NuAxiSemiconductor trade intelligence

[News] Google Marks First 4-TPU Year with 8tx and 8i Max, as 8i Max Doubles HBM Bandwidth to Tackle Memory Wall

What happened

On November 9, Google unveiled TPU 8tx and 8i Max, expanding its eighth-generation lineup, marking the first time Google has introduced four TPU systems in a single year. The 8tx integrates Google's proprietary chip technology with AMD processors, offering enhanced protection for sensitive information even while data is being processed via hardware-level data encryption, delivering up to twice the speed of its predecessor with virtually no performance penalty. The 8i Max combines two existing 8i chips in a single package, raising HBM capacity from 288 GB to 384 GB, doubling memory bandwidth from 8.6 TB/s to 17.2 TB/s, lifting SRAM capacity to 768 MB and doubling compute performance, effectively delivering twice the compute density without increasing rack space; however, the report notes Google has yet to officially confirm the new chip's total HBM capacity. The move comes amid surging AI computing demand as Google tackles AI's growing memory bottleneck (memory wall). In background, TPU sales made a meaningful contribution to Alphabet's September-quarter results, with Google Cloud revenue more than doubling to $30.71 billion; industry analysts point to TSMC as a key beneficiary, Samsung and SK hynix reportedly share Google's HBM3E orders nearly evenly at roughly 50% each, analysts expect Samsung's HBM shipments to Google to double this year, Samsung Electronics is expanding collaboration with Google's TPU development partner Broadcom across HBM, foundry services and advanced packaging, and SK hynix is pursuing joint development of custom HBM optimized for Google's TPUs.