Tuesday night's Trump-Iran ceasefire announcement triggered one of the sharpest single-day risk reversals in months. The S&P 500 broke its 200-day moving average for the first time since the conflict began. Crude collapsed on confirmation that Iran would allow passage through the Strait of Hormuz for a two-week truce period.

Crypto caught the bid. BTC surged from $68K to $72.7K, triggering nearly $600M in leveraged liquidations — over $400M from shorts. Equities proxies tracked the rally; CRCL and HOOD lagged on CLARITY Act uncertainty.

The Nilson Report put global card fraud losses at $33.4 billion in 2024. A vendor-commissioned Oxford Economics study measured $50.7 billion in direct merchant revenue lost to false declines across just four markets in a single year. The industry has built an entire discipline around fraud. It has built almost nothing around the revenue it throws away.

The asymmetry is structural. The risk stack — identity verification, fraud scoring, authentication — was designed layer by layer, measured against its own metrics, incentivised by its own penalties. Nobody designed the layers to work together. The seams between them are where legitimate customers get turned away.

A returning customer, recognised device, matching billing region, three prior purchases. The acquirer's ML model scores it low risk. High approval likelihood. Then the acquirer builds the ISO 8583 authorization message — a format originally built in 1987 and still reflecting implementation guides updated as recently as March 2025.

Each layer in the authorization chain — acquirer, network, issuer — sees a different slice of reality and passes on a compressed version. The basis-point gap between best-in-class and average authorization performance lives not in any single layer, but in the seams between them. The data richness that the acquirer holds never reaches the issuer.

Forrester's latest merchant payments wave confirms a growing capability gap between the three leading global processors and the rest. Nuvei remains a contender. Worldpay earns credit for relationship quality. No one else is competitive for enterprise buyers. Enterprise focus has shifted: high acceptance rates and low prices are now baseline expectations, not differentiators. The new battleground is pace of capability addition and quality of relationship.

Legacy players are under pressure. Worldline raised €392M from three French banks as part of a €500M capital round after last summer's near-collapse. ANZ is moving to unwind its JV. Nexi lost its CEO following a capital markets day that sent the stock down 16%; the company is now yielding 9% as it pivots from M&A growth to cash generation.

Neobanks are entering the acquiring market. Revolut Merchant is growing; Monzo and N26 are watching.

Revolut's 2025 results confirm it as the most valuable private technology company in Europe. The a16z "Rule of 75" — revenue growth rate plus net profit margin — puts Revolut at 75 (46% growth + 29% margin). Few financial institutions, modern or established, have reached that level.

Revenue is unusually diversified for a neobank: 76% fee-based, with no single segment above 22% of total turnover. Card payments lead at 22%, followed by interest income at 21.6%. Subscription revenue at 15.7% is the highest-quality component — recurring, predictable, expanding. Revolut enters its fifth consecutive year of net profit.

Markets are pricing Circle as a rate-sensitive money market fund on blockchain rails. The Artemis case is that this misses the point. USDC supply grew 72% in 2025 even as the Fed cut rates 75bps — demand is utility-driven, not yield-driven. The GENIUS Act has created a federal stablecoin framework that advantages regulated compliant issuers. Circle Payments Network has enrolled 55 financial institutions with $5.7 billion in annualised TPV.

The structural cost risk remains: Coinbase takes 100% of reserve income on USDC held on its platform, and 50% off-platform. In 2025, Coinbase received $1.35 billion — 51% of Circle's gross reserve income. Distribution costs at 60%+ of gross are the constraint, not the revenue trajectory.

On 18 March 2026, Stripe and Tempo Labs launched the Machine Payments Protocol (MPP) and the Tempo mainnet. The core premise: AI agents do not use browsers. They do not click buttons or solve CAPTCHAs. When an agent needs a resource, payment is a programmatic prerequisite. MPP formalises this by reviving the HTTP 402 "Payment Required" status code into a standardised machine-readable challenge-response framework.

For thirty years the internet lacked a native value layer. Card rails were designed for human psychology — friction that reduces cart abandonment is an anti-feature for an agent that has no psychology to manipulate. MPP is a direct attempt to replace that friction with a protocol layer.

Artemis published a weekly-rebalanced long-short factor built on stablecoin flows between chains. The thesis: when dollars in stablecoin form flow onto a blockchain, native token demand follows. The factor longs the top quintile by inflow and shorts the weakest. Five-year Sharpe of 1.67, annualised return of 83.6%, market beta of effectively zero.

The long leg generates 84% of returns — this is fundamentally a positive-flow detector for mid-cap L1s and L2s, not a structural short on Ethereum. Five chains (Polygon, Mantle, Optimism, BSC, Sei) account for 84% of total returns.