CONNECT WITH US
Jun 25
Europe's AI infrastructure: the cost gap that policy cannot paper over
The EU currently hosts roughly 5% of the world's AI compute capacity. The US holds close to 75%. McKinsey projects European data center demand will grow from 10 GW of IT load in 2024 to 35 GW by 2030 — a tripling driven almost entirely by AI. The infrastructure to close that gap does not yet exist, and building it in Europe costs substantially more than building the equivalent in the US.
IEI Integration said its first-quarter margin was squeezed by higher component costs and a changing product mix, but it expects a second-quarter rebound as high-end video conferencing shipments ramp up. The outlook matters for global industrial and enterprise technology buyers facing tighter supply, shifting production, and uneven demand across markets.

Acer channel arm Weblink International remains upbeat on the second half of 2026, even as the PC industry looks increasingly likely to see weaker demand than in the first half.

Qualcomm has unveiled its latest AI data center platform, Dragonfly, at its annual investor day, highlighting a new technology it calls High Bandwidth Compute, or HBC, as a key weapon in its challenge to Nvidia, AMD and AI chip startups.

Chinese AI startup Z.ai, also known as Zhipu AI, has turned GLM-5.2 into the clearest test yet of China's frontier AI gap with Anthropic and OpenAI. Its stock surge is now raising a second question: whether investors are pricing a real technology shift or a thin-float capital-market fever.
Global demand for intelligent eyewear is rising fast, with shipments jumping 83% in the first quarter of 2026, according to Counterpoint Research. For readers worldwide, the shift signals a faster move toward AI-enabled wearables, while VR headsets continue to struggle with weak demand and slower product refreshes.
AI-driven hyperscale data center expansion is pushing the global AI race beyond raw GPU computing power into a broader contest over high-speed interconnects and electro-optical integration. As Nvidia's next-generation AI factory moves toward petabyte (PB)-scale data transfers, co-packaged optics (CPO) is heading toward commercialization, lifting the importance of key components such as fiber array units (FAU) and external laser sources (ELS).
OpenAI CEO Sam Altman told staff that its latest model, GPT 5.6, would be released on a staggered basis, with a small group of entities first gaining preview access to it after approval by the US government. The case highlights the regulatory uncertainty many local AI developers are facing after US President Donald Trump announced a voluntary review system before the release of frontier models, with the industry remaining concerned that it is not so voluntary after all.
Academia Sinica has built a 20-qubit quantum chip from scratch, and the team now says the next phase is less about lab breakthroughs and more about engineering a scalable, reproducible manufacturing system. Chii-Dong Chen, executive director of Academia Sinica's Center for Quantum Computer, says Taiwan's quantum chip effort is moving into that engineering battleground.
AI companies are racing to hire forward deployed engineers, or FDEs, as the industry shifts from model-building to implementation. For global readers, the trend signals that the next wave of AI competition may hinge less on smarter systems and more on whether businesses can actually use them effectively at scale.
Apple raised prices across its Mac and iPad lineup on June 25, and the timing wasn't a choice so much as a deadline.
Apple is preparing its biggest shift yet to the release strategy for its in-house Mac processors, opting to skip high-end M6 Pro and M6 Max chips and instead bring more powerful AI-focused M7 Pro and M7 Max processors to market in 2027, according to a Bloomberg report citing people familiar with the matter.