Alipay Lets AI Agents Spend Your Money Autonomously

Alipay Just Gave AI Agents a Credit Card — And It’s a Big Deal

Why it matters: AI agents can now complete purchases on your behalf — no human confirmation required. Alipay’s new AI Pay service lets OpenClaw-style autonomous agents initiate and complete real financial transactions based on user instructions. This is the missing link between “AI assistant” and “AI employee with a budget.”

The service is designed specifically for agentic workflows — think an AI that books your travel, orders supplies, renews SaaS subscriptions, or pays vendor invoices without you lifting a finger. Alipay is framing it as a trust layer: users set spending limits and categories; the agent operates within those guardrails.

For business owners: This is the moment agentic commerce becomes real. The same way we gave employees corporate cards with spending limits, we’re now giving AI agents payment authority. The productivity unlock is enormous — but so is the governance question. What policies does your business need before your AI agent starts spending?

⚡ Quick Hit

$40M Bet on AI Agents That Actually Learn on the Job

NeoCognition just closed a $40M seed round to solve AI’s dirty secret: today’s agents only succeed about 50% of the time. The Ohio State spinout, backed by Cambium Capital, Walden Catalyst, Vista Equity, and Intel CEO Lip-Bu Tan, is building agents that self-specialize — learning a domain the way a new hire learns a job, without custom engineering for every vertical. Their pitch: stop deploying generalist agents that require a leap of faith on every task.

Business takeaway: If your AI agents are failing half the time, it’s not just you. The industry is racing to fix reliability — and the money is now serious.

📰 Read the full story →

⚡ Quick Hit

No Code, No Problem: Anyone Can Build an AI Agent Now

No-code platforms like Toolhouse are eliminating the “you need a developer” barrier entirely. Voice commands, drag-and-drop workflows, and pre-built templates now let non-technical business owners deploy agents for email triage, customer support, research, and document management — in minutes, not months. Integrations with Gmail, Zapier, and Pipedream mean these agents plug into tools you already use.

Business takeaway: The “we don’t have an IT team” excuse is expiring fast. If your competitor’s office manager can now deploy an AI agent before lunch, the playing field just shifted.

📰 Read the full story →

🔴 Security Alert

Anthropic’s Most Powerful AI Was Leaked Before It Even Launched

Anthropic’s new Claude Mythos — a cybersecurity-focused model that found 271 vulnerabilities in Firefox alone and “thousands” of flaws across every major OS and browser — was accessed by unauthorized users via a private online forum on April 21, the same day Anthropic announced its limited testing release. Access reportedly came through a third-party vendor. The model was being distributed through a tightly controlled program called Project Glasswing, with access granted only to Amazon, Microsoft, Nvidia, Apple, and ~40 critical infrastructure organizations.

Central banks in Australia and New Zealand are now actively monitoring the situation. Germany’s Bundesbank called Mythos a “double-edged sword” — brilliant at finding vulnerabilities, but equally capable of exploiting them. OpenAI responded by widening access to its own cybersecurity model, GPT-5.4-Cyber.

Business takeaway: The most advanced AI is now in a cyberattack arms race. If you haven’t reviewed your software and vendor security posture recently, this is your reminder.

📰 Read the full story →

💰 Big Money Move

SpaceX Locks In $60B Option to Buy Cursor. AI Coding Is Now a $50B Business.

SpaceX announced it has secured the right to acquire AI coding platform Cursor for $60 billion — nearly double its $29.3B valuation from just six months ago. The deal pairs Cursor’s 1 million+ daily users (including 50%+ of Fortune 500 companies) with SpaceX’s Colossus supercomputer. Cursor’s annualized revenue already tops $2B and is projected to hit $6B by year-end. Separately, Cursor is closing a $2B funding round at a $50B valuation led by Andreessen Horowitz and Thrive.

Business takeaway: AI coding tools aren’t a niche dev perk anymore — they’re the fastest-growing enterprise software category on the planet. If your team isn’t using one, your competitors’ teams are writing code twice as fast.

📰 Read the full story →

🏗️ Infrastructure

Anthropic Commits $100 Billion to AWS — A Decade-Long Cloud Bet

Anthropic has signed a 10-year, $100+ billion commitment to run its Claude models on Amazon Web Services — one of the largest cloud infrastructure deals ever announced. This locks AWS in as Anthropic’s primary compute backbone for training and inference, deepening a partnership that includes Amazon’s $4B investment in the company. The deal signals that the AI infrastructure war isn’t just being fought in data centers — it’s being won in long-term contracts.

Business takeaway: Anthropic’s Claude isn’t going anywhere — and it’s going to get faster and cheaper as AWS scale kicks in. If you’re building on Claude via AWS Bedrock, your foundation just got more stable.

📰 Read the full story →

Quick Plug

Still working out where AI fits for you? The Lab is our free community — what's changing, what's worth your time, and how people are actually running agents day to day.

aitokenlabs.com/the-lab

This newsletter? Written by an AI Employee, approved by a human — so our team stays focused on what only humans can do.