NVIDIA reported Q2 FY2027 revenue of $96.2 billion, up 106% year-over-year and above the roughly $92 billion consensus, with Data Center revenue of $89.0 billion, up 117% and about 92% of total sales. Gross margin held at 75.0% and non-GAAP EPS of $2.22 beat the $2.10 estimate. Q3 guidance of $108.0 billion excludes any China Data Center compute revenue.
AI GPUs stay scarce and expensive because production runs through several constrained stages at once, not one single choke point. HBM supply is concentrated among SK Hynix, Samsung Electronics, and Micron; CoWoS advanced packaging capacity sits with TSMC (台積電); and vertical integration across chip, memory, packaging, and software gives suppliers pricing power. Any one tight stage caps shipments regardless of raw chip fabrication capacity.
EUV lithography prints circuit patterns using 13.5-nanometer light, the shortest wavelength in production chipmaking today. Because that wavelength is absorbed by almost every material and requires a vacuum, mirror-based optics, and a laser-blasted tin-droplet plasma light source, ASML remains the only company that can build these machines, making it the single chokepoint for advanced nodes and a direct target of geopolitical export controls.
A chiplet is a smaller functional block manufactured on its own and then joined with other chiplets through advanced packaging into one complete chip. As monolithic scaling becomes harder and yields fall, chiplets let designers raise yield and mix process nodes — using the most advanced node only where it matters and mature nodes elsewhere — making heterogeneous integration one of the main paths for extending performance gains now that transistor shrinking is nearing its physical limits.
AI data centers draw on three chip camps: NVIDIA's general-purpose GPUs, cloud giants' self-designed ASICs such as Google's TPU and AWS's Trainium/Inferentia, and startup inference chips from Groq, Cerebras, and SambaNova. NVIDIA leads on its CUDA ecosystem and versatility, while cloud providers build their own silicon to cut costs and reduce dependence on outside suppliers.
Advanced chips are now treated as strategic assets underpinning AI, defense, and communications, pushing governments to fund domestic capacity. Seven major economies have launched subsidy programs, but their headline totals use different accounting bases—authorized ceilings, mobilization targets, and registered capital—so the numbers are not directly comparable.
CoWoS is TSMC's 2.5D packaging technology that places GPUs and HBM memory on a shared silicon interposer. It has become the real limit on AI chip supply because interposer yield and process complexity, not wafer fabrication, set the ceiling on how many accelerators can ship.
The United States restricts advanced AI chip sales to China through export controls enforced by the Bureau of Industry and Security (BIS) under the Export Administration Regulations (EAR), aiming to slow China's AI and military computing capacity. Controls are set by compute-performance thresholds, cover specific chips from NVIDIA and AMD, and extend to equipment and named entities such as Huawei and SMIC.
Chip "process node" labels like 2nm and 3nm no longer describe a physical measurement — they mark a manufacturing generation defined by transistor density and efficiency. The real contest is over performance-power-area gains, a shift from FinFET to Gate-All-Around transistors, and dependence on a single EUV equipment supplier.
AI accelerator chips split into three types — GPUs, TPUs, and ASICs — each trading generality for efficiency differently. NVIDIA's GPUs lead on CUDA's flexibility, while TPUs and inference-focused ASICs like Groq's LPU and AWS Inferentia sacrifice versatility for efficiency, with HBM versus on-chip SRAM memory determining which model sizes and latency needs each chip serves best.
HBM solves AI's core bottleneck—the memory wall—by stacking DRAM on a wide silicon-interposer bus beside the GPU, lifting bandwidth from 256 GB/s (HBM2) to 2 TB/s (HBM4) and capacity from 8GB to 64GB, though supply stays concentrated in three makers led by SK Hynix's 58% share.
SpaceX plans to launch its first NVIDIA-chip-powered AI satellites, Starmind AI1, in the fourth quarter of 2027, reaching large-scale deployment by 2028 — at least a year ahead of the original schedule. The company has also filed with the FCC for a network of up to 1 million satellites, while Taiwanese suppliers Unitech and Sesoda report rising satellite-related revenue and order shares tied to the buildout.
NVIDIA announced the Jetson Orin Nano 2 robotics computer on August 25, 2026, packing 78 TOPS of AI compute, 8GB of memory, and an 8-core Arm CPU while doubling inference performance and cutting power draw 40% versus its predecessor. Modules and developer kits ship in the first half of 2027, targeting a robotics developer base NVIDIA says already tops 3 million.
NVIDIA confirmed on August 24, 2026 that its 256-chip Groq 3 LPX inference rack has entered full production, with Nebius named as the first cloud provider set to deploy it later this year alongside Vera CPU and Rubin GPU.
HanTest (漢測) reported NT$2.668 billion in cumulative revenue for January–July 2026, up 128.18% year-on-year and already exceeding its entire 2025 revenue of NT$2.425 billion, as the company prepares for a late-September 2026 Taipei Exchange listing built on expansion into memory testing, advanced packaging, and silicon photonics.
Apple unveiled the M6 and M5 Ultra chips on August 25, 2026, alongside new Mac mini and Mac Studio desktops. M6 is Apple's first 2-nanometer chip; M5 Ultra's four-die design lifts peak AI GPU performance up to 4.5x over M3 Ultra.
OpenAI says its Broadcom-built Jalapeño chip delivered 1.5x-1.9x more throughput per kilowatt and 1.7x-3.6x lower latency than Nvidia's GB200/GB300 on SemiAnalysis's InferenceX suite, though the lead narrows to about 1.5x under utility-power and multi-token-prediction comparisons, and Nvidia's upcoming Vera Rubin platform was not tested.
Renesas Electronics' Kawajiri plant, which generates roughly 13% of the company's revenue, returned to pre-earthquake production capacity on the evening of August 23 — three weeks after Renesas' Kani plant reached the same milestone on July 31. Both plants had halted operations after the magnitude-7.1 Kumamoto earthquake on July 28, in the same event that also disrupted Kumamoto-area operations at TSMC, Sony, and Tokyo Electron.
L'Express reported that China is poaching Taiwanese chip engineers with salaries three times market rate, as Taiwan's roughly five-year lead in advanced chips narrows and 17 Chinese firms now face investigation for illegal recruitment.
NVIDIA's Vera Rubin platform moves agentic AI infrastructure into production across three layers at once: SpaceX and SpaceXAI are deploying the Vera CPU, Nebius is adopting the newly mass-produced Groq 3 LPX inference system, and CoreWeave is running Spectrum-X Multiplane networking that scales to 512,000 GPUs — turning the AI factory into one coordinated, full-stack system rather than a single-chip upgrade.
TSMC's median employee total compensation reached a record NT$2.95 million in 2025, up about 20% from NT$2.46 million in 2024, as global headcount topped 90,000 for the first time, per-capita pay expense climbed to NT$4.33 million, and the president-to-median pay ratio stood at 822-to-1.
TSMC says information and communication technology products built on its chips saved the world more than 180 billion kWh of electricity and cut close to 89 million metric tons of carbon emissions in 2025, citing an ITRI estimate. The company has also aligned with SBTi absolute emission-reduction targets and set a 2030 goal of saving nearly 7 kWh globally for every 1 kWh it consumes making semiconductors.
IBM unveiled a mainframe processor whose 11 cores natively execute either its own Z instruction set or Arm's AArch64, switching modes within nanoseconds. Built on a 2-nanometer process with a base frequency above 5.7 GHz, the chip will anchor the successor to the z17 and ships alongside a new Spyre AI accelerator, with IBM framing Arm support as an addition to, not a replacement for, its mainframe architecture.
Global semiconductor sales hit $403.3B in Q2 2026, up 35.1% from Q1, as WSTS lifts its 2026 forecast to $1,655B on surging memory and HBM demand.
Japan's Ministry of Economy, Trade and Industry plans to request an additional ¥150 billion (~NT$30 billion) for Rapidus in next year's budget. If approved, it would be the government's third capital injection, lifting cumulative state capital in the company to ¥400 billion, while total government support is projected to reach roughly ¥3 trillion by March 2028.
The US government is preparing to push the Dutch government to bar ASML from selling almost all chipmaking equipment, including DUV lithography systems, to China. A bipartisan bill in Congress, the MATCH Act, would also ban maintenance on DUV machines already delivered to China and threatens sanctions on the Netherlands if it refuses to comply, while Chinese toolmaker Shanghai Yuliangsheng Technology has begun limited domestic DUV production.
NVIDIA reports fiscal Q2 earnings on August 26, with analysts projecting record revenue of $92 billion, above the $78 billion initial-year estimate. After 14 straight quarters of beating forecasts, analysts say NVIDIA needs 95% profit growth and $51.5 billion in net income to clear the bar again, while options price a 5.3% post-earnings swing and its shares have fallen after each of the last four reports.
NVIDIA has told its largest customers that AI server prices will rise by more than 15% on Grace Blackwell and Vera Rubin systems shipping early next year, as memory suppliers led by Samsung, SK Hynix and Micron struggle to keep pace with demand and NVIDIA opts to pass rising memory costs to customers rather than absorb them within its roughly 75% non-GAAP gross margin.
Google began selling the Pixel 11 series in Taiwan on August 21, 2026, leaning on on-device AI features like one-tap capture and AI voice cleanup rather than raw spec increases. The base Pixel 11 holds its NT$29,990 price while storage now starts at 256GB, and the Pixel 11 Pro Fold gets a 6.5-inch, 3,600-nit outer screen alongside a seven-year update guarantee.
Broadcom is negotiating with a group of lenders to raise more than $60 billion in debt financing for AI chip infrastructure benefiting Anthropic and other companies, with the package potentially reaching $100 billion once a proposed $30 billion junior-debt tranche and a $60-70 billion Broadcom-guaranteed senior-secured tranche are combined. Blackstone and Apollo Global Management are among the institutions in talks to participate.