- MENU
- Home
- Special Reports
- Systems & Design
- Low Power-High Performance
- Manufacturing, Packaging & Materials
- Test, Measurement & Analytics
- Auto, Security & Edge AI
- Knowledge Center
- Videos
- Startup Corner
- Business & Startups
- Jobs
- Technical Papers
- Events
- Webinars
- Industry Research
- Newsletters
- Store
- Special Reports
Google Details Five Generations Of TPU Training Supercomputers
Researchers from Google and University of California, Berkeley published a technical paper titled “Google’s Training Supercomputers from TPU v2 to Ironwood: Architectural Stability, Scale, Resilience, Power Efficiency, and Sustainability Across Five Generations.”
The paper summarizes five generations of Google TPUs, from TPU v2 through Ironwood, and examines how the systems evolved into scalable, resilient, power-efficient, and more sustainable supercomputers for AI training. It describes the architectural stability of the TPU platform across rapidly changing neural-network workloads, including Transformers. Across eight years, the paper reports major gains in HBM capacity and bandwidth per node, peak node performance, and total supercomputer performance. It also discusses optical circuit switches, built-in self test, and hardware replay for resilience, as well as improvements in performance per watt and carbon emissions per floating-point operation. The paper concludes with six features the authors identify as likely characteristics of successful training accelerators in this decade.
Find the technical paper here. June 2026.
arXiv:2606.15870v1
Jouppi, Norman P., Sridhar Lakshmanamurthy, Cliff Young, and David Patterson. “Google’s Training Supercomputers from TPU v2 to Ironwood: Architectural Stability, Scale, Resilience, Power Efficiency, and Sustainability Across Five Generations.” IEEE Micro, July/August 2026. https://doi.org/10.1109/MM.2026.3699647.
Technical Papers
- Google Details Five Generations Of TPU Training Supercomputers June 16, 2026 by Technical Paper Link
- Modeling Multi-GPU Traffic For Distributed AI Workloads (UW Madison, AMD) June 15, 2026 by
- Physical Neural Networks: A Survey (U. of Lübeck, TU Hamburg) June 15, 2026 by
- Fault Injection Framework Targets RISC-V Security Weak Spots June 14, 2026 by
- Cross-Validated Timing Analysis for Automotive CAN Networks (NYCU et al.) June 12, 2026 by Technical Paper Link
Related Articles
Can A Computer Science Student Be Taught To Design Hardware?
Advanced Packaging Limits Come Into Focus
Startup Funding: Q1 2026
All AI Data Center Interconnects Will Be Optical Within 5 Years
Making Hybrid Bonding Better
When Semiconductor Materials Misbehave
Backside Power Delivery Creates Fab Tool, Thermal Dissipation Barriers
The Sub-2nm Paradox
-
Sponsors
-
Newsletter Signup
Popular Tags
Recent Comments
- Brian Bailey on Disturbance In Verification
- David Murray on Disturbance In Verification
- Ramesh Narayanaswamy on Disturbance In Verification
- Cliff Greenberg on Research Bits: Jun. 2
- Peter Ateshian on Detect, Diagnose, And Debug Using Sensors And Functional Monitoring
- Dr. Dev Gupta on Chip Industry Week In Review
- Ed Sperling on The Sub-2nm Paradox
- Jan Hoppe on The Sub-2nm Paradox
- Jan Hoppe on Toward Agentic Verification
- Riko R on With Chiplets, What Role Does Economics Play?
- C Lytle on Flash Getting Stacked High-Bandwidth Version
- Saket Goyal on HBM Shifts Testing Left To Preserve AI Chip Yield
- Debamitro Chakraborti on Creating Agentic EDA Methodologies
- Wim Bakker on Batteries Charge To The Edge
- Ron Richter on TSV Complexity Leads To Manufacturing Bottleneck
- François on The Petabyte Problem: How AI Is Finally Making Semiconductor Manufacturing Data Actionable
- Anonymous on The One Bit Problem That Can Break a System
- JC Bouzigues - IROC Technologies on The One Bit Problem That Can Break a System
- Tom Fitzpatrick on All Software Is Hardware-Dependent
- Anonymous on Memory Wall Gets Higher
- Gregory Haley on Detecting Chemical Variability At Advanced Nodes
- Matt Groves on A Framework That Generates Chip Layouts Directly From Natural Language Specifications (U. of Bristol, RAL)
- AICHIPS on How AI Will Impact Chip Design And Designers
- Martijn Fransen on Detecting Chemical Variability At Advanced Nodes
- Narek on Digital Twins: The Cloud’s The Limit
- Freelancer on Chip Industry Talent Shortage Drives Academic Partnerships
- Chris Riches on Backside Power Delivery Creates Fab Tool, Thermal Dissipation Barriers
- Jon on Robust Dynamic Voltage Droop Mitigation And Power Management
- Ashfaq on Minimum Energy Per Query
- Robert Boissy on Can A Computer Science Student Be Taught To Design Hardware?
- Hank Walker on Can A Computer Science Student Be Taught To Design Hardware?
- fieke khan on Bump Co-Planarity And Inconsistencies Cause Yield, Reliability Issues
- Rajiv on Chip Industry Week in Review
- Guy van der Walt - Thermco on Chiplets And 3D-ICs Add New Electrical And Mechanical Challenges
- Stephen W on Chiplet Fundamentals For Engineers: eBook
- ShrdMem256 on The Verification Conundrum
- Brendan McGuire on Wafer Probe Struggles To Adapt To Multi-Die Assemblies
- NobodyTestedThis on The Verification Conundrum
- Fiber Techs San Francisco on Agentic AI In Chip Manufacturing
- SF Fiber Techs on Oxides Bring Low Leakage Transistors To Leading-Edge Memories
- humic on Addressing Critical Tradeoffs In NPU Design
- Dr. Dev Gupta on Liquid Cooling Gains Traction In Data Centers
- Mikko Utriainen, Chipmetrics on Metrology Digs Deep To Produce Next-Generation 3D NAND
- Raymond Doerr on Wafer Bonding Mechanisms Using SiCN Films For Hybrid Bonding Applications In 3D Integration
- Hoyong Lee on Chiplets Vs. Soft IP: Different In Almost Every Way
- Nano Banana AI on Blog Review: Dec. 17
- JS Paek on Thermal Management In 3D-IC: Modeling Hotspots, Materials, & Cooling Strategies
- Larry K on Chiplets Vs. Soft IP: Different In Almost Every Way
- Seceon.Inc on Resilient And Optimized GenAI Systems
- Linda Christensen on AI With Open And Scaled Data Sharing in IC Manufacturing (NIST)
- komputer on AI With Open And Scaled Data Sharing in IC Manufacturing (NIST)
- Sree on The Thermal Trap: How Dielectrics Limit Device Performance
- David Muncier on China GenAI: Who Will Fill The Vacuum?
- acp on Is There Still a Future for Hard Disk Drives?
- Carlos Borer on Legacy Process Nodes Going Strong
- Raj Sodhi on AI Meets Device Modeling: Transforming Compact Modeling With Machine Learning
- Jaynarayan Tudu on Semiconductor Test Faces Technology Shifts In The AI Era
- Mihai Buta on Why In-Memory Computation Is So Important For Edge AI
- Silver Luis on The Competitive Advantage Of SRAM PUF Technology
- Gretchen Patti on Advanced Packaging: A Key Technology For The Next Generation Of Electronics
- Ev roach on AI Bubble Or Boom?
- Kathleen Murphy-Weber on Chip Industry Week in Review
- Raz on Multi-Die Verification
- Paul C on Scaling Memory With Molybdenum
- JWS on Scaling Memory With Molybdenum
- Klein miller on Electrifying Everything: Power Moves Toward ICs
- Team at Rise Design Automation on What Does Semiconductor Disruption Look Like?
- masa mizuno on Physics Limits Interposer Line Lengths
- Lou Covey on What Does Semiconductor Disruption Look Like?
- Ricardo Shiroma on Complex Mix Of Processors At The Edge
- Cor on Glass Substrates Gain Momentum
- Daya Young on New Antennas And Advanced ICs Needed For 6G
- Subramanian Srikanteswara Iyer on Manufacturing At The Limits
- Doug La Tulipe on Coloring Optical Signals For More Bandwidth In Data Centers
- WH on The Hidden Cost Of Contact Resistance
- Raymond Doerr on Metrology Under Pressure: Detecting Defects in Fine-Pitch Hybrid Bonding
- Raymond Doerr on Manufacturing At The Limits
- Mallikarjun on The Evolution of DRAM
- Klein Miller on Democratizing Design: How The CHIPS Act Is Reshaping EDA And Semiconductor Innovation
- Klein Miller on Lessons From 30 Years In The Trenches On The Future Of Semiconductor Manufacturing
- Jonathan Kolber on Preparing For The Quantum Computing Age
- Pramod Gupta on Challenges In Stacking HBM
- Madeline Lueilwitz on Reconfigurable Single-Walled CNT FeFET (Univ. of Pennsylvania, Yonsei et al.)
- Anne Meixner on Need For KGD Drives Singulated Die Screening
- Capt kirk on Designing A Better Clock Network
- Alex Martin on AI Effort And Money Misplaced
- Hamilton Carter on AI Effort And Money Misplaced
- Ilmu Komunikasi on Reconfigurable Single-Walled CNT FeFET (Univ. of Pennsylvania, Yonsei et al.)
- Giovanni Lostumbo on Will New Processor Architectures Raise Energy Efficiency?
- Brendon Berg on ESD Guns, Transients And Testing…Oh My!
- GD on All About Interconnects
- Raimo on Crisis Ahead: Power Consumption In AI Data Centers
- Fred Chen on Laser-Focused Results: Improving EUV Line Edge Roughness With Ion Beam Etching
- Fred Chen on Many Options For EUV Photoresists, No Clear Winner
- Amba Prasad on Advanced Packaging Fundamentals for Semiconductor Engineers: eBook
- Shakir Ullah on The Race To Replace Silicon
- Hwaiyu Geng, P.E. on Accelerating Semiconductor Process Development Using Virtual Design Of Experiments
- Larry Gus Christiansen on Reducing Risk In The Semiconductor Supply Chain
- Vikas Sharma on RTL Signoff vs. Functional Signoff: What’s The Difference?
- Simone on Crisis Ahead: Power Consumption In AI Data Centers
- Ron Lavallee on Can Today’s Processor Architectures Be More Efficient?
- Rashid on Can Today’s Processor Architectures Be More Efficient?
- Joseph Fjelstad on When Can I Buy A Chiplet?
- Craig Lytle on Can Today’s Processor Architectures Be More Efficient?
- Partha Thirumalai on Cognichip: Using AI To Speed Complex Chip Design
- Gopal raju on Advanced Packaging Fundamentals for Semiconductor Engineers: eBook
- Lawrence Kushner on Democratizing Design: How The CHIPS Act Is Reshaping EDA And Semiconductor Innovation
- Ken Rygler on Disruptive Changes Ahead For Photomasks?
- Jan Hoppe on Need For KGD Drives Singulated Die Screening
- B.S. DeepakSubramanyan on HBM Roadmap: Next-Gen High-Bandwidth Memory Architectures (KAIST’s TERALAB)
- Dr. Subhash L. Shinde , Booz Allen, Ex. Univ. of Notre Dame, Ex. Sandia National Labs., Ex. IBM. on The Race To Glass Substrates
- Traian MUNTEAN (Honorary Professor) on A Balanced Approach To Verification
- Amba Prasad on HBM Roadmap: Next-Gen High-Bandwidth Memory Architectures (KAIST’s TERALAB)
- sam on HBM Roadmap: Next-Gen High-Bandwidth Memory Architectures (KAIST’s TERALAB)
- Lentz on Agentic AI In Chip Design
- adam on RISC-V’s Increasing Influence
- Jon Taylor on RISC-V’s Increasing Influence
- Roi Mit on The DAC Valuation
- Janis Robin Schamberger on Chip Industry Week in Review
- Rashid Kukkady on New Data Center Protocols Tackle AI
Navigation
Privacy Overview
By continuing to use our website, you consent to our Privacy Policy. If you access other websites using the links provided, please be aware they may have their own privacy policies, and we do not accept any responsibility or liability for these policies or for any personal data which may be collected through these sites. Please check these policies before you submit any personal information to these sites.



Leave a Reply