🚨 China's AI Heist • 🇦🇪 The Autonomous Government • 💸 OpenAI's $600B Problem
Plus: OpenAI's secret plan to build an AI-first smartphone.
🎵 Podcast
Don’t feel like reading? Listen to it instead.
📰 Latest News
This week’s image aesthetic (Flux 2 Pro): Rorschach test
The Great AI Heist: How Foreign Actors Are "Distilling" America's Frontier Models
The White House has issued an urgent warning that foreign entities, primarily based in China, are conducting “industrial-scale distillation” campaigns to illicitly copy advanced artificial intelligence systems from United States frontier labs. In a recent memo, the administration accused these actors of using tens of thousands of proxy accounts and jailbreaking techniques to send a massive volume of queries to large AI models, using the extracted data to train smaller knockoff versions. Major US firms such as OpenAI and Anthropic have already reported detecting this adversarial activity on their platforms. In response, the US administration plans to share critical threat information with domestic AI companies, help coordinate industry defences, and explore strict measures to hold the foreign actors accountable.
Why it Matters
This systemic intellectual property theft poses a severe threat to American technological dominance. By relying on illicit distillation, foreign entities can create highly capable AI products that perform comparably on key benchmarks for a fraction of the original cost, actively eroding the competitive advantage of US firms that have invested heavily in research and development. More alarmingly, these distillation campaigns allow adversaries to intentionally strip away critical safety and security protocols embedded in the original models. The resulting unsecured systems could then be weaponised for malicious purposes, including offensive cyber operations and disinformation campaigns, drastically intensifying the technological rivalry between the US and China.
No More Apps: Why OpenAI is Teaming Up with Qualcomm for a 2028 Smartphone
According to prominent industry analyst Ming-Chi Kuo, OpenAI is actively developing a revolutionary new smartphone centred entirely on AI agents rather than traditional applications. This ambitious device aims to replace the familiar app grid with a task-based interface, where users simply interact with an AI to seamlessly complete objectives like booking services or sending messages. To build this hardware, OpenAI is reportedly collaborating with major chipmakers Qualcomm and MediaTek to engineer custom mobile processors, while tapping Luxshare as the exclusive manufacturing partner. The phone will rely on a sophisticated hybrid of on-device and cloud-based AI, designed to continuously monitor and understand the user’s real-time context. Mass production for the handset is reportedly targeted for 2028, with core specifications and vendor agreements expected to be finalised by early 2027.
Why it Matters
This massive development signals a profound shift in how we interact with mobile devices, moving away from a cluttered, app-centric model to a proactive, task-oriented ecosystem. By completely controlling both the physical hardware and the operating system, OpenAI can create a deeply integrated AI experience that is simply impossible to achieve through standalone software on existing platforms. This aggressive approach could fundamentally alter consumer behaviour, as intelligent agents take over complex, multi-step tasks in the background and eliminate the need to manually navigate between different applications. Ultimately, this project positions OpenAI not just as a software provider, but as a direct, formidable competitor to established smartphone giants.
The Autonomous State: UAE to Hand 50% of Government Services to AI Agents
The United Arab Emirates is aggressively pursuing a massive overhaul of its public sector, announcing plans to transition 50% of its federal government services to agentic artificial intelligence within just two years. Unlike basic digital chatbots, agentic AI is capable of independently executing complex workflows from start to finish without human intervention. For residents, this means notoriously bureaucratic processes, such as visa renewals or utility connections, could soon be completed autonomously with a single request, as the AI manages all background checks, data cross-referencing, and approvals. To support this massive operational shift, the UAE has mandated specialised AI training for all federal government employees to build the necessary oversight skills. In parallel, Dubai has launched a specific, targeted programme to upskill 50,000 of its own public sector workers for the new technological reality.
Why it Matters
This initiative represents a monumental shift in public administration: moving from a “digital government,” where citizens are still burdened with filling out online forms, to an “autonomous government,” where the system proactively completes the service for them. By effectively deploying AI as an executive partner, the UAE aims to drastically reduce the administrative burden on residents and significantly cut operational costs. Furthermore, the mandatory, large-scale upskilling of the federal workforce is a direct and necessary attempt to safely integrate human oversight with highly automated decision-making. Ultimately, this two-year sprint establishes a radical new benchmark for public service delivery globally, proving that the future of governance relies on fully autonomous systems rather than just the mere availability of online portals.
Musk Takes the Stand: The $134 Billion Trial to Dismantle OpenAI
Elon Musk has officially taken the stand in a highly anticipated courtroom showdown against OpenAI and its CEO, Sam Altman. The trial, which kicked off with jury selection in late April 2026 at a federal courthouse in Oakland, California, centres on allegations that OpenAI completely betrayed its founding mission. Musk claims that the organisation, initially established as a non-profit to develop artificial intelligence for the benefit of humanity, was improperly transformed into a massive for-profit enterprise. Accusing the leadership of a breach of charitable trust and unjust enrichment, Musk is demanding the immediate removal of Altman and President Greg Brockman. Furthermore, he is seeking to force the company to return to its original non-profit structure, with up to $134 billion in potential damages to be redistributed to OpenAI’s charitable foundation.
Why it Matters
The outcome of this trial carries potentially catastrophic stakes for one of the world’s most valuable tech companies. A verdict in Musk’s favour could force a massive restructuring of OpenAI, severely impacting its multi-billion dollar partnership with Microsoft and derailing its expected 2026 initial public offering. Beyond the immediate financial fallout, the case raises fundamental questions about the governance of powerful AI technologies and the legal responsibilities of organisations founded with altruistic aims that later pivot to lucrative commercial models. Ultimately, this legal battle exposes the extreme tension between the stated goals of developing AI for the good of humanity and the brutal, profit-driven pressures of Silicon Valley.
The $600 Billion Problem: OpenAI Misses Targets and Sparks Internal Civil War
OpenAI’s seemingly invincible growth narrative has suddenly hit a wall. According to a bombshell report, the company missed several crucial internal revenue targets in early 2026, alongside a massive failure to reach its goal of one billion weekly active users for ChatGPT by the end of 2025. With user growth plateauing at around 900 million and rivals like Anthropic and Google Gemini aggressively eating into its enterprise and consumer market share, internal tensions are reportedly boiling over. Chief Financial Officer Sarah Friar has explicitly warned colleagues that if revenue does not rapidly accelerate, OpenAI may struggle to pay for its staggering $600 billion in future data centre computing contracts. Furthermore, Friar has reportedly clashed with CEO Sam Altman over the timing of a potential initial public offering; while Altman is pushing for an aggressive Q4 2026 listing, Friar cautions that the company is simply not organisationally ready. Despite Altman and Friar issuing a joint statement calling the division “ridiculous” and claiming the business is “firing on all cylinders”, the market reaction was brutal.
Why it Matters
This massive stumble proves that even the leader of the AI boom is not immune to harsh financial realities. The missed targets expose a terrifying gap between OpenAI’s astronomical infrastructure commitments and its actual revenue trajectory, raising severe questions about the company’s long-term financial sustainability. Furthermore, the landscape has fundamentally shifted: with competitors like Anthropic reportedly overtaking OpenAI in annualised revenue while spending a fraction on training, the generative AI market is no longer a one-horse race. The fallout from these internal struggles is already sending shockwaves through the broader tech ecosystem. Investors immediately wiped tens of billions of dollars off the valuations of key OpenAI infrastructure partners, sending shares of Oracle, CoreWeave, and SoftBank plummeting. Ultimately, this crisis signals that the blind “buy everything” era of AI infrastructure is over, replaced by a harsh new reality where massive capital expenditures must be justified by actual, dominant market performance.
🔗 More from the Wall Street Journal
Last week’s newsletter: