Hi all, happy Wednesday! Today, we break down how the softest CPI print of the year completely reversed Monday's geopolitical selloff, sparking a broad risk-on rally led by a massive surge in Privacy. 

Plus, we dive into the data behind Hyperliquid to explain why its shrinking revenues and market share don't tell the whole story.

Let’s get into it.

The softest consumer price index (CPI) print of 2026 flipped Monday's geopolitical selloff into a broad rally: BTC +4.3%, everything green except miners. Privacy stole the session at +10.3%, with ZEC ripping into its July 28 Ironwood upgrade. We break down the move below.

Across the majors the week is one round trip. Everything bled into Monday's Hormuz panic, BTC briefly 3% underwater, before Tuesday's CPI reversed the move in a session. BTC leads the benchmarks at +2.1% on the week, ahead of NASDAQ (+1.2%), the S&P 500 (+0.5%), and gold (-1.5%), which gave back its geopolitical bid as inflation cooled. Weekly leadership from an asset that has traded heavy for months reads as either an oversold snapback or early regime change; Fear & Greed at 29 says the market has not decided.

Tuesday saw broad green with Privacy the clear standout at +10.3%; most sectors clustered in the 2% to 7% range and miners were the lone red at -1.0%. The bid was June CPI: headline fell 0.4% on the month, the largest drop since April 2020, with annual inflation at 3.5% against 3.8% expected, erasing July 29 Fed hike risk and liquidating $68M of shorts within hours.

Inside the Privacy index, ZEC (48.3% of the index) did the heavy lifting at +14.3% on the day and +15.7% on the week; DCR quietly leads weekly at +24.4% without a clear catalyst, and XMR (41.1% weight) lagged at +3.1%.

ZEC's bid is upgrade-driven. Ironwood activates July 28 at block 3,428,143, and developers say they are close to mathematically proving the next-generation privacy system free of a critical counterfeiting vulnerability class, restoring confidence after last month's flaw disclosure. Futures volume above $1.35B in 24 hours says the move carries leverage.

One week doesn't make a bottom, but the setup is there. A geopolitical shock failed to stick, BTC is leading the benchmarks at +2.1%, and 20 of 21 sectors closed green while Fear & Greed still reads 29. Lows tend to form when the market rallies and nobody trusts it.

— AJC

Declining revenues, shrinking market share and price near an all-time high

Hyperliquid has recorded three consecutive quarters of declining revenue.

Its perp volume has been on a downward trend…

…falling alongside its share of open interest.

As for its hot new sector, RWAs? 

Increased competition has led to three months of declining volume share.

Yet HYPE is near all-time highs.

And from headline figures, it isn't cheap compared to many peers. 

So why the hype?

Well, that declining revenue is industry-wide, and against that backdrop, Hyperliquid's share of quarterly revenue has actually risen to an all-time high.

Nearly one-third of every dollar spent on apps goes to Hyperliquid.

Meanwhile, open interest has been rebuilding since February 2026. At $11B, it's still roughly 30% below the 2025 summer peak, but the composition has improved. OI is more diversified across products, even if that comes at the cost of blended margins.

Thanks to this second material revenue stream, Hyperliquid has gone from a high beta play in crypto bull runs and added defensibility during crypto winters. 

Even as RWA volume share declines, nominal volume continues to grow.

Moreover, the category itself is expanding quickly, both in volume and competition.

Competition isn't necessarily a headwind here. It can legitimize a category, de-risk the decision for institutions to participate, and grow the total revenue pool. 

Pre-existing crypto ETF issuers are a precedent: CoinShares' revenue and EBITDA rose roughly 90% (shown below) in the same year BlackRock, Fidelity, and other large institutions launched competing products. The same pattern played out in tokenized treasuries (Franklin Templeton, BlackRock), stablecoins (PayPal, Circle), and prediction markets (Robinhood, Kalshi). 

Hyperliquid also carries optionality beyond the current revenue mix. HIP-3 builder-deployed perps already generate ~10% of revenue. HIP-2 contributes ~2% while HIP-4 outcome markets contribute just 0.014% - a potential third revenue leg if adoption carries over.

There's also optionality from HIP-5s through 10, whatever they may be, and execution is sourced from builders, which makes it operationally scalable.

On the valuation front, earnings are the fairer comparison because Hyperliquid is already profitable despite its age. On this basis, HYPE trades around 19x, cheaper than many of its peers, but with a growth rate they don't have. 

The market is not paying for Hyperliquid’s last few quarters, it’s pricing the durability of its next cycle. Revenue has softened, but its share of industry spend is at a record high, open interest is rebuilding on a broader product base, and new revenue streams are beginning to take shape. If perps activity recovers and Hyperliquid retains its current share, today’s valuation may prove less demanding than the headline multiples suggest. 

— Marc

Inference Capital Markets

As global AI demand shifts from resource-heavy training to everyday execution, a new decentralized layer is emerging: onchain inference capital markets. Rather than routing queries through centralized tech giants, these networks utilize crypto rails to turn AI compute into a tradable commodity. This financialization is unfolding across three key fronts: tokenizing access to turn future compute into ownable, tradeable assets (like Venice's DIEM); tokenizing production through useful proof-of-work networks that subsidize and verify model outputs (like Pearl and Ambient); and financing physical hardware by turning real-world GPU fleets into yield-bearing opportunities for stablecoin depositors (like USD.AI). While the ecosystem currently wrestles with high speculation and the challenge of bridging token utility with organic demand, it is rapidly laying a permissionless, programmatic financial foundation uniquely primed to power the next generation of autonomous AI agents.

TradFi doesn’t want DeFi. It wants blockchain

Rather than merging with open DeFi, traditional financial institutions are selectively adopting blockchain technology as a cost of goods sold (COGS) play to compress costs, improve atomic settlement, and expand distribution. By stripping away native DeFi's core features such as open access, pseudonymity, and trustless execution, TradFi is reconfiguring these primitives into a highly controlled, compliant, and permissioned category of programmable financial infrastructure. For Web3 builders, this bifurcated reality creates two distinct, non-competing opportunities: constructing compliant, enterprise-grade rails optimized for institutional procurement, or continuing to pioneer open, permissionless crypto systems. Ultimately, these parallel paths reinforce one another, as the open DeFi sandbox will continue to yield the cutting-edge primitives that TradFi eventually adapts and integrates onto shared underlying blockchain rails.