TSMCActive
TSMC's 2.5D and 3D packaging technologies that integrate GPU dies with stacks of high-bandwidth memory on silicon interposers - the technology every leading AI accelerator depends on, and for two years the binding constraint on global AI chip supply. Capacity is being expanded roughly fourfold from late-2024 levels, including the new AP7 plant in Chiayi and two planned Arizona packaging facilities.
Latest update Aug 17, 2026
01
Aggressive expansion toward ~130-140k CoWoS wafers/month by end of 2026, up from ~35k in late 2024 (10–20% of the 2026 capital budget); as of August 2026, TSMC is also outsourcing the CoW front-end step to OSAT partners (ASE and others) to ease capacity pressure driven largely by Nvidia's reservation of an estimated 800,000-850,000 CoWoS wafers for 2026
02
03
TSMC's 2nm-class node and its first with gate-all-around nanosheet transistors, delivering 10–15% more speed or 25–30% less power than N3E. Volume production began on schedule in Q4 2025 at Fab 22 (Kaohsiung) and Fab 20 (Hsinchu), with the enhanced N2P variant following in H2 2026 - the process powering the next generation of flagship smartphone and AI silicon.
TSMC's US manufacturing site in Phoenix, now backed by a ~$265B commitment - the largest foreign direct investment in US history - after TSMC pledged another ~$100B on its July 16, 2026 earnings call to add at least four more 2nm-and-below fabs. Phase 1 has produced 4nm chips at Taiwan-par yields since Q4 2024 for customers including Apple and Nvidia; Phase 2 brings 3nm production, with equipment move-in from Q3 2026, pilot production in early 2027, and volume production targeted for Q3 2027 (pulled in about a year from the original plan); Phase 3, targeting N2 and A16, has broken ground for ~2029. TSMC expects ~30% of its 2nm-and-below capacity to eventually sit in the US.
TSMC's 1.6nm-class node combining gate-all-around nanosheets with Super Power Rail backside power delivery, targeting AI and HPC chips with 8–10% more speed or 15–20% less power than N2P. Volume production is now slated for 2027 after slipping from the original H2 2026 target; the follow-on A14, A12, and A13 nodes extend the roadmap to 2029 - all planned without High-NA EUV.
04
SpaceX
Partially reusable two-stage medium-lift launch vehicle. The world's most frequently launched orbital rocket.
SpaceX
Heavy-lift launch vehicle consisting of a strengthened Falcon 9 core with two additional boosters.
SpaceX
Reusable spacecraft for cargo and crew transport to the International Space Station.
SpaceX
Fully reusable super-heavy-lift launch system designed for missions to the Moon, Mars, and beyond. V3 (Block 3) variant with Raptor 3 engines is the production version.