The tariff war between the United States and China has entered a new phase of calculated reversals and strategic brinksmanship. After imposing sweeping duties across trading partners, Donald Trump paused tariffs for 90 days—a move that signals less ideological certainty than pragmatic response to market signals. When bond yields rose instead of falling, suggesting the market had read through his stated rationale, Trump pivoted. The consumer electronics exemption, including iPhones, came after pressure from Apple's Tim Cook; the company faces a $30 billion bill to relocate just 10 percent of its supply chain domestically over three years, making the political cost of price shocks too high.
China, meanwhile, has adopted a different playbook. Rather than negotiate for tariff relief, it escalated. With 15 percent of its exports flowing to the United States, the country faces genuine pressure—but it has asymmetric weapons. Rare earths tell the story: China controls 90 percent of global rare-earth processing, materials essential to semiconductors, defence systems, and pharmaceuticals. By banning their export or imposing restrictions through triangulation detection, China can weaponize the very industrial autonomy Trump seeks. It is also playing a longer game of attrition, betting that American inflation and voter dissatisfaction with rising costs will break faster than Chinese factory closures and unemployment.
The episode also touched on credible concerns about insider trading tied to tariff announcements. Options traders made 2,000 percent returns on same-day bets that the S&P 500 would rise following Trump's pause announcement—trades placed before public disclosure. The SEC and Department of Justice, both under Trump's influence, face obvious incentive conflicts in investigating. If true, it represents precisely the kind of market manipulation that free-market capitalism is supposed to prevent.
On the technology front, OpenAI released GPT 4.1, a model with similar capability to the delayed GPT 4.5 but at lower cost, featuring a 1 million token context window—eight times larger than GPT 4O. The company also hinted at building a Twitter-like social network, reportedly drawing on Sora's image generation. Whether a new social platform centred on AI-generated content can shift user behaviour remains doubtful; network effects favour incumbents, and Meta and Elon Musk's X already embed AI features. The more immediate question is whether product innovation alone can overcome inertia.
Finally, Meta faces antitrust hearings over its $1 billion Instagram and $19 billion WhatsApp acquisitions. Forced divestiture remains possible, though regulators increasingly favour preventive merger-blocking over post-facto breakups. The contrasts are telling: Instagram and WhatsApp became trillion-dollar assets; Tumblr, acquired by Yahoo for $1.2 billion a year earlier, sold for $3 million. Strategic M&A, not regulation, has always been the real arbiter of value.