Block Inc. Cuts 40% of Workforce Using AI Agent “Goose”: The First Empirical Case of Mass AI-Driven Labor Displacement

Restructuring From Strength, Not Distress

In late February 2026, Block Inc. (formerly Square) executed what analysts at Evercore ISI described as a “seminal moment” in the AI era [2]. CEO Jack Dorsey announced the reduction of Block’s global workforce from over 10,000 employees to fewer than 6,000—a structural cut of approximately 40% that eliminated more than 4,000 positions in a single announcement [1].

What distinguishes this event from conventional corporate layoffs is the financial context. Block executed this mass termination from a position of profound operational strength. For the full year 2025, Block reported gross profit of $10.36 billion, up 17% year-over-year [1]. The fourth quarter was particularly robust, with gross profit jumping 24% to $2.87 billion, heavily fueled by a 33% surge in the Cash App ecosystem, which generated $1.83 billion on its own [1].

The company simultaneously raised its forward guidance. For 2026, Block expects to generate $12.2 billion in gross profit with adjusted EPS of $3.66, alongside a Q1 2026 operating income projection of $600 million that easily surpassed the $574 million consensus [1].

The Technological Catalyst: Autonomous Agent “Goose”

The primary driver enabling this restructuring is not outsourced labor but an internal, open-source AI developer agent named “Goose” [3]. Created by Block engineers and released under an open-source license, Goose represents a fundamental evolution beyond first-generation AI coding assistants that merely suggest code snippets.

Goose operates as a fully autonomous developer agent. It functions locally via a command-line interface (CLI) or desktop application, interacts directly with the Model Context Protocol (MCP), and integrates with any preferred Large Language Model (LLM) to prevent vendor lock-in [4]. The agent independently runs tests, interprets error codes from vendor APIs, executes iterative debugging workflows, and pushes code to production environments without human intervention [5].

Unlike traditional automation tools that require explicit programming for each task, Goose reasons about complex software engineering problems and chains multiple actions together autonomously. It can navigate unfamiliar codebases, identify bugs from stack traces, propose and test fixes, and validate deployment pipelines end-to-end [4].

The Market’s Response: Rewarding Efficiency Over Employment

The market reaction was explosive. Block shares surged between 22% and 25% in after-hours trading, rapidly approaching $69.00 per share [1]. Investors aggressively rewarded the immediate margin expansion implicitly guaranteed by the permanent removal of over 4,000 salary obligations, calculating that long-term operational expenditure savings will vastly outweigh the estimated $450 million to $500 million one-time restructuring charge [1].

The severance package, while substantial by technology industry standards, was designed for a clean, singular break: 20 weeks of base pay, accelerated equity vesting, and six months of COBRA health coverage [1]. Dorsey explicitly opted for a “single deep round of cuts” rather than gradual attrition, betting that a rapid, decisive restructuring would minimize organizational uncertainty and allow the leaner company to immediately increase product velocity [1].

Dorsey’s Thesis: Intelligence Tools Have Changed What It Means to Build a Company

In his shareholder communication, Dorsey tied the layoffs directly to AI capabilities: “Intelligence tools have changed what it means to build and run a company… A significantly smaller team, using the tools we’re building, can do more and do it better” [2]. This statement represents one of the most explicit public acknowledgments by a Fortune 500 CEO that artificial intelligence has fundamentally altered the labor economics of software engineering and corporate operations.

Dorsey’s framing positions the restructuring not as a cost-cutting measure during financial hardship, but as an affirmative strategic pivot toward an AI-augmented operating model. The implicit argument is that the combination of a smaller, highly skilled human workforce and autonomous AI agents will produce superior engineering output at dramatically lower cost.

This positions Block as the first major public company to empirically demonstrate the full labor arbitrage potential of generative AI and autonomous agents at corporate scale. The immediate stock market reward creates a powerful incentive structure: boards of directors across the Fortune 500 will face intense pressure to replicate Block’s AI-driven margin expansion, potentially triggering a wave of similar restructurings throughout the technology and financial services sectors.

Goose Architecture: How the Agent Replaces Human Workflows

Understanding the technical architecture of Goose is essential for evaluating the durability and scalability of Block’s labor arbitrage strategy. Goose is built on the Model Context Protocol (MCP), an open standard that allows AI agents to interface with development tools, databases, APIs, and deployment pipelines through a unified protocol layer [4].

The agent’s key capabilities include [5]:

• Autonomous test execution: Goose runs entire test suites, interprets failures, generates fixes, and re-validates until tests pass.
• Multi-step debugging: When encountering an error from a vendor API or internal service, Goose analyzes the error context, searches documentation, and implements corrective code changes iteratively.
• LLM-agnostic operation: The agent supports any compatible LLM backend, preventing vendor lock-in and allowing Block to switch between models as capabilities and pricing evolve.
• Local-first execution: Unlike cloud-hosted copilots, Goose runs locally on developer machines, reducing latency and maintaining security by keeping source code on-premises.
• Production deployment: Goose can validate CI/CD pipelines and push approved changes to production environments autonomously.

Block has further institutionalized AI adoption through the Goose Grant Program, providing financial incentives for open-source developers to extend the agent’s capabilities and build integrations with external toolchains [3].

Macroeconomic Implications: The Deflationary Labor Spiral

The macroeconomic implications of Block’s restructuring extend far beyond one company’s operating margin. This event validates a class of highly speculative economic warnings—including the “2028 Global Intelligence Crisis” thesis by Citrini Research and Alap Shah—that hypothesize widespread AI automation will trigger structural displacement of white-collar workers, eventually producing a deflationary economic spiral [6].

The mechanism is straightforward: consumer spending drives approximately 70% to 80% of the U.S. economy. As major technology companies systematically substitute human capital with AI agents, aggregate purchasing power diminishes. The efficiency gains realized by individual corporations through automation are beginning to disconnect from—and actively undermine—the long-term health of the broader economy.

Block’s implementation of Goose represents the first empirical case study of this macroeconomic theory executing in real time. The stock market’s enthusiastic 24% reward for mass terminations guarantees that similar actions will proliferate across industries, accelerating the timeline for structural labor market disruption.