“
Claude Leaked, OpenAI Hits $852B: Every Major AI Story From April 1, 2026
April Fool’s Day 2026 arrived with no jokes — only breaking news. Anthropic confirmed a source-code packaging leak exposing Claude’s internals, OpenAI closed the largest private funding round in history, and the US Department of Labor launched a national AI apprenticeship initiative. Here’s what this actually means for the industry.
In This Article
01Anthropic’s Claude Code Source Leak: 1,900 Files, 512,000 Lines
02OpenAI Closes $122B Round at $852B Valuation
03Anthropic Finalising $380B Valuation Funding Round
04AWS Frontier Agents Reach General Availability
05US Department of Labor Launches National AI Apprenticeship Initiative
06AI Funding Rounds — April 1, 2026
07Frequently Asked Questions
Anthropic’s Claude Code Source Leak: 1,900 Files, 512,000 Lines
On April 1, Anthropic issued an official statement confirming what security researchers had already flagged: a packaging configuration error had exposed the source tree for Claude Code, the company’s AI-powered coding assistant. The leak comprised 1,900 files and 512,000 lines of code, including internal toolchains, prompt-handling logic, and scaffolding that sits between the model API and the user-facing interface.
Anthropic described the event as a “release packaging issue” — meaning the source bundle was accidentally included in a distributable package rather than being stripped at build time. The company confirmed no model weights were exposed and no customer data was compromised. But the leak handed competitors and security researchers an unusually clear view of how Anthropic structures its production tooling.
Separately, leaked internal documents circulating in AI research circles referenced two upcoming Anthropic model codenames: “Mythos” and “Capybara.” Anthropic has not confirmed or denied the codenames. Wired’s coverage of the Claude Code leak provides additional technical context.
When 512,000 lines of a frontier lab’s production toolchain become public, it doesn’t just affect Anthropic — it gives every rival and every malicious actor a blueprint. The packaging error is fixable in hours. The competitive intelligence exposure lasts years.
OpenAI Closes $122B Round at $852B Valuation
OpenAI confirmed the closing of a $122 billion funding round that values the company at $852 billion — making it one of the most valuable private companies ever. The round was co-led by SoftBank, Andreessen Horowitz (a16z), and D.E. Shaw.
The most consequential element: Amazon committed $50 billion, but the commitment is reportedly contingent on OpenAI reaching defined AGI milestones — a clause that could reshape how investors and regulators define artificial general intelligence. TechCrunch’s full breakdown of the round includes the term structure.
The round positions OpenAI to accelerate compute infrastructure, hire aggressively, and challenge hyperscalers on their own turf. With SoftBank’s backing already deep in GPU infrastructure through SoftBank Vision Fund plays, the strategic alignment is clear.
Anthropic Finalising $380B Valuation Funding Round
Despite the source-code leak making headlines, Anthropic’s business trajectory appears unaffected. The company is finalising a Series G raise that would value it at $380 billion, placing it in the same stratospheric tier as OpenAI. Multiple sources cited a $14 billion annual revenue run-rate as the basis for the valuation.
That revenue figure — if accurate — suggests Anthropic has found meaningful enterprise traction with Claude, particularly through its API and enterprise contracts. The $380B valuation marks a significant leap from its $60B valuation in early 2025. Reuters’ report on Anthropic’s funding round covers the investor lineup.
$122B into OpenAI. $380B valuation for Anthropic. These are not startup rounds — they are infrastructure bets. The money is chasing compute, talent, and the possibility that one of these labs reaches AGI first and redefines every competitive moat in tech.
AWS Frontier Agents Reach General Availability
Amazon Web Services announced that AWS Frontier Agents — its managed AI agent orchestration layer — has reached General Availability. The service allows enterprise developers to deploy long-running, tool-using AI agents on AWS infrastructure with built-in memory, state management, and tool integrations.
The GA announcement comes as AWS deepens its relationship with Anthropic. Claude models are natively available in Frontier Agents via Amazon Bedrock, giving AWS a competitive answer to Azure’s OpenAI-powered agent infrastructure. For enterprises already running on AWS, the GA release removes a key friction point in deploying agentic workloads. AWS official blog on Frontier Agents GA.
US Department of Labor Launches National AI Apprenticeship Initiative
The US Department of Labor officially launched a national AI apprenticeship initiative, formalising pathways for workers to enter AI-adjacent roles through structured, employer-sponsored apprenticeships. The programme targets roles in AI data labelling, model evaluation, AI system administration, and prompt engineering.
The announcement represents a policy acknowledgment that AI adoption is disrupting traditional employment pathways and that new entry-level credentials are needed. The initiative is designed to work alongside community colleges, apprenticeship intermediaries, and major tech employers. DOL official press release.
A national apprenticeship initiative signals that AI workforce displacement is now a mainstream policy concern, not a fringe prediction. The question is whether the pace of programme rollout can match the pace of model capability improvement — historically, it cannot.
AI Funding Rounds — April 1, 2026
| Company | Round | Valuation | Lead Investors |
|---|---|---|---|
| OpenAI | $122B | $852B | SoftBank, a16z, D.E. Shaw |
| Anthropic | Series G (finalising) | $380B | TBC |
| Amazon (contingent) | $50B into OpenAI | On AGI milestone | Amazon |
Frequently Asked Questions
A release packaging error caused Anthropic’s internal Claude Code source tree — 1,900 files and 512,000 lines of code — to be included in a distributable package, making it publicly accessible. No model weights or customer data were exposed.
Internal documents circulating after the leak referenced upcoming models codenamed “Mythos” and “Capybara.” Anthropic has not officially confirmed these codenames or their release timelines.
Amazon’s $50B commitment is reportedly structured around OpenAI reaching defined AGI milestones — effectively tying a massive capital deployment to a technical threshold that remains contested. This creates an unusual incentive structure and raises questions about how AGI will be formally defined for contractual purposes.
AWS Frontier Agents is Amazon’s managed service for deploying long-running AI agents with tool use, memory, and state management. It reached General Availability on April 1, 2026, and natively supports Claude models via Amazon Bedrock.
The US Department of Labor’s national AI apprenticeship initiative creates formal employer-sponsored pathways into AI-adjacent roles including data labelling, model evaluation, AI system administration, and prompt engineering — targeting workers displaced or underserved by traditional credential pathways.
Every week, the biggest AI stories distilled into clear analysis you can act on.