Samsung Electronics' decision to place a key image sensor production line under its System LSI division highlights a broader question facing the CMOS image sensor (CIS) industry: whether chip designers should own manufacturing capacity or rely on external foundries as competition intensifies across smartphones, automotive systems, and AI-enabled devices.

Nvidia CEO Jensen Huang introduced RTX Spark, co-developed with MediaTek, at Nvidia GTC Taipei, with PC brands expected to launch products in the third quarter of 2026. While widely seen as Nvidia's return to Windows on Arm after a 15-year absence and a challenge to Qualcomm, RTX Spark points to a larger fight over AI-era endpoints.

In a move set to reshape the global robotics landscape, Nvidia announced at GTC Taipei that it will partner with Chinese humanoid robot pioneer Unitree Robotics to launch the world's first open humanoid robot reference design. By deeply integrating advanced AI with physical hardware, the collaboration aims to drastically lower development barriers across the industry.

At the earnings call on June 3, Broadcom said its AI semiconductor business continued to accelerate in the fiscal second quarter, with revenue reaching a record US$10.8 billion, up 143% from a year earlier, and bookings exceeding US$30 billion. CEO Hock Tan said demand for custom accelerators and networking remained "simply insatiable," adding that the company expects AI semiconductor revenue to double in the second half of fiscal 2026 from the first half.

Broadcom said its AI semiconductor business remained in an "insatiable" demand environment on its second quarter of fiscal 2026 earnings call on June 3, as it outlined progress with its six core customers, including Google, OpenAI, Meta, Anthropic, and Apollo-backed programs. Chief Executive Hock Tan said the company's AI semiconductor revenue reached a record US$10.8 billion in the quarter, up 143% year on year, and added that networking made up almost 40% of AI revenue. He also said bookings topped US$30 billion, while shipments totaled US$10.8 billion.

Broadcom reported record second-quarter revenue and raised expectations for the current quarter, but its outlook for artificial intelligence (AI) semiconductor sales fell short of elevated market expectations, prompting a sharp decline in the company's shares in after-hours trading.

Andrea Gallo, CEO of RISC-V International, used his appearance at the MIPS Forum during Computex 2026 to declare that RISC-V has completed its transition from academic project to industrial standard — and that its dominance in physical AI is no longer a future prediction but a present reality.

Huawei's recently proposed Tau Law has drawn attention across the semiconductor industry, with the company arguing that chip performance can be improved by reducing the internal signal transmission time constant, known as τ, through multi-layer optimization across devices, circuits, architectures, systems, and algorithms.

As robots and autonomous systems move from factory floors into hospitals, warehouses, and public spaces, NXP Semiconductors CEO Rafael Sotomayor used his Computex 2026 keynote to argue that the defining challenge of physical AI is not raw intelligence, but the ability to act in milliseconds without waiting for instructions from the cloud.

MediaTek denied a foreign report about a timeline for adopting Intel Corp.'s embedded multi-die interconnect bridge packaging at a recent investor event and said its supply-chain strategy remains focused on TSMC. The clarification came after a Goldman Sachs research note, and media coverage suggested MediaTek had set tape-out and mass-production dates for a project using Intel EMIB-T packaging.

Microsoft unveiled Majorana 2 on June 2, 2026, its latest quantum chip and the follow-up to Majorana 1, announced on February 19, 2025. The new chip continues Microsoft's bet on a distinctive approach to quantum computing, even as parts of the physics community continue to press for more independent evidence.

Acer and Taipei Computer Association (TCA) Chairman Jason Chen said Nvidia's AI PCs, designed specifically for agentic AI functions, point to a new usage model in the AI era and could create fresh demand in the PC market. He said the PC industry, which had been stuck in stagnation or decline for years following a pandemic-era boom due to a shift toward working and studying from home, now has a chance to rebound as AI shifts from training to inference.

Pegatron announced a new generation of AI infrastructure products and an AI Factory validation framework at Computex 2026, showcasing the Nvidia Vera Rubin platform, HGX Rubin NVL8, and RTX PRO servers integrated with the Nvidia DSX AI Factory reference design to support design, verification, and deployment from digital twin simulation to mass production. The firm said the portfolio aims to advance AI Factory capabilities for manufacturing customers by combining compute, cooling, and validation workflows.

Nvidia CEO Jensen Huang said the computing industry is entering an era of "agentic computing," in which data centers, personal computers, autonomous vehicles, humanoid robots, and satellite systems will share a common AI architecture. He framed the shift as a broad reworking of how future devices will operate.

Nvidia's new N1X processor, developed with MediaTek, signals a broader shift in PC computing as AI agents gain traction worldwide. The Arm-based chip could boost supply choices, intensify competition with Intel and AMD, and reshape demand patterns across notebooks, servers, and consumer devices.

On the first day of Computex 2026 in Taipei, Marvell Technology CEO Matt Murphy made a pointed argument to an audience of industry insiders: the next major bottleneck in AI infrastructure is not compute or memory; it is connectivity.

Nvidia used GTC Taipei on June 1, 2026, to unveil RTX Spark, also known as N1X, a new AI PC system-on-chip designed for native AI agent workloads rather than mainstream Windows PCs. The chip appears designed to fill a gap in consumer hardware that cannot reliably handle local, autonomous AI tasks.

Intel is expanding its data center portfolio with new Xeon 6+ processors, Ethernet E835 networking products, and fresh details on its Crescent Island AI accelerator, positioning the CPU as a central control point for the next phase of AI infrastructure.

When Marvell CEO Matt Murphy asked ASE Group CEO Tien Wu what makes Taiwan's technology ecosystem unique — and why it is so hard to replicate elsewhere — the answer was less about technology than about time.

Nvidia CEO Jensen Huang made a surprise guest appearance during Marvell Technology CEO Matt Murphy's keynote at Computex 2026 in Taipei on June 2, joining Murphy on stage for about 10 minutes in what became the most-talked-about moment of the one-hour presentation.

Nvidia and MediaTek have formally entered the AI PC and Windows on Arm market with the unveiling of RTX Spark at Computex 2026, ending two years of low-profile development. The first products are expected from Dell, HP, Lenovo, and MSI in autumn 2026.

Nvidia CEO Jensen Huang held a global media session on June 2 with an unusual guest: MediaTek vice chairman and CEO Rick Tsai, who joined Huang to share the inside story of how the two companies developed RTX Spark, the AI PC platform that Huang described as the beginning of the agent computing era.

Nvidia CEO Jensen Huang announced at GTC Taipei on June 1st that the company is launching Vera, a new CPU designed specifically for agentic AI workloads. The announcement reframes the CPU market around a simple but consequential argument: every CPU built until now was designed for human users. Vera is designed for AI agents.