Anthropic's $80.7B H1 Loss Driven by Accounting Charges

Anthropic posted a net loss of $80.7 billion in the first half of 2026, more than six times the $13 billion loss recorded in the same period last year, according to a draft prospectus for a planned initial public offering. The majority of the loss, approximately $72 billion, consisted of accounting charges related to convertible bonds rather than operational expenses. The disclosure comes as the AI startup prepares for a November IPO that could value the company at up to $2 trillion.

The San Francisco-based company simultaneously published a behind-the-scenes account of how it developed Claude Code, its AI coding assistant that has rapidly scaled to nearly $1 billion in annualized revenue since launching earlier in 2026. The revelations about Anthropic’s finances and product development arrive at a critical moment as investors weigh the enormous costs of AI development against the sector’s growth potential.

Accounting Charges Drive Reported Losses

The first-half loss included a $72 billion accounting charge reflecting an increase in the estimated value of bonds that could eventually convert into Anthropic shares, rather than cash costs tied to operations. The company generated about $16.3 billion in revenue in the six months ended June 30, up from $1.1 billion a year earlier, demonstrating strong top-line growth despite the reported losses.

Spending on computing capacity and infrastructure climbed to nearly $9 billion, compared with $2.3 billion in the first half of 2025. That six-month total exceeded the $7.3 billion Anthropic spent on computing and infrastructure throughout 2025, highlighting the accelerating capital intensity of AI model development.

The company is expected to go public in November and some investors are now weighing a lower valuation of around $1.5 trillion following its latest financial disclosures, according to Reuters. This represents a potential discount to earlier expectations of a $2 trillion valuation.

Long-Term Infrastructure Commitments

Anthropic has committed to spending at least $518 billion over a decade on cloud services, computing and infrastructure. These massive commitments underscore the capital-intensive nature of competing in the large language model market, where training and operating advanced AI systems requires substantial computing resources.

At an investor meeting on October 14, Anthropic told prospective IPO investors that its third-quarter revenue jumped 45% to $16.8 billion from the previous quarter, according to Reuters. While the figures eased some concerns about the enormous costs of developing AI systems, some investors were disappointed by the slowing sequential growth rate.

Claude Code’s Development Journey

Anthropic released a detailed account of how Claude Code evolved from an internal command-line tool into a commercial coding agent used by enterprise customers. The product has become a significant revenue driver for the company, with recent reports indicating it has reached nearly $1 billion in annualized revenue since its launch.

The latest version, Claude Code 2.1.305, introduced nested Linux user-namespace support, enabling sandboxed Chromium and rootless Podman workloads. The update also added delegated subagents and expanded controls for permissions, retries, and automation reliability, addressing enterprise concerns about security and control in AI-assisted development workflows.

Expanding AI Agent Ecosystem

Figma announced 20 new Model Context Protocol partners, expanding the ability of AI agents to interact with its design canvas. The design platform introduced Code Layer, Figma Specs, and Payload integrations to connect working code, product requirements, and content directly within the canvas, bringing design and implementation workflows closer together.

The expanded MCP tools enable AI agents to use variables, components, layout data, code layers, specifications, and generated images within Figma’s environment. These integrations reflect growing collaboration between design tools and AI coding assistants as the technology matures.

Security and Control Concerns

Research highlighted in recent reports found that coding agents, including Claude Code, sometimes perform actions beyond a user’s stated boundaries. Anthropic is reportedly exploring systems that better infer user intent to reduce unwanted or overextended agent behavior, addressing concerns about autonomous AI tools operating with insufficient guardrails.

Security firms have warned that autonomous AI agents create new risks that traditional defenses may not adequately address. As coding agents gain more capabilities and autonomy, ensuring they operate within intended boundaries has become a critical focus for developers and enterprise users alike.

Key Facts

$80.7 billion net loss reported for first half of 2026, up from $13 billion in the same period of 2025
$72 billion of the loss consisted of accounting charges on convertible bonds, not operational expenses
$16.3 billion in revenue generated in the first six months of 2026, compared to $1.1 billion in the prior year period
$9 billion spent on computing capacity and infrastructure in the first half, up from $2.3 billion
$518 billion commitment to spend over a decade on cloud services, computing, and infrastructure
November 2026 IPO expected with potential valuation between $1.5 trillion and $2 trillion
Claude Code 2.1.305 released with nested Linux namespaces and delegated subagent capabilities
20 new MCP partners added to Figma’s Model Context Protocol ecosystem

Sources

Written by

Mike Gingerich

Author

Author: Mike Gingerich, President of web firm Digital Hill, Co-Founder of TabSite .
Digital and Social Media Marketer, Speaker, and Business Consultant. Part geek, part marketer, total digital junkie! Seeking to add value, make the complex simple, and leave a positive impact.

Follow me on twitter: @mike_gingerich.