Presented by
Good morning, AI enthusiasts. Amazon just made the biggest compute bet in AI history, locking in $100B to power Anthropic's next generation of models.
With 5GW of capacity and Trainium chips at the core, this deal isn't just about scale. It's a direct challenge to Nvidia's grip on AI infrastructure.
In today's recap:
Anthropic and Amazon lock in $100B compute deal
Tim Cook steps down, John Ternus becomes Apple CEO
Set up Codex Chronicle to remember your coding context
NSA caught using Anthropic AI on the federal blacklist
4 new AI tools, prompts, and more
ANTHROPIC & AMAZON
Anthropic locks in landmark $100B Amazon compute deal
Amazon
Recaply: Anthropic just sealed a massive expansion of its Amazon partnership, adding 5GW of compute and a $100B investment commitment to back its AI growth.
Key details:
Amazon will supply compute through Trainium2 and Trainium3 chips. Claude will run on AWS for enterprise customers and consumer apps.
Amazon has invested $5B in Anthropic so far, with $20B more pledged. Anthropic's annual revenue target rises from $9B to $30B.
The deal moves Anthropic's chip strategy away from Nvidia. Trainium is Amazon's own chip, not the standard industry choice.
The first gigawatt of capacity arrives by end of 2026. The full 5GW rollout continues over the years ahead.
Why it matters: Amazon's $100B commitment is the biggest single compute deal in AI history. The more interesting story is the chip strategy. By choosing Trainium over Nvidia's GPUs, Amazon and Anthropic are building a parallel AI stack. If it works, Amazon can shape its own AI roadmap without depending on Nvidia's supply chain.
PRESENTED BY NEO
AI help, without the trust tax.
Most AI tools ask you to trade your data for intelligence. Norton Neo doesn't. It's the first safe AI-native browser built by Norton, and it gives you powerful built-in AI without handing your privacy over to get it. Search, summarize, and write with AI built directly into your browser. Your data stays yours. Your context stays private.
Built-in VPN, anti-fingerprinting, and ad blocking come standard. No add-ons. No setup. No compromises.
Fast. Safe. Intelligent. That's Neo.
APPLE
Tim Cook steps down, hardware chief takes over
Apple
Recaply: Apple officially confirmed Tim Cook's departure as CEO, with hardware lead John Ternus stepping in to run the company as Apple pushes deeper into chips and on-device AI.
Key details:
Cook will move to Executive Chairman and Ternus becomes CEO. Johny Srouji, who leads Apple's chip design team, becomes Chief Hardware Officer.
Under Cook, Apple's market cap grew from $350B to over $4T. Annual revenue jumped from $108B to $416B.
Ternus led the M-series chip program, Apple's shift to custom silicon. His rise signals Apple's next chapter is built around hardware and on-device AI.
The transition takes effect September 1, 2026. Ternus has been with Apple since 2001.
Why it matters: Tim Cook built Apple into the most valuable company in history. But the next era is a hardware story. Ternus led Apple's move to M-series chips, one of the biggest strategic shifts in the company's history. Under his leadership, Apple's on-device AI push is likely to become the central product strategy. That means less dependence on cloud AI, and more power staying on your device.
GUIDES
Set up Codex Chronicle to remember your coding context
Recaply: In this tutorial, you will learn how to set up Codex Chronicle, OpenAI's new memory tool for its coding agent. Once active, it remembers your tech stack and past decisions across every session.
Step-by-step:
Open Codex on macOS. Go to Settings, then Memory, and turn Chronicle on. This feature is in research preview for ChatGPT Pro users on macOS only.
Start a new Codex session. Describe your project: your coding language, framework, and testing library. Codex will save an initial memory based on what you share.
As you work, Codex saves your decisions automatically. It tracks file structures, naming patterns, and commands you use often. Check your stored memories in the Memory panel at any time.
Before each new task, open the Memory panel to review saved context. Codex loads this into the session so you don't have to repeat yourself each time.
Clean up your memory store by removing outdated entries from the Memory panel. Add rules like "always use TypeScript strict mode" and Codex will follow them on every task.
Pro tip: Connect Codex to your GitHub repository so Chronicle can store branch-level context. When you switch branches, Codex knows which decisions apply to each feature or fix.
TOOLS
Trending AI Tools
🤖 Kimi K2.6 - Moonshot AI's new open-source reasoning model with SOTA coding benchmarks and support for 300 parallel agent runs
🤖 Qwen 3.6-Max-Preview - Alibaba's new preview model with stronger agentic coding and improved instruction following
🎨 Claude Design - Anthropic's AI tool for building branded prototypes, slides, and visual assets
🎨 Canva AI 2.0 - Canva's AI-native creative suite with 6 new design workflows and a custom design model in research preview
NEWS
What Matters in AI Right Now?
Moonshot AI released Kimi K2.6, its latest open-source model with strong coding and reasoning benchmarks, including a top score on SWE-Bench Pro and support for 300 parallel agent runs.
Adobe launched CX Enterprise, an AI agent suite for marketing teams, with support for MCP and A2A protocols to connect into tools used by 20,000+ brands.
Google is developing new inference chips to speed up AI results and reduce reliance on Nvidia, with some rivals already buying Google's TPUs as an alternative to Nvidia's hardware.
Google rolling out Gemini integration in Chrome to Asia Pacific countries, extending the AI browser assistant to new markets across the region.
Lovable, the AI app builder, exposed user data through a security flaw, with researchers finding that the issue allowed access to sensitive information across accounts.
Alibaba released Qwen 3.6-Max-Preview, a new model with improved agentic coding, stronger world knowledge, and better instruction following than earlier Qwen 3 models.
OpenAI released a preview of memories in Codex, called Chronicle, which lets the coding agent remember your stack and past decisions across sessions, currently available for ChatGPT Pro users on macOS only.
The NSA is using Anthropic's Mythos model despite it appearing on the US federal blacklist, with Anthropic restricting the model to roughly 40 organizations due to its offensive cyber capabilities.
Deezer revealed that 44% of songs uploaded to its platform daily are now AI-generated, with roughly 75,000 AI tracks added each day as AI music volumes continue to grow.
OpenAI ad partner StackAdapt is selling ChatGPT ad placements based on "prompt relevance," with a leaked deck showing CPMs as low as $15 through the "OpenAI x StackAdapt Limited Pilot Program."
EVENTS
Notion Hackathon: May 16-17 • San Francisco, CA
Google Cloud Next 2026: April 22-24, 2026 • Las Vegas, NV
Anthropic Code with Claude: May 6, 2026 • San Francisco, CA
🧡 Enjoyed this issue?
🤝 Recommend our newsletter or leave a feedback.
How'd you like today's newsletter?
Cheers, Jason
