk type. Read down a column to see whether a question is settled or split; read across a row for one desk's whole position. Empty rings are desks in the source base that took no public position on that question — and most cells are empty, which is itself the finding: 61 reports, few desks answering more than two of these six questions.

---

Source base

61 reports cross-read, 29 Mar – 27 Jun 2026 (49 archived locally — the source documents are not in this repo; the page reproduces figures as reported, with dates).

• Banks — Citi · J.P. Morgan · Morgan Stanley · Standard Chartered · Deutsche Bank · HSBC · Barclays · Nomura
• Managers & issuers — BlackRock · Franklin Templeton · Securitize · Grayscale · 21Shares · Bitwise
• Consultancies — BCG · McKinsey · PwC · Oliver Wyman · Deloitte · KPMG
• Crypto research — Binance · Pantera · Messari · Galaxy · DefiLlama · CoinGecko · BeInCrypto · RWA.xyz · Keyrock
• Official sector — BIS · IMF · GFMA · OMFIF

---

Repo layout

index.html       the entire site — markup, design tokens, CSS, chart data, chart engine
wrangler.jsonc   Cloudflare static-assets config (worker name: rwa-snapshot)
.gitignore       wrangler local state + env files
README.md        this file

Run it locally

Any static server works — there's nothing to build.

python3 -m http.server 8000

Then open http://localhost:8000. Or with Wrangler, matching the deploy target:

npx wrangler dev

Opening index.html directly off the filesystem also works; only the Google Fonts request needs network.

Deploy

The page ships as static assets on Cloudflare. wrangler.jsonc serves the repo root (assets.directory: ".") under the worker name rwa-snapshot:

npx wrangler deploy

Git-integration deploys from main work equally well — there is no build command and no output directory to configure.

Two things worth knowing before you touch that config: serving . publishes every file in the root, including this README and wrangler.jsonc; and the nodejs_compat flag is inert here, since there is no worker script to run.

---

Editing the content

Chart data lives in the inline <script> — the DS object at index.html:841, split into forecast and today arrays. One entry per study:

{n:'Citi', v:5500e9, cat:'secur', pt:1,
 scope:'Tokenized securities — equities + Treasuries led', src:'Jun 2026'}

• v — headline figure in dollars · lo/hi — published scenario band · alt/altLbl — same desk, wider scope (renders as the hollow dot)
• cat — the scope bucket, keyed to CATS at index.html:867; it drives both the colour and the legend, which is generated from whichever categories are actually present
• pt: 1 — flags a point estimate · scope and src — the hover/tap tooltip

The prose tables under each chart are hand-written HTML, not generated from DS. Adding or revising a study means editing both, or the chart and the table beneath it will disagree — which would be an unfortunate way for this particular page to fail.

Design tokens are CSS custom properties in :root at index.html:12 — paper #F2EBDD, ink #1A1816, accent #C44A36, highlight #F5E69A, plus Fraunces (display) and Inter (text) from Google Fonts. Scope colours used by the chart and the matrix pills come from the same set, so the two sections stay legible against each other.

---

Caveats

Figures are reproduced as reported by each source on the date cited. Scopes and definitions differ materially between studies and are not directly comparable — that incomparability is the subject of the brief, not a defect in it. Not investment advice.

---

Prepared by Mike Blank · June 2026. Design after trancheprotocol.com.