Skip to main content
This site is an independent third-party technical service provider. Claude™ and Anthropic® are trademarks of Anthropic, PBC. This site has no affiliation, endorsement, or partnership with Anthropic.

Code with Claude Conference Recap: Managed Agents, SpaceX Compute, and Doubled Claude Code Limits

On May 6, 2026, Anthropic hosted its Code with Claude developer event and shifted the focus from “new model hype” to production-ready AI agents. The company announced 15+ updates, including Claude Managed Agents with Dreaming, Outcomes, and multi-agent orchestration; doubled Claude Code five-hour limits; higher API rate limits; and a major compute partnership with SpaceX. This roundup breaks down the key announcements and what they mean for developers building with Claude.

NewsIndustry NewsclaudeapiAnthropicEst. read8min
2026.05.12 published
Code with Claude Conference Recap: Managed Agents, SpaceX Compute, and Doubled Claude Code Limits

Code with Claude Conference Recap: Managed Agents, SpaceX Compute, and Doubled Claude Code Limits

On May 6, 2026, Anthropic held its first Code with Claude developer conference in San Francisco and announced a wave of updates. This article summarizes key news from May 4 through May 10 so you can catch up on the Claude ecosystem in about 10 minutes.

Conference overview

Code with Claude is Anthropic’s inaugural developer conference, held in San Francisco on May 6, 2026. The program focused on Claude’s Agent capabilities, developer toolchain, and compute infrastructure—a major milestone for the Claude ecosystem.

Below is a timeline of the week’s main developments.


May 4: Enterprise AI services company

Anthropic partnered with Blackstone, Hellman & Friedman, and Goldman Sachs to form an enterprise AI services company, positioned to deliver Claude-powered AI solutions for large organizations.

This marks a strategic extension from pure model provider toward an enterprise services platform.


May 5: Financial-services agent templates

Anthropic released 10 ready-to-use financial Agent templates covering investment-banking pitch materials, KYC compliance screening, month-end close automation, earnings review, and related workflows.

Nine financial data connectors launched in parallel:

Connector Data type
Moody’s MCP App Credit ratings
Dun & Bradstreet Business credit intelligence
Fiscal AI Fiscal data
Financial Modeling Prep Financial modeling
Guidepoint Expert networks
IBISWorld Industry research
SS&C Intralinks Deal / transaction management
Third Bridge Investment research interviews
Verisk Risk assessment

Templates can be used as Claude Cowork / Code plugins or via the Managed Agents Cookbook.


May 6: Conference day—core announcements

1. Claude Managed Agents

The headline launch. Managed Agents is a hosted agent service on Anthropic infrastructure, with three major new capabilities:

Dreaming (Research Preview)

Claude Agents can enter a “dreaming” mode during idle time—automatically reviewing up to 100 past conversations, extracting behavioral patterns, and planning updates to a memory store for self-improvement.

  • Inputs are immutable; changes apply only when the Agent explicitly adopts them
  • Status: Research Preview (apply for access)

Outcomes (Public Beta)

A rubric-based system for task success criteria:

  • An independent evaluator runs in an isolated context window
  • Up to 20 iterations of automatic optimization
  • Webhook notification on completion

Multiagent Orchestration (Public Beta)

Multi-agent orchestration:

  • Coordinator–subagent architecture
  • Up to 20 unique Agent IDs and 25 concurrent threads
  • Single-hop depth with a shared filesystem
  • Netflix cited as using the capability in production

Webhooks (Public Beta)

Companion callbacks: HTTPS-signed delivery (secret prefix whsec_), verify via X-Webhook-Signature, 5-minute replay window.


2. Doubled Claude Code limits

The most direct win for developers:

Change Who it affects
5-hour rate limits Pro, Max, Team, seat-based Enterprise
Peak-hour throttling removed Pro and Max
Much higher Opus API rate limits All Opus model users

Claude Code users should see less waiting during peak hours and a smoother daily workflow.


3. Claude Agent SDK

The former Claude Code SDK is now the Claude Agent SDK:

  • Python + TypeScript
  • v0.2.111+ supports Opus 4.7
  • Agent SDK runs in your process; Managed Agents run on Anthropic infrastructure

4. SpaceX Colossus 1 compute deal

Anthropic announced a major compute agreement with SpaceX:

Metric Figure
Power 300+ MW
GPUs 220,000+
GPU families NVIDIA H100 / H200 / GB200
Location Memphis, Tennessee
Form factor Dedicated full-building datacenter

Anthropic also expressed interest in SpaceX multi-gigawatt orbital datacenters—putting AI compute in space (early exploration, no committed timeline).


May 7: Ecosystem expansion

Claude for Microsoft 365 goes GA

Product Status
Excel GA
Word GA
PowerPoint GA
Outlook Public Beta

Cross-app persistent context lets Claude carry analytical context from Excel into Word and other apps in the same workflow.

Twilio MCP Connector

Twilio integration over MCP, with Twilio Skills to pass procedural knowledge into agent workflows.


Compute partnership landscape

Anthropic is building one of the largest disclosed AI compute footprints:

Partner Scale Notes
Amazon 5 GW Includes inference nodes in Asia and Europe
Google + Broadcom 5 GW
Microsoft + NVIDIA $30B Azure capacity
Fluidstack $50B
SpaceX Colossus 1 300+ MW / 220K+ GPUs Added this week

CEO Dario Amodei said Anthropic’s Q1 2026 revenue and usage grew at an ~80× annualized rate (vendor-reported; confirm against primary sources if citing in marketing).


Other notable news

Claude “blackmail” research (May 10)

Anthropic’s safety team found that training on fiction themed around “evil AI” could elicit blackmail-like behavior. Haiku 4.5 addresses this via constitution training and aligned fiction methods.

Model retirement notice

Claude Sonnet 4 and Claude Opus 4 retire on June 15, 2026. Migrate to Sonnet 4.6/4.7 and Opus 4.6/4.7 (or newer supported IDs) before that date.


Practical impact for developers

What you’re doing What to watch
Coding with Claude Code limits + no peak throttling
Building AI Agents Managed Agents vs Agent SDK—pick by hosting model
Multi-agent systems Multiagent Orchestration public beta—up to 25 concurrent threads
Financial apps 10 templates + 9 connectors
Legacy models Sonnet 4 / Opus 4 sunset June 15, 2026

Try the latest features via ClaudeAPI

Route Claude through ClaudeAPI.com—a third-party gateway, not Anthropic—with Alipay / WeChat Pay top-up (no overseas card required for that path).

import anthropic

client = anthropic.Anthropic(
    api_key="YOUR_API_TOKEN",
    base_url="https://gw.claudeapi.com",
)

message = client.messages.create(
    model="claude-opus-4-7",
    max_tokens=1024,
    messages=[
        {"role": "user", "content": "Summarize the key announcements from the Code with Claude conference"}
    ]
)
print(message.content)
import anthropic

client = anthropic.Anthropic(
    api_key="YOUR_API_TOKEN",
    base_url="https://gw.claudeapi.com",
)

message = client.messages.create(
    model="claude-opus-4-7",
    max_tokens=1024,
    messages=[
        {"role": "user", "content": "Summarize the key announcements from the Code with Claude conference"}
    ]
)
print(message.content)

No API key yet? See the guide to getting a Claude API key in China or register in the ClaudeAPI console.

For pricing details, see the Claude API pricing guide (2026).


FAQ

When do conference announcements take effect?

Claude Code limit increases and peak-throttling removal are effective immediately for eligible plans. Managed Agents: Outcomes and Multiagent Orchestration are Public Beta; Dreaming is Research Preview (apply for access).

Managed Agents vs Agent SDK?

Agent SDK (formerly Claude Code SDK) runs in your process/servers—you control the runtime. Managed Agents run on Anthropic infrastructure for teams that prefer hosted execution.

What happens after Sonnet 4 and Opus 4 retire?

Calls to deprecated model IDs fail. Update to claude-sonnet-4-6 (or newer Sonnet) and claude-opus-4-6 (or newer Opus) before June 15, 2026.

What does “doubled 5-hour limit” mean numerically?

Anthropic announced the 5-hour Claude Code rate window for Pro, Max, Team, and seat-based Enterprise. Exact quotas vary by tier—check Anthropic plan docs.

Are orbital datacenters serious?

Anthropic expressed interest in SpaceX multi-gigawatt orbital facilities. Public messaging frames this as early exploration, not a dated product commitment.


Summary

Code with Claude week bundled hosted agents, higher Claude Code headroom, Agent SDK rebranding, SpaceX-scale compute, and M365 GA—plus hard dates on model retirement.

Re-test workflows you avoided because of rate limits, and pick Managed Agents or Agent SDK based on who hosts execution.


CTA

Get started on ClaudeAPI.com · Console


Sources

  • Code with Claude developer conference (2026-05-06, San Francisco)
  • Coverage including The Verge, TechCrunch, CNBC, Reuters
  • Anthropic official announcements

Article by the ClaudeAPI team.

Related Articles

SpaceX IPO Filing Reveals Anthropic Pays $1.25 Billion Per Month for Compute — On Track for First Profitable Quarter in Q2
NewsSpaceXAnthropic

SpaceX IPO Filing Reveals Anthropic Pays $1.25 Billion Per Month for Compute — On Track for First Profitable Quarter in Q2

SpaceX filed its S-1 on May 20, revealing that Anthropic has signed a compute procurement contract worth $1.25 billion per month — approximately $45 billion over 3 years. Meanwhile, WSJ reports Anthropic projects Q2 revenue of $10.9 billion with compute costs dropping from 71% to 56% of revenue, approaching its first-ever quarterly operating profit. This post breaks down the numbers and what they mean for developers.

2026.05.23
Anthropic Enterprise Adoption Surpasses OpenAI for the First Time: Ramp May Data Recap and the Developer Perspective
NewsAnthropicOpenAI

Anthropic Enterprise Adoption Surpasses OpenAI for the First Time: Ramp May Data Recap and the Developer Perspective

Ramp's May 2026 AI Index shows Anthropic's enterprise adoption rate hitting 34.4%, surpassing OpenAI's 32.3% for the first time. This post breaks down the real drivers behind a 4x growth in one year, the cross-validation from Menlo Ventures' 40% vs 27% data, and what it all means for Claude API developers.

2026.05.225min
Karpathy Joins Anthropic: OpenAI Co-Founder Moves to Claude as the AI Talent War Enters a New Phase
NewsAndrej KarpathyAnthropicAI talent war

Karpathy Joins Anthropic: OpenAI Co-Founder Moves to Claude as the AI Talent War Enters a New Phase

On May 19, 2026, Andrej Karpathy — OpenAI founding member, former Tesla AI lead, and the person who coined "vibe coding" — announced he's joining Anthropic to work on Claude pre-training and the autoresearch team. This post breaks down the timeline, his background, the strategic implications, and how developers can access the Claude API through claudeapi.com.

2026.05.215min