Category
Samsung has launched an experienced-hire recruitment drive spanning HBM design, packaging, and customer qualification. With an eye on scaling up HBM4 production and developing next-generation products, the company aims to strengthen coordination across processes to shorten development timelines and improve mass-production yields, boosting its competitiveness in the AI memory market.
The memory semiconductor market keeps posting record earnings on the AI boom, yet Samsung's stock fell instead. With more than three years needed from groundbreaking to mass production at new fabs, and SK hynix targeting full-scale capacity expansion only by 2034, this piece examines the supply lag and compares the current surge quantitatively against the two previous boom-and-bust cycles.
OpenAI and Anthropic are rolling out measures to ease usage restrictions on their top-tier models and improve work efficiency within fixed-rate plans. OpenAI is curbing consumption through reasoning optimizations, while Anthropic is temporarily expanding access to its premium model—intensifying competition on the practical front.
On July 13, 2026, Microsoft announced it had begun rolling out improvements to the Windows search box for the Experimental channel, removing promotional content, adding source labeling, and introducing new settings. This is the latest step in a gradual course correction that began with a policy statement in March, though no timeline for GA or quantitative speed improvements has been disclosed.
TSMC's revenue for June 2026 set a new all-time high, driven by robust demand for AI-related advanced semiconductors. Quarterly revenue also grew sharply, up 36.0% year-on-year, and with supply constraints expected to persist through 2027, the company is accelerating capacity investments and next-generation process rollouts.
In September 2026, Mozilla will experiment with biweekly major updates for Firefox Desktop and Android. Here's a breakdown of how this differs from weekly minor updates and ESR, and what it means for enterprise deployments.
Databricks' internal benchmark revealed a cost reversal: Sonnet 5, despite its lower per-token price, cost more per task than Opus 4.8. We unpack the structure—driven by completion rates and harness selection—alongside Uber's budget overrun case.
Reports have surfaced that the Pixel 11's Tensor G6 will use TSMC's 2nm process. Drawing on clues about a 7-core CPU and a MediaTek-based modem, we examine how power might be allocated across the SoC as a whole.
Intel's space-oriented SoC "Starfire" targets up to 75 TOPS using 18A and Foveros, but radiation tolerance figures and spaceflight qualification remain undetermined. Judging its real capability will have to wait for Q3 2026 samples and test results.
Meta is expanding its Hyperion facility in Louisiana to over $50 billion and 5GW. Here's a breakdown of the investment structure that incorporates outside capital, the ten planned gas power plants, and the regulatory review over rate protections.
