Trade-press brief (Payments Dive, Dive Brief format, August 6, 2026) covering Block's quarterly earnings release: the company has already rolled out several AI tools to its customers — Moneybot (Cash App) and Managerbot (Square) — and has not yet decided how to charge for them. Jack Dorsey on the analyst call: "We're in a fortunate position where we can experiment with a number of models, and then choose the right one that's going to align all of our incentives with our customers."The financial backdrop illuminates that stance. Six months earlier, Block had laid off roughly 4,000 people, about 40% of its workforce, in a reorganization explicitly framed around AI.
A Payments Dive brief from August 6, 2026 on Block's quarterly earnings, the parent of Cash App, Square, and Afterpay.
The stated topic. Block has rolled out several AI tools to its customers and has not yet decided how to charge for them. Jack Dorsey, on the analyst call: "We're in a fortunate position where we can experiment with a number of models, and then choose the right one that's going to align all of our incentives with our customers." The company is consulting Square merchants on their needs. Evercore ISI analysts list four possible paths — SaaS bundles, direct subscriptions, enterprise offerings, usage-based pricing — noting that Block is prioritizing "product quality, distribution, and adoption" first.
The real story, left for the reader to piece together.Six months earlier, Block laid off roughly 4,000 people, ~40% of its workforce, in a reorganization centered on AI. In Q2 2026, gross profit rose 25% to $3.2B while revenue climbed only 10% to $6.62B; net income fell to $89M, down 83%, weighed down by severance costs; 2026 guidance was raised. Not a single dollar of AI has been billed to customers: the value has already been captured through the cost structure. The "fortunate position" that lets Dorsey take his time on pricing is exactly what the workforce cut bought.
The buried number. In the shareholder letter: "Starting in June, agentic AI helped write and review nearly all of our production code changes." Writing and reviewing nearly all production code changes, at a publicly traded payments company. A self-reported claim to investors, with no definition of "nearly all" or of "review."
The tooling.Goose, an "agnostic" internal system built two years earlier, plugging in several commercial models for employees. Buzz, launched the previous month, for agent collaboration, communication, and code repositories. On the customer side, Moneybot (Cash App) tracks activity, surfaces accounts, balances, and transactions, and has passed one million weekly active accounts; Managerbot runs automated marketing and margin analysis for Square merchants.
Two distribution facts.Square is entering Google Maps with a conversational discovery-and-ordering experience, "the first step of a broader partnership" with Google. And the Tags device (NFC) shows three million people on the waitlist.
Key takeaways
The reversal to make while reading this headline. The article is titled "Block explores how to price AI" and, without saying so, describes how Block has already monetized its AI — through costs, not through price: | | Q2 2026 | vs 2025 | |---|---|---| | Gross profit | $3.2B | +25% | | Revenue | $6.62B | +10% | | Net income | $89M | −83% (severance) | | Headcount | — | ≈ −4,000, about 40%, six months earlier | → Gross profit is growing two and a half times faster than revenue, without a single AI dollar billed to customers. The "fortunate position" Dorsey claims for taking his time on pricing is exactly what the workforce cut bought. That's Block's real AI business model in Q2 2026, and it doesn't appear in any of the four monetization paths listed by analysts.
The number to pull from this brief, if only one."Starting in June, agentic AI helped write and review nearly all of our production code changes" (Dorsey, shareholder letter). Two verbs, not one: writing and reviewing. At a publicly traded payments company, on production code, six months after a 40% headcount cut. Three caveats to carry alongside the number: (1) self-reported to investors, on the day of a favorable earnings release; (2) "nearly all" is undefined, and "helped" leaves open the entire range between autocomplete and autonomous agent; (3) "review" in a chain where the agent also wrote the code raises the question of who is reviewing what. Cite it with the source and date, never as a measurement. Compare with [[cherny-wu-reflecting-year-claude-code-2026-07-17]] and [[dora-google-cloud-roi-ai-assisted-software-development-j-curve-2026-04-21]] on the gap between share of AI-generated code and actual gains.
The four monetization paths listed by Evercore ISI — and the fifth that's missing.SaaS bundles · direct subscriptions · enterprise offerings · usage-based pricing. All four are classic software billing forms. None of them is tied to outcomes. The industry debate over outcome-based pricing ([[greenwald-sierra-outcome-based-pricing-ai-agents-2024-12-10]], [[kamelman-thoughtworks-service-as-software-economic-model-ai-agents-2025-12-03]], orq-ai-finops-ai-agents-cost-per-outcome-hosseini-2026-04-15) has not reached the sell-side analyst note on a mainstream deal. A useful observation for anyone building an offering: the sell side is still thinking in terms of SaaS billing forms.
The stated sequence, reusable as a framework.product quality → distribution → adoption → pricing model. Block openly puts billing last. The unstated corollary: this sequence only holds if someone else is paying for the interim — here, the cost structure. A company without slack in its cost base cannot copy this sequencing, even if it copies the talking points.
Dorsey's phrase, decoded."choose the right one that's going to align all of our incentives with our customers." Incentive alignment is the standard argument for outcome-based pricing — the one path missing from the analysts' list. Worth watching: if Block eventually announces pricing tied to value produced (margin recaptured at the merchant, for instance), this August 2026 line will have been the early signal.
The four building blocks, distinguished. | Tool | For whom | What it does | Stated traction | |---|---|---|---| | Moneybot | Cash App customers | monitors activity, surfaces accounts, balances, transactions, suggests features | > 1M weekly active accounts | | Managerbot | Square merchants | automated marketing, margin analysis, "operational fixes" | — | | Goose | employees | "agnostic" internal system plugging in several commercial models; built two years earlier | — | | Buzz | teams | agent collaboration, communication, code repositories; launched the previous month | — | → Only Moneybot carries a usage figure. The other three are described by intent alone.
Buzz seen from here vs. seen from elsewhere — the same object, two narratives. In girard-acp-deux-protocoles-un-sigle-2026-08-02, Buzz is a self-hostable workspace under Apache-2.0 built on Nostr, valued for demonstrating agent/client decoupling via ACP (Goose, Claude Code, and Codex plugged into the same harness). Here, in a shareholder letter, it's "an internally developed AI system" that saves teams time and speeds up delivery. Same artifact, two audiences, two framings of value: protocol openness on one side, productivity on the other. A good reminder that how an announcement reads depends first on the document it's found in.
The distribution fact not to miss.Square is entering Google Maps with a conversational experience where the customer discovers a restaurant and orders by asking questions — and Block frames it as "the first step in a broader partnership between Square and Google." This isn't a product detail: it's merchant discovery and transaction moved onto a third party's surface, worth reading alongside the agentic-commerce corpus (cloudflare-wallets-agentic-commerce-2026-08-04, google-agentic-commerce-ap2-payment-protocol-2025-09-16, nrf-2026-commerce-agentique-ucp-deep-research-2026-01-13).
Citation hygiene for this document. this is a trade-press brief, not an analysis. Everything it reports on AI comes from Dorsey, the shareholder letter, or sell-side analyst notes, with no counterpoint or independent verification. Cite it as "Dorsey, Q2 2026 shareholder letter, via Payments Dive," never as "Block measured that." The only auditable figure in the article is the income statement.
Meta / cross-references. on Block and Goose in this corpus, block-goose-mcp-ui-future-agentic-interfaces-2025-08-25 and mcp-ui-future-agentic-interfaces-goose-2025-08-25; on AI as a justification for headcount cuts, sfeir-ia-emploi-risque-decrochage-2026-07-23 and sternfels-mckinsey-60000-people-20000-agents-officechai-2026-01-14; on publisher margin structure in the AI era, bain-ai-rule-of-40-headwinds-tailwinds-saas-2026-04 and bain-100b-saas-opportunity-cross-system-labor-agentic-ai-2026-05.
Key figures
gross profit up 25% to $3.2B in Q2 2026, on revenue up 10% to $6.62B
Block is in a comfortable position to experiment with several pricing models before choosing one that aligns its incentives with those of its customers
— Jack Dorsey
since June, agentic AI has helped write and review nearly all of Block's production code changes
— Jack Dorsey
les outils d'IA de Block ont quatre voies de monétisation possibles : bundles SaaS, abonnements directs, offres entreprise et tarification à l'usage
— Evercore ISI
The knowledge graph extracted from this fiche — 11 entities, 26 relations.
In this graph :Block · monétisation par les coûts · Moneybot · Managerbot · Goose · Buzz · Cash App · Square · Tags · Jack Dorsey · Evercore ISI