On-Chain Private Credit TVL Rises 7.2% to $2.47B as Stablecoin Yields Fall Below 3.9%

AltStreet's tracked on-chain private-credit TVL increased by $165.6 million from the prior print, while the large-pool stablecoin yield median fell 21.2 basis points to 3.888%. WTI eased, natural gas rose and RTX 4090 rental pricing moved higher.

· Reviewed by michael
Headline card for On-Chain Private Credit TVL Rises 7.2% to $2.47B as Stablecoin Yields Fall Below 3.9%

Key facts

25
  1. AltStreet Tracked On-Chain Private Credit TVL was $2,467,318,006 as of August 7, 2026.
  2. The prior July 29 observation was $2,301,742,197.
  3. Tracked on-chain private-credit TVL therefore increased by approximately $165.6 million, or 7.2%, between observations.
  4. Maple Finance accounted for $2,465,501,293 of the August 7 private-credit aggregate.
  5. Goldfinch accounted for $1,816,713.
  6. Maple represented more than 99.9% of AltStreet's current two-protocol private-credit TVL universe.
  7. AltStreet Tracked RWA TVL was $5,116,357,968 as of August 7.
  8. Ondo Finance accounted for $3,478,734,013 of tracked RWA TVL.
  9. Centrifuge accounted for $1,637,623,955.
  10. The AltStreet Large-Pool Stablecoin Yield Median was 3.888% across 101 qualifying pools.
  11. The prior stablecoin-yield median was 4.100%.
  12. The median declined by 21.2 basis points between weekly observations.
  13. WTI Cushing spot was $81.96 per barrel on August 3 versus $84.25 on July 27.
  14. Henry Hub natural gas spot was $2.81 per MMBtu on August 3 versus $2.63 on July 27.
  15. Lower-48 working gas in storage was 3,117 Bcf as of July 31.
  16. The H100 on-demand median remained $4.19 per GPU-hour.
  17. The A100 80GB on-demand median remained $3.23 per GPU-hour.
  18. The RTX 4090 spot median increased from $0.30 to $0.34 per GPU-hour.
  19. No seven-day Royalty Exchange median was calculated because fewer than three usable closings were recorded.
  20. The trailing 28-day Royalty Exchange sample contained 10 deals with a median LTM entry yield of approximately 16.1%.
  21. Fund FG-GWS reported cumulative sales unchanged at $4,215,007 and 24 investors in its August 5 Form D/A.
  22. Fund FG-GWS changed its reported Investment Company Act exclusion from Section 3(c)(1) in April to Section 3(c)(7) in August.
  23. AltStreet detected five entity-growth signals, six Exchange Act filings, two new Form 1-U filings, two offering-circular supplements, one offering-stalled signal and one amendment-spike signal during the trailing seven days.
  24. The Hiive50 Index stood at 165.68 on August 7 and was reported up 52.1% over twelve months.
  25. AltStreet's TVL, yield, GPU and EDGAR-flow measures represent defined tracked universes rather than whole-market measurements.

AltStreet Tracked On-Chain Private Credit TVL rose to $2.467 billion as of August 7, up from $2.302 billion on July 29 — an increase of approximately $165.6 million, or 7.2%, between weekly observations.

The move was almost entirely attributable to Maple Finance, which accounted for $2.466 billion of the $2.467 billion tracked aggregate. Goldfinch contributed approximately $1.82 million, leaving Maple with more than 99.9% of the two-protocol universe.

At the same time, the AltStreet Large-Pool Stablecoin Yield Median fell to 3.888% across 101 qualifying pools with at least $50 million in TVL, down from 4.100% the prior week. That is a decline of 21.2 basis points.

Elsewhere, WTI Cushing spot declined to $81.96 per barrel while Henry Hub natural gas rose to $2.81 per MMBtu. H100 and A100 cloud GPU medians were unchanged, but the tracked RTX 4090 spot median increased from $0.30 to $0.34 per GPU-hour.

Three things that moved this week

1. Tracked on-chain private-credit TVL increased 7.2% to $2.467 billion, but Maple Finance represented more than 99.9% of the aggregate, making the headline highly sensitive to a single protocol.

2. The large-pool stablecoin yield median fell 21.2 basis points to 3.888% even as the qualifying universe expanded to 101 pools, the largest pool count in AltStreet's recent observations.

3. Energy diverged: WTI Cushing spot fell approximately 2.7% between the July 27 and August 3 EIA observations, while Henry Hub natural gas rose approximately 6.8%.

Private-credit TVL rose $165.6M — but Maple dominates the measurement

AltStreet Tracked On-Chain Private Credit TVL reached $2,467,318,006 on August 7, compared with $2,301,742,197 on July 29.

The difference between those observations is $165,575,809, equivalent to approximately 7.2%. The figures are stock-of-assets readings rather than capital-flow data, so the change should not be described as $165.6 million of new investor inflows.

Maple Finance accounted for $2,465,501,293 of the current reading. Goldfinch accounted for $1,816,713.

That concentration is analytically important. Although the aggregate is useful for tracking the protocols currently inside AltStreet's defined universe, it should not be treated as a diversified measure of the entire on-chain private-credit market.

A movement in Maple alone can move essentially the entire index. Until additional protocols enter the tracked universe with economically meaningful TVL, protocol-level composition is more informative than the aggregate headline by itself.

Tracked RWA TVL holds above $5.1B

AltStreet Tracked RWA TVL stood at $5,116,357,968 as of August 7 across two protocols with active DefiLlama mappings.

Ondo Finance accounted for $3,478,734,013, or approximately 68% of the aggregate, while Centrifuge contributed $1,637,623,955, or approximately 32%.

Like the private-credit series, this is a deliberately scoped AltStreet measurement rather than an estimate of the entire tokenized-real-world-asset market.

The two-protocol composition means changes at either Ondo or Centrifuge can materially move the aggregate. Readers should therefore use protocol composition alongside the top-line TVL rather than treating the $5.12 billion figure as a broad-market index.

AltStreet's analysis of tokenized-equity pricing and wrapper risk provides additional context for evaluating on-chain representations of traditional assets: https://altstreet.investments/blog/tokenized-equities-valuation-marks-guide

Large-pool stablecoin yields fall 21 basis points

The AltStreet Large-Pool Stablecoin Yield Median declined to 3.888% from 4.100% the prior week, a decrease of 0.212 percentage points, or 21.2 basis points.

The latest calculation included 101 qualifying pools with at least $50 million in TVL, the highest qualifying count in AltStreet's recent observations.

Among the ten largest qualifying pools examined this week, quoted yields ranged from 3.16% for USYC on BSC through Circle to 4.87% for a Maple USDC pool on Ethereum.

The group included tokenized Treasury exposures such as BlackRock BUIDL, Invesco USTB and Ondo USDY alongside lending and structured-yield products from Maple, Sky, Ethena and Centrifuge.

The median decline is descriptive rather than causal. Changes in pool composition, utilization, protocol incentives and underlying short-term rates can all affect the observed yield distribution.

AltStreet's current yield data and cross-platform comparisons are available through the Yield Monitor: https://altstreet.investments/data/yields

WTI falls while Henry Hub gas rises

U.S. energy benchmarks moved in opposite directions in the latest official EIA observations.

WTI Cushing spot was $81.96 per barrel on August 3, down from $84.25 on July 27. The $2.29 decline represents approximately 2.7% between the two observations.

Henry Hub natural gas spot increased to $2.81 per MMBtu from $2.63, an increase of approximately 6.8%.

Lower-48 working natural gas in storage stood at 3,117 Bcf for the week ended July 31, compared with 3,084 Bcf one week earlier.

The weekly movements provide macro context for AltStreet's energy and real-asset coverage but should not be interpreted as direct performance measures for private energy investments, mineral interests or agricultural assets.

H100 pricing holds steady as RTX 4090 spot rates rise

The AltStreet GPU Rental Index showed little movement in higher-end on-demand accelerators.

The H100 on-demand median remained $4.19 per GPU-hour across nine tracked offers from Azure, CoreWeave, Lambda Labs and RunPod. Prices ranged from $2.89 to $12.29.

The A100 80GB on-demand median also held at $3.23 per GPU-hour across four tracked offers, with a range of $2.70 to $4.10.

Consumer-class GPU capacity moved differently. The RTX 4090 spot median increased to $0.34 per GPU-hour from $0.30 on July 29, an increase of approximately 13.3%, across offers from RunPod, SaladCloud and Vast.ai.

With only three qualifying RTX 4090 observations in the current print, that change should be treated as a thin-market pricing signal rather than a broad measure of GPU compute inflation.

Decentralized-compute tokens were mixed

Akash traded at $0.4883 on August 7 compared with $0.4220 on July 29, a gain of approximately 15.7% between AltStreet's two weekly observations.

Bittensor traded at $194.12 compared with $188.12, an increase of approximately 3.2%, while Render moved to $1.32 from $1.38, a decline of approximately 4.3%.

CoinGecko reported market capitalizations of approximately $1.86 billion for Bittensor, $683 million for Render and $145 million for Akash as of the August 7 observation.

These weekly comparisons are separate from each token's reported 24-hour percentage change. AltStreet uses the dated observations to preserve consistency across weekly market wraps.

Carbon-token liquidity remains thin

Nature Carbon Tonne was effectively unchanged between AltStreet's weekly observations, trading at $0.2238 on August 7 compared with $0.2237 on July 29.

Klima Protocol K2 declined to $0.01422 from $0.01543, while kVCM declined to $0.02275 from $0.02425.

The kVCM observation remains particularly liquidity-sensitive. The token carried only about $1,000 of reported daily trading volume across five decentralized-exchange pairs.

Prices in markets of that depth can move materially on small transactions, so AltStreet treats kVCM as a thin-liquidity indication rather than a robust mark for the value of the underlying carbon portfolio.

Music-royalty deal flow falls below the weekly median threshold

AltStreet did not publish a seven-day Royalty Exchange median entry yield for the week ending August 7 because fewer than three usable closings were recorded, below the minimum sample threshold for the weekly series.

The trailing 28-day window contained 10 usable deals with a median last-twelve-month entry yield of approximately 16.1%.

Using a minimum-observation threshold prevents a single small closing from being presented as a representative weekly market yield when transaction activity is sparse.

AltStreet's broader marketplace study covers 2,683 economically distinct music-royalty transactions and explains how catalog age, duration and deal size relate to observed pricing: https://altstreet.investments/blog/what-music-royalties-actually-sell-for-marketplace-data

Forge changed FG-GWS to a 3(c)(7) structure with sales unchanged

The week's most analytically significant Form D amendment was not a capital-raise increase.

Fund FG-GWS, a series of Forge Investments, LLC, filed a Form D/A on August 5 with cumulative amount sold unchanged at $4,215,007 and its investor count unchanged at 24.

The structural change was in the fund's claimed Investment Company Act exclusion: the August amendment identifies Section 3(c)(7), while the preceding April filing identified Section 3(c)(1).

A traditional 3(c)(1) fund is generally subject to a 100-beneficial-owner framework, while a 3(c)(7) vehicle generally limits ownership to qualified purchasers. The filing does not explain whether the change represents revised governing documents, a correction of the earlier filing or a change in future investor eligibility.

AltStreet's current Forge review documents the platform's fund structures, investor-eligibility differences and broader Form D reporting history: https://altstreet.investments/platforms/reviews/forge-global

EDGAR activity remained moderate

Across AltStreet's tracked entity universe, the trailing seven-day filing flow included five entity-growth signals, six Exchange Act filings, two new Form 1-U filings, two offering-circular supplements, one offering-stalled signal and one amendment-spike signal.

Those counts describe AltStreet's monitored universe rather than all SEC filings submitted during the week.

The distinction matters because the module is designed to measure filing activity among entities already mapped into AltStreet's alternative-investment database. It should not be interpreted as a census of U.S. private-market issuance.

Capital-formation signals are also based largely on issuer-reported cumulative Form D amounts sold. Changes between filings can identify reporting movement but do not by themselves establish transaction dates, net inflows or portfolio performance.

Hiive50 remains up more than 50% over twelve months

The Hiive50 Index stood at 165.68 as of August 7, representing a reported 52.1% increase over twelve months.

Hiive describes the index as an equal-weight measure of 50 of the most liquid securities trading on its private-company marketplace.

AltStreet uses the published index level as an attributed private-market reference and does not ingest or republish Hiive constituent-level pricing.

The index should not be treated as a broad valuation measure for all venture-backed or private companies because its composition is intentionally concentrated in the more actively traded names on Hiive.

How to read this week's numbers

AltStreet Tracked On-Chain Private Credit TVL and AltStreet Tracked RWA TVL are scoped aggregates, not measurements of the entire private-credit or tokenized-RWA markets.

Both current TVL universes contain only two protocols, and the private-credit aggregate is more than 99.9% Maple Finance. Concentration should therefore be considered alongside every headline change.

TVL is a stock measure and should not be described as net investor inflow. Week-to-week changes can reflect deposits, withdrawals, asset prices, accounting methodology and protocol-level composition.

Stablecoin yields are observed pool rates rather than guaranteed investor returns. They can change with utilization, incentives, underlying collateral yields and market conditions.

GPU rental medians are based on qualifying observed offers and are particularly sensitive to sample size in thinner categories such as RTX 4090 spot capacity.

Weekly token comparisons use AltStreet's dated observations and are distinct from 24-hour percentage changes reported by market-data providers.

Form D amount-sold figures are cumulative issuer-reported values, not AUM, NAV, transaction-date flows or audited cash movements.

Data and methodology

AltStreet Weekly Markets is a descriptive data publication. AltStreet Tracked RWA TVL, Tracked On-Chain Private Credit TVL, Large-Pool Stablecoin Yield Median and GPU Rental Index are derived from defined coverage universes and should be interpreted within those scopes.

Energy data is sourced from U.S. Energy Information Administration series. Crypto-market observations are sourced from CoinGecko. Royalty Exchange data is derived from publicly disclosed past-deal information. The Hiive50 level is used solely as an attributed index reference.

Nothing in this publication constitutes investment advice, a recommendation or a prediction of future market performance.

See the opportunity. Understand the risk.

Get the latest on AltStreet Weekly Markets

Follow the latest on AltStreet Weekly Markets, discover new deals, and keep an eye on the risks. Get market news, platform updates, and independent research that helps you see beyond the pitch, delivered to your inbox.

New dealsMarket newsPlatform updates

Make your next move an informed one

Get the latest on AltStreet Weekly Markets and the wider alternative markets. Join free.

Independent intelligence from AltStreet. No hype. No sponsor spin.

Sources

  1. Maple Finance TVL — DefiLlama
  2. Goldfinch TVL — DefiLlama
  3. Ondo Finance TVL — DefiLlama
  4. Centrifuge TVL — DefiLlama
  5. Cushing, Oklahoma WTI Spot Price FOB — U.S. Energy Information Administration
  6. Henry Hub Natural Gas Spot Price — U.S. Energy Information Administration
  7. Weekly Natural Gas Storage — U.S. Energy Information Administration
  8. Akash Network Market Data — CoinGecko
  9. Bittensor Market Data — CoinGecko
  10. Render Market Data — CoinGecko
  11. Fund FG-GWS Form D/A — August 5, 2026 — SEC EDGAR
  12. Fund FG-GWS Form D/A — April 21, 2026 — SEC EDGAR
  13. Forge Global Review — AltStreet Research
  14. AltStreet Yield Monitor — AltStreet Research
  15. What Music Royalties Actually Sell For: Data From 2,683 Marketplace Deals — AltStreet Research
  16. Tokenized Equities in 2026: Valuation Marks, Venue Spreads, and What Tokenized Stocks Are Really Worth — AltStreet Research

This article was produced by AltStreet's filing-monitoring systems and reviewed by a human editor before publication. Figures sourced from SEC EDGAR filings are cited by accession number; platform-reported figures are unaudited. See our editorial policy and verification policy. Nothing here is investment advice.

Questions about this article, corrections, or additional information — including from the entities covered: [email protected]. Substantive responses from covered entities are added to articles per our editorial policy.

Related: Comparison terminal