Intel is developing a new memory architecture aimed at challenging the dominance of high-bandwidth memory (HBM), with commercialization targeted for around 2030. Although the path is fraught with ecosystem barriers and compatibility hurdles, Intel's parallel development of Z-angle memory (ZAM) and cross-batch memory (XBM) underscores its determination to re-enter the DRAM market, as it simultaneously bets on AI compute and storage.
AI computing demand continues to fuel growth in the global memory market, but the industry's attention is shifting beyond short-term price movements. Increasingly, the focus is on longer-term variables, including the pace of capacity expansion, the sustainability of AI-driven demand, and whether emerging AI applications can achieve commercial scale.
Reports that Meta is considering leasing out idle AI computing capacity have rattled investors. But treating Meta's predicament as a warning sign for the entire AI industry is a classic case of overgeneralization.
The US can no longer close its artificial-intelligence talent gap with China through visa curbs or export controls alone, a new Hoover Institution and Stanford study argues, because China is now producing frontier-model researchers who never trained, worked, or published abroad, even as it also reclaims talent that spent years in American institutions.
The US National Science Foundation will prohibit the researchers it funds from collaborating with organizations on Washington's restricted-party lists, a roster heavily populated by Chinese firms and institutions, under a Dear Colleague Letter dated July 8, 2026. The agency said it intends to implement the prohibition in fiscal 2027.
At the Humanoids Summit in Tokyo at the end of May 2026, a conference that had originally focused on technology and commercialization also set aside a stage for government delegates and policy watchers. An executive at a major US robotics company said bluntly at the event: "Government intervention is no longer optional."
As the world enters an AI-centric era, the global race for technological leadership is no longer defined only by who can build the most advanced models. It is increasingly shaped by who can secure compute, deploy infrastructure at scale, reduce energy constraints, and turn research into commercial capability.
Reports have emerged that Apple may have managed to avoid 100% tariffs imposed by US President Donald Trump, partly by agreeing to partner with Intel to manufacture its chips. While Apple could benefit from expanding its chip suppliers, the episode also shows the power of Intel's government backing as the US seeks to reshore its semiconductor industry.
Tencent is in talks to become the largest shareholder in Manus, leading a proposed buyback of the AI agent startup after Chinese regulators ordered Meta Platforms to unwind its US$2 billion acquisition.
AI demand is still outrunning advanced semiconductor capacity, putting foundry output, HBM supply, packaging and server infrastructure at the centre of this week's tech agenda.
China is preparing to allow a limited number of Nvidia H200 AI accelerators into the country, giving Alibaba, ByteDance, and DeepSeek access to advanced computing power while preserving Beijing's broader campaign for semiconductor self-reliance.
Foxconn is using AI and coordinated security teams across Asia, Europe and the Americas to defend against cyberattacks as enterprises increasingly adopt AI. Chief Information Security Officer Wei-Bin Lee said that the company's global footprint makes it a prime target.
