Microsoft's Majorana 2 Targets 2029; ARM CEO Flags China CPU Export Limits; UK Reins In Google AI

Three signals from the quantum computing race, AI chip export policy, and Google's regulatory front in the UK.

🔬 Microsoft's Majorana 2 Chip Delivers 1,000x Reliability Gain, Sets 2029 Quantum Target

Decoded: Microsoft CEO Satya Nadella unveiled the Majorana 2 quantum processor at Build 2026 on June 2 — the successor to Majorana 1 announced in 2025. The new chip's topological qubits are 1,000x more reliable than the prior generation and maintain quantum state substantially longer, according to Microsoft's official Build 2026 live blog and reporting by The Verge and Reuters. Microsoft set a stated goal of a working, commercially useful quantum system by 2029. (Microsoft Build 2026 Blog, The Verge, Reuters, June 2–3, 2026)

Why it matters: Qubit reliability is the decisive barrier in quantum computing — unstable qubits generate errors faster than correction can address, making large computations unworkable. A 1,000x improvement is a qualitative step, not incremental, and if independent review confirms the topological qubit results, Microsoft (MSFT) would hold a technical lead over IBM and Google in the race to quantum utility. A working system by 2029 arrives at the height of the AI infrastructure cycle, where quantum-accelerated workloads could reshape drug discovery, financial modeling, and cryptography — and differentiate Azure in the cloud platform race.

🏛️ ARM CEO Says Banning AI CPU Exports to China Is Practically Unworkable

Decoded: Arm Holdings CEO Rene Haas told Reuters on June 2 that blocking AI CPU chip exports to China would be nearly impossible — because CPUs lack the discrete performance thresholds and memory bandwidth limits that make GPU export controls enforceable. Haas said CPUs are broadly used across applications and described them as "like oil relative to the application space." The comments came as Senators Warren and others called on the Trump administration to close loopholes allowing AI chips to reach overseas affiliates of Chinese firms. (Reuters, June 2, 2026)

Why it matters: The Haas statement is a direct intervention in the export control policy debate at a critical juncture. GPU-based export controls on Nvidia chips have a defined enforcement mechanism; CPU-based controls do not, and ARM (ARM) architecture CPUs power the vast majority of mobile and data center devices entering China's AI build-out. If Haas is correct, attempts to extend chip export controls from GPUs to CPUs would face significant implementation problems — which benefits ARM's China licensing revenue and undercuts the policy rationale for broader compute restrictions. The political dynamic creates uncertainty: if Congress pushes CPU controls regardless, ARM faces both revenue risk and operational complexity in its China licensing agreements.

🏛️ UK CMA Orders Google to Give Publishers Opt-Out From AI Search Content Use

Decoded: Britain's Competition and Markets Authority imposed new conduct requirements on Google's search services on June 3, mandating that publishers be permitted to stop their content from powering Google's AI features — including AI Overviews — without losing visibility in standard search rankings. The CMA said the rules give publishers stronger bargaining power; Google holds more than 90% of UK search queries. News publishers have seen click-through rates fall as users rely on AI-generated overviews instead of visiting source pages. (Reuters, June 3, 2026)

Why it matters: The CMA ruling is the first binding regulatory enforcement action globally that restricts how a dominant search engine can use publisher content for AI-generated answers. If publishers exercise the opt-out at scale — the financial incentive is strong as AI Overviews continue to suppress referral traffic — Google's AI search quality degrades in the UK, and the regulatory template expands: the EU and U.S. FTC are both monitoring AI search content practices. For Alphabet (GOOGL) investors, mandatory opt-out rights across multiple markets represent a compounding structural risk to the AI Overviews product that Google is counting on to defend search market share against new AI-native competitors.

Stay decoded. See you tomorrow.

— The Get AI Decoded Team