Alphabet Inc. (NASDAQ: GOOGL) shares rose over 3% to $294.50 in Monday morning trading following Google’s unveiling of two specialized AI computing systems, the TPU 8tx and TPU 8i Max. These architectures aim to enhance data security and handle demanding AI workloads.
The additions broaden Google’s AI hardware portfolio as it contends with NVIDIA in the accelerator market. Notably, both systems leverage existing eighth-generation Tensor Processing Unit (TPU) chips within new hardware configurations, avoiding the development of entirely new processor architectures.
TPU 8tx Integrates AMD EPYC Processors for Confidential Computing
The TPU 8tx combines Google’s TSX 8t technology with fifth-generation AMD EPYC processors. This integration supports confidential computing, ensuring that sensitive information and AI models remain protected while being processed.
These security enhancements specifically target government agencies and regulated enterprises handling proprietary models or sensitive data. Google reported that the TPU 8tx can reduce processing latency by up to 50% compared to the TPU 8t, delivering substantial speed improvements with minimal performance overhead.
Meanwhile, the TPU 8i Max targets high-demand workloads requiring greater computational power and memory. By housing two TPU 8i chips within a single package, the design doubles computing power, on-chip SRAM, and high-bandwidth memory bandwidth without increasing rack footprint.
Potential applications span processing large software codebases, extensive video content, and vast datasets utilized by autonomous AI agents.
Collectively, these releases expand Google’s eighth-generation TPU portfolio to four distinct systems, complementing the TPU 8t and TPU 8i models introduced earlier this year. Google Chief Technologist Amin Vahdat, speaking to Bloomberg, predicted the new configurations would launch in 2027.
Complementing the hardware, Google unveiled software updates to streamline TPU system management and optimization. The company announced open-sourcing TPU Sync, a tool for managing AI model data and memory across TPU clusters, alongside an open-source TPU agent suite that assists developers in deploying and tuning these systems.
Additionally, Cloud Pathways and dynamic TPU slicing are now generally available on Google Kubernetes Engine. Cloud Pathways assists customers in managing large TPU clusters, while dynamic slicing enables real-time reorganization of computing resources as workload demands fluctuate. Google reported that dynamic TPU slicing can accelerate job startup by up to fivefold without necessitating the recreation of computing groups.
TPU System Sales Contributed to Google Cloud’s Third-Quarter Revenue Growth
Google reported that its TPU business serves nearly 100 customers, ranging from AI developers and financial institutions to retailers and universities.
Furthermore, Alphabet supplies TPU systems for deployment in customers’ own data centers, expanding beyond its cloud-only model. In its third-quarter SEC filing, the company disclosed that it had concluded several agreements to provide TPU solutions for specialized, high-scale workloads requiring on-premises infrastructure.
Alphabet disclosed Google Cloud revenue of $30.71 billion for the third quarter of 2026, marking a 103% increase year-over-year. The company highlighted TPU system sales as a significant driver of this segment’s revenue expansion.
The prominence of TPU hardware within Alphabet’s cloud business is further evidenced in its quarterly SEC filing. The company indicated that TPU systems represent the majority of product sales within Google Cloud, a category distinct from its standard cloud-service fees and subscriptions. Specific revenue figures for TPU systems individually were not disclosed.
During the third-quarter earnings call on October 28, CFO Anat Ashkenazi anticipated continued growth for TPU system sales in the fourth quarter, projecting a meaningful revenue increase in 2027.
Alphabet Stock and Tokenized Trading
Alphabet shares closed at $284.78 on Friday, posting a gain of roughly 1.8% over the preceding five trading sessions. Notably, the $294.50 morning price reflected a pre-market dip that had left the stock approximately 9.4% below its closing value of $324.89 on October 28, prior to the release of third-quarter financial results.
Separately, Alphabet’s tokenized trading instrument (GOOGLB) traded at $293.35, rising 3% in the last 24 hours per CoinMarketCap data. Unlike the standard Nasdaq-listed shares, this tokenized asset operates independently and tracks a distinct 24-hour value curve.
Also Read
- Europe Can Still Hold Up a $500 Crypto Transfer as the U.S. Retreat Leaves Limits Intact
- Bybit Expands RWA Earn Into Japanese Equities With New JPE Product
- Fed Intervention Expected to Stabilize S&P 500 Amid Market Volatility
- Nvidia Shares Surge Down Over 5% After Board Launches Investigation into Alleged Espionage Against CEO