Topic
Explore English articles about AI Semiconductor.
AMD and Cerebras plan to launch a division-of-labor AI inference platform in the second half of 2026, with Helios handling input processing and WSE handling token generation. The claimed up-to-5x figure comes from internal modeling comparing against WSE alone, and disclosure of measured latency and pricing will be the next thing to watch.
AMD has fleshed out the Instinct MI455X, equipped with 432GB of HBM4, and the Helios system with a 72-GPU configuration. After shipping begins in the second half of 2026, the question will be whether theoretical performance can be translated into real-world speed and operating costs.
Reuters reports that Intel and AMD are shifting to long-term contracts with Chinese server customers that lock in purchase volumes but not prices. Some server CPU prices in China have already risen more than 40% since the start of the year, and AMD has raised its 2030 TAM forecast from $60 billion to over $120 billion.
DRAM inflation has surfaced in GM's combined cost increases and in the price of BYD's driver-assistance option. Automakers and suppliers are expanding long-term supply contracts to prepare for shifts in both supply and price.
The new MCP spec, slated for publication on July 28, 2026, eliminates initialization and the protocol session. While it lightens cloud deployments, authorizing state handles and designing for old-new compatibility become challenges for implementers.
NVIDIA has reportedly raised procurement prices again for GeForce GPUs and VRAM, this time expanding the scope to include GDDR6 products. Model numbers and the size of the price increase remain unannounced, and it is unclear when the change might reach retail.
While Google begins selling TPUs externally, it has set cutting-edge AGI development as the top priority in its compute resource allocation. The company is weathering supply shortages through sales to customer facilities and procurement of external capacity.
Micro-transfer printing technology, which fabricates light-emitting and modulator materials incompatible with silicon separately before transferring them, is drawing attention. It leverages the advantages of CMOS processes while enabling high-precision integration of high-performance lasers and modulators on the same chip, positioning it as a promising foundation for next-generation high-speed communications.
Nanya Technology forecasts the 2026 memory market will exceed $800 billion. Based on WSTS outlook and the company's earnings, we track how AI inference demand, price increases, and 2028 new fab plans will move the market.
Google Cloud revenue grew 82% and operating margin expanded to 35.6%. Meanwhile, the payback on capex of up to $205 billion and shared AI expenses must be measured through backlog-to-revenue conversion and free cash flow recovery.
