Damnang
Semiconductor Engineer in Silicon Valley, AI Chip Investment Research
STMicroelectronics is a hidden common thread across semiconductors, and the market is waking up to it in 2026. It turns up in silicon carbide and GaN power devices, its STM32 is effectively the MCU standard, and it leads MEMS sensor share. Its chips sit in SpaceX Starlink satellites, AWS data center infrastructure and a humanoid robot shown at NVIDIA's headquarters. A "European semiconductor" label long kept interest lukewarm, but that perception is breaking down fast. x.comÂ
Navitas, Infineon, Texas Instruments and STMicroelectronics each have a strong position in GaN and SiC power chips. Navitas is one of only three vendors with integrated-driver GaN and a credible SiC part, so it beats Texas Instruments when customers want dual-source GaN plus SiC. Texas Instruments has top-tier integrated-drive GaN but no SiC. Infineon has competitive GaN and very strong SiC. STMicroelectronics has a good SiC part, with GaN in development. x.comÂ
STMicroelectronics is an early turnaround story because AI infrastructure has given it a new growth engine. On June 2 it doubled its 2026 data center revenue target to about $1B, saying that could double again in 2027 if trends hold, on high-volume silicon photonics anchored by a multi-year AWS partnership. Q1 revenue rose 23% to $3.1B, it holds about $2.8B net cash, and Bank of America upgraded it to Buy, saying the market underestimates its earnings power. x.comÂ
Pure or heavy silicon carbide players such as STMicroelectronics, onsemi and Wolfspeed look higher-risk and currently unappealing in the AI data-center shift to 800V DC power and solid-state transformers. The better risk/reward in that transition sits with power-electronics names positioned for it, such as Enphase and SolarEdge, whose bidirectional power-conversion know-how suits transformers that step 13 kV utility power down to 800V DC and must handle sudden GPU load swings. x.comÂ
STMicroelectronics is one of the top names in power semiconductors and physical AI. Per @wallstengine, it raised its 2026 data-center revenue outlook to about $1B from "nicely above $500M" and expects that to double to about $2B in 2027 based on current customer engagements. About two-thirds is expected from optical components, including MCUs, silicon photonic PICs and BiCMOS EICs, with power chips the rest. x.comÂ
Navitas is a one-customer concentration story facing heavy competition from tier-one suppliers. Its GaN portfolio winning Nvidia adoption played out, but STMicroelectronics and Infineon also supply Nvidia and are strong in GaN and SiC. GaN output may quadruple by 2030, yet SiC prices were hammered by Chinese capacity, unresolved, so revenue and margins could diverge sharply. Power and thermal efficiency are better played across these names than through Navitas alone. x.comÂ
Infineon, STMicroelectronics and NXP have little value upside left from humanoid robotics after rallying hard on the AI datacenter theme. Infineon's CEO said humanoids could eventually rival AI datacenters as a revenue opportunity, about 20% of sales, yet Infineon saw insider sales this week. Humanoids may create few real bottlenecks: about 60% of parts come from the existing auto supply chain, most value will likely accrue in China and OEMs will integrate vertically, leaving heavy competition. … Show more
Infineon is a better pick than STMicroelectronics because data-center power chip demand is set to inflect while its smartphone exposure is much lower. About a fifth of STM's revenue still comes from smartphones, which are vulnerable to soaring memory prices. Rising rack power densities and 800V rollouts should lift demand, and Infineon has very competitive SiC and GaN products. Data-center power chips could top a quarter of Infineon's sales, more counting energy storage. x.com x.comÂ
STMicroelectronics offers rare asymmetry for a large cap: a bear case in line with the S&P 500, a slightly better base case and a wildly better bull case. 2026 marked an inflection in its data-center business, with a multi-year, multi-billion-dollar Amazon agreement and guidance raised from $1B before 2030 to well above $2B in 2027. A base case with a modest AI business implies 15% annual growth through 2030; if photonics, power and robotics reach a third of revenue, 30-50%. x.com x.com x.comÂ
Silicon carbide's years of oversupply are turning, as China's Yangjie Tech says its capacity is fully booked, per SCMP, a positive sign for Infineon, Wolfspeed and STMicroelectronics. As more makers fail to meet rising demand, orders should also flow to Wolfspeed, whose heavily underutilized New York SiC fab would gain tremendous operating leverage. x.comÂ
Power semiconductors should deliver index-beating returns, though the trade is not early and the easy money was made long ago. The power-semi revenue cycle already started last year, and the ecosystem is broad enough that the whole sector should do well. STMicroelectronics and Infineon are the buys. x.comÂ
