The tokenized real-world asset market is bigger than most people realize and worse-distributed than most coverage admits. As of May 27, 2026, RealWorldTokenSpace fully rates 49 tokenized assets covering $101.85 billion of tracked TVL, and the distribution surfaces three uncomfortable facts that "RWA market size hits $X billion" headlines miss.
This piece publishes the entire corpus, the distribution by tier and Trust Score, and the downloadable dataset so researchers, journalists, and operators can do their own slicing. It is the second-most cited piece of RWTS data after the methodology page itself.
The headline numbers
- 49 fully-rated assets with complete Trust Score breakdowns
- $101.85 billion total tracked TVL across the corpus
- 2 assets score 90+/100 (KAU and KAG, both from Kinesis — same issuer)
- 11 assets score below 60/100 — the cautionary band
- Median Trust Score: 71/100; mean 70.4
- Tier 3 (Secured DeFi) holds 73.8% of all TVL despite being the second-lowest-trust-score tier on average
Distribution by tier
| Tier | Definition | Asset count | Total TVL | % of TVL |
|------|------------|-------------|-----------|----------|
| Tier 1 | Physically-Backed RWA (allocated gold, silver) | 5 | $5.40B | 5.3% |
| Tier 2 | Treasury & Fiat-Backed (BUIDL, USYC, USDM, OUSG) | 10 | $16.48B | 16.2% |
| Tier 3 | Secured DeFi (wstETH, sDAI, vaults, lending) | 25 | $75.18B | 73.8% |
| Tier 4 | Synthetic & Structured (sUSDe, Pendle PTs, leveraged) | 9 | $4.79B | 4.7% |
The headline number — $101.85B — masks a sharply non-uniform distribution. Tier 3 alone is more than three-quarters of the entire on-chain RWA market by TVL, which means most retail and institutional capital is sitting in lending vaults and liquid staking tokens, not in the tokenized-treasury and tokenized-gold products the financial press writes about most. Tier 1 (the "safest" allocated-physical category) is a rounding error in TVL terms even though it scores the highest Trust Scores.
The Tier 4 share — under 5% — is informative for a different reason. Synthetic and structured yield (sUSDe, Pendle PTs, Ember vaults) is where the highest APYs live (median 9.2%, max 14.5% in our corpus), and the small TVL share suggests institutional capital is voting with its feet against the risk-adjusted return. Retail capital is heavier here as a fraction of holdings, but it's not where the institutional money sits.
Distribution by Trust Score
| Score band | Description | Count | % of corpus |
|------------|-------------|-------|-------------|
| 90-100 | Exceptional | 2 | 4.1% |
| 80-89 | Strong | 10 | 20.4% |
| 70-79 | Solid | 16 | 32.7% |
| 60-69 | Acceptable | 10 | 20.4% |
| < 60 | Cautionary | 11 | 22.4% |
Less than a quarter of the corpus scores in the "Strong" or "Exceptional" bands. Almost a quarter sits in the "Cautionary" band (under 60/100), and the median asset is a 71/100 — solid but not without questions on at least one of the six methodology dimensions.
The shape of the distribution is closer to bimodal than bell-curved. There's a cluster of strong-to-exceptional assets at the top (the well-documented physical metals, the institutional treasury tokens, the regulated stablecoin earn products), a fat middle of solid-but-imperfect DeFi positions, and a long tail of cautionary scores — almost all of which are Tier 4 synthetic strategies whose backing is harder to verify or whose track record is too short to rate higher.
The top 10 by Trust Score
| Rank | Ticker | Asset | Tier | Score | TVL |
|------|--------|-------|------|-------|-----|
| 1 | KAU | Kinesis Gold | T1 | 97/100 | $349M |
| 2 | KAG | Kinesis Silver | T1 | 97/100 | $283M |
| 3 | PAXG | PAX Gold | T1 | 89/100 | $2.18B |
| 4 | USDC | Coinbase USDC Rewards | T2 | 88/100 | $5.00B |
| 5 | BUIDL | BlackRock USD Institutional Digital Liquidity | T2 | 87/100 | $2.86B |
| 6 | cbETH | Coinbase ETH Staking | T3 | 85/100 | $6.36B |
| 7 | USYC | Hashnote US Yield Coin | T2 | 84/100 | $2.99B |
| 8 | USDM | Mountain Protocol USD | T2 | 84/100 | $348M |
| 9 | XAUT | Tether Gold | T1 | 83/100 | $2.54B |
| 10 | XAUm | Matrixdock XAUm | T1 | 83/100 | $50M |
Two patterns stand out. First, Kinesis dominates the top by virtue of running both ends of its own infrastructure — they custody the metal AND operate the chain, which removes most third-party-attestation friction. The trade-off (no third-party validator) is real but the dimension scores reflect it. Second, the institutional issuers (BlackRock, Coinbase, Franklin, Hashnote, Mountain) cluster in the 84-88 band — high but not maxed, because the 90+ threshold requires both perfect operational documentation AND a multi-year incident-free track record, and most institutional tokenized products are too new for the latter.
The bottom 5
| Ticker | Asset | Tier | Score | TVL |
|--------|-------|------|-------|-----|
| PT-USDe | Pendle PT-USDe | T4 | 48/100 | $1.10B |
| rcUSD | Ember rcUSD Vault | T4 | 49/100 | $15M |
| SolvBTC | Solv BTC | T4 | 50/100 | $498M |
| sFRAX | Staked FRAX | T4 | 50/100 | $65M |
| KVT | Kinesis Velocity Token | T4 | 55/100 | $130M |
All Tier 4. All synthetic or structured. The single largest position in the bottom 5 by TVL is Pendle PT-USDe at $1.1 billion — meaning a non-trivial amount of capital is sitting in a position that scored below the corpus median, presumably because the yield premium (~9%+ APY) justifies it for those holders. That's not a problem in itself; it's a problem only when investors don't know where their position sits in the distribution.
APY range by tier
| Tier | Asset count | Min APY | Median APY | Max APY |
|------|-------------|---------|------------|---------|
| T1 | 5 | 0.00% | 0.00% | 0.45% |
| T2 | 10 | 3.02% | 3.55% | 7.50% |
| T3 | 25 | 0.36% | 4.07% | 9.50% |
| T4 | 9 | 0.60% | 9.20% | 14.50% |
The risk-return curve is exactly what theory predicts: yield rises monotonically with tier, with each step adding roughly 1.5-3 percentage points of median APY in exchange for additional complexity, custody risk, or structural risk. The interesting tells:
- T1 max is 0.45% (Kinesis velocity yield). The other four T1 assets pay nothing.
- T2 median (3.55%) closely tracks 3-month US T-bill yields — these are tokenized treasuries doing what they say on the tin.
- T3 has the widest range (0.36% to 9.50%) because the tier groups very different products: rebasing LSTs at 2-3%, leveraged DeFi vaults at 6-9%, USDC lending at 4%.
- T4 has the smallest sample size (9 assets) but the highest median (9.20%), reflecting that synthetic yield IS the value proposition of this tier — without the yield premium, no one would accept the additional structure risk.
Chain distribution
| Chain | Asset count | % of corpus |
|-------|-------------|-------------|
| Ethereum | 34 | 69% |
| Multi-chain | 17 | 35%* |
| Solana | 6 | 12% |
| Stellar | 3 | 6% |
| Sui | 2 | 4% |
| Arbitrum | 1 | 2% |
*Multi-chain assets count in multiple buckets, so percentages do not sum to 100.
Ethereum hosts 69% of the corpus by asset count, consistent with where institutional issuance still lands first. Solana's presence (12% of assets) is concentrated in DeFi yield products (jitoSOL, mSOL, Kamino vaults) rather than tokenized RWAs. Stellar appears purely because of Kinesis (KAU, KAG, KVT), which is an interesting concentration: the three highest-velocity-yield instruments on the chain are all the same issuer's products.
Download the full corpus
The full machine-readable dataset is published at /data/rwts-corpus-2026-05-27.csv. Columns: id, ticker, name, category, chain, tier, trust_score, apy_percent, tvl_usd. 49 rows.
Free to use for research, journalism, and product development with attribution to RealWorldTokenSpace and a link back to this article or the methodology page. If you want programmatic access with live-updating data instead of a frozen snapshot, the Trust Score API covers the same corpus.
Notes on methodology
Trust Scores are reviewed quarterly and on any material event (incident, attestation gap, regulatory action, peg break). The May 2026 corpus snapshot reflects the most recent quarterly cycle plus any event-driven re-scores between then and publication. Each asset's score breakdown — including the dimensional weights and rating rationale — is on its respective /directory/{id} page.
Inclusion is constrained by what we can independently verify. Assets without public backing documentation, asset-side audits, or on-chain proof-of-reserves are excluded outright regardless of TVL — opacity is disqualifying in our methodology. That's why some products you see in marketing decks aren't in the corpus: we couldn't verify enough to rate them honestly.
What the data does not show
Three things the corpus snapshot deliberately does not surface, because they're outside what a static distribution can answer:
- Trust Score volatility over time. Some assets have stable scores across quarters; others move materially when issuer disclosures change. The snapshot is a point-in-time view. The methodology page documents the re-scoring triggers.
- Correlations between assets. Two T2 stablecoins might look identical at the asset level but be highly correlated through shared treasury counterparties. The correlation matrix tool covers this dimension explicitly.
- Issuer concentration risk. Five of the top ten Trust Scores are from three issuers (Kinesis × 2, Coinbase × 2, Paxos). The corpus is more concentrated by issuer than by asset count alone suggests. This is a follow-up piece — flag in the comments or email partnerships if you want it next.
If you cite this dataset, please link back so readers can see the live methodology and the latest re-scores rather than a frozen snapshot.
