Meta Needs To Rethink Manus

Apr 27, 2026 19m

AI Summary Available

Get the full experience! Sign up to access transcripts, personalized summaries, and more features.

Episode Description

China blocked Meta's $2B Manus acquisition and ordered both sides to unwind the deal, closing the "Singapore washing" loophole for Chinese AI startups. OpenAI is developing smartphone chips with Qualcomm and MediaTek, Google controls ~25% of global AI compute, and SaaS pricing shifts to usage-based. China blocks Meta's $2B Manus acquisition, after reviewing whether it violated investment rules, and tells both to cancel it; Manus moved to Singapore in 2025 (FT) Kuo: OpenAI is working with MediaTek and Qualcomm to develop smartphone chips, with Luxshare handling the system co-design; mass production is expected in 2028 (Ming-Chi Kuo) Epoch AI: Google controls ~25% of global AI compute, with ~3.8M TPUs and 1.3M GPUs; Google Cloud CEO Thomas Kurian says demand and revenue justify the spend (FT) Analysis: as of late 2025, 79 of 500 tracked software companies including HubSpot, Adobe, and Salesforce adopted usage-based AI fees, more than doubling on 2024 (The Information) Anthropic details Project Deal, a marketplace experiment where Claude models bought, sold, and negotiated personal belongings on behalf of Anthropic employees (Anthropic) Learn more about your ad choices. Visit megaphone.fm/adchoices

AI-Generated Summary

China Blocks Meta's Manus Acquisition, Shutting Down the Singapore Washing Loophole

China has blocked Meta's planned $2 billion acquisition of Manus, citing concerns over foreign investment rules and technology export regulations. Despite Manus relocating to Singapore in 2025 to access Western capital, Chinese authorities have decided to prohibit Chinese companies from being involved in the transaction and instructed both Meta and Manus to unwind the deal. This action signals China's resolve to tighten control over foreign AI investments and demonstrates that offshore jurisdictions like Singapore are no longer beyond China's regulatory reach. The move could significantly impact Chinese AI startups' ability to access global funding and could influence the strategic choices of Chinese tech founders, emphasizing a move away from offshore entities for fear of regulatory risk.

OpenAI Collaborates with Qualcomm and MediaTek on Smartphone Chips for 2028 Release

OpenAI is working with MediaTek and Qualcomm to develop AI-optimized smartphone chips, with mass production anticipated in 2028. The collaboration involves Luxshare as the exclusive system co-design and manufacturing partner. This initiative aims to integrate advanced AI capabilities directly into future smartphones, focusing on faster AI performance rather than raw power. OpenAI's involvement suggests a strategic move to control hardware for AI agent ecosystems, potentially offering subscription services bundled with devices. This development aligns with broader trends of AI integration into consumer hardware and signifies a shift towards AI-first smartphones.

Google Commands a Quarter of Global AI Compute Power, Invests in Proprietary Chips

Google currently controls approximately 25% of the world's AI compute power, with around 3.8 million TPUs and 1.3 million GPUs, making it a dominant player in AI infrastructure. Google's CEO Thomas Kurian highlighted the company's shift towards in-house full-stack AI solutions, including chips, data centers, and models, which has paid off with strong revenue growth. Google's strategy emphasizes reducing dependence on third-party hardware providers like Nvidia, aiming for cost efficiency and higher margins. This move positions Google as a leader in AI hardware, competing with other tech giants and shaping the future landscape of AI computing.

Shift Towards Usage-Based SaaS Pricing and AI-Enabled Enterprise Software

An increasing number of enterprise software companies are shifting from flat per-user subscription models to usage-based pricing for AI features. By the end of 2025, nearly 80 out of 500 tracked companies, including major players like HubSpot, Adobe, and Salesforce, adopted this model amid rising AI adoption and costs. This shift reflects efforts to better align pricing with actual AI usage but introduces risks of escalating costs for customers. Companies like Uber have already experienced significant budget overruns. The trend indicates an industry move towards monetizing AI consumption more granularly, which may impact customer satisfaction and vendor strategies.

Anthropic's AI Marketplace Experiment Demonstrates AI Negotiation Power and Risks

Anthropic conducted a week-long experiment called Project Deal, where AI agents using Claude models bought and sold personal items autonomously via Slack. The results showed that the more advanced Opus 4.5 agents closed more deals and obtained higher prices compared to smaller Haiku 4.5 agents. Interestingly, participants rated deals as fair regardless of AI agent strength, raising concerns about market fairness and transparency. The experiment highlights the potential of AI-powered commerce but also points to risks such as unknowing disadvantages for users and increased security issues like prompt injections. It underscores that AI agents could soon play significant roles in autonomous market transactions with complex societal implications.