Semi Doped
Semi Doped

Tensordyne's R K Anand: HPE Juniper Fabric, Logarithmic Math, MoE Inference, Air Cooling, 3nm

August 18, 2026

AI Summary

5 min read

The Logarithmic Bet: How Tensordyne Is Rethinking AI Inference

In the summer of 2023, a small team at a company then called Recognite faced a realization that would reshape their trajectory. ChatGPT had just exploded, and their building blocks—designed for automotive image recognition with convolutional neural networks—were unexpectedly well-suited for large language models. The market was "50 or 100 times larger," the path to revenue was far shorter, and their fundamental insight about logarithmic math suddenly had a much bigger stage.

R.K. Anand, co-founder and chief product officer, came to this moment with an unusual combination of experience. His 37 years in Silicon Valley had spanned both compute and networking—first at Sun Microsystems designing microprocessors for high-end servers, then as a founding engineer at Juniper Networks, where he spent 17 years building the silicon behind the routers that powered the early internet. That dual background, he argues, is precisely what the AI inference problem demands.

Why Logarithmic Math Changes the Silicon Equation

The core insight behind Tensordyne's approach is deceptively simple. AI inference—whether for image recognition or LLMs—runs on matrix multiplications: trillions of them. In standard arithmetic, multiplication is expensive in silicon. It requires more gates, more area, more power. Addition is cheap.

Continue reading the full summary in the app — free to try.

Read Full Summary →

Free • No credit card required

Listen to Audio Summary Open in App

What you'll learn

  • 1 (00:00) **Introduction & Guest Background** - Austin introduces R.K. Anand, co-founder and CPO of Tensordyne, and his unusual background spanning compute and networking.
  • 2 (03:04) **The Juniper vs. Cisco Analogy for AI** - Anand draws a direct historical parallel between Juniper disrupting Cisco and Tensordyne’s strategy against GPU incumbents.
  • 3 (09:03) **The Inference Window & Tensordyne’s Origin** - Anand explains how Tensordyne’s founding technology (logarithmic math) was originally developed for automotive edge inference before pivoting to data center LLMs.
  • 4 (17:44) **Pivot from Automotive to LLMs** - Anand describes how ChatGPT triggered the pivot and why the data center market was compelling.
  • 5 (21:08) **The Networking Problem & HPE Juniper Partnership** - Anand explains why they partnered with HPE Juniper for the scale-up fabric rather than building from scratch.
  • 6 (28:02) **Scale-Up Fabric vs. NVLink / UA-Link** - Anand contrasts their fabric with current alternatives and explains the latency advantage.
  • 7 (33:27) **The Disaggregation Era vs. Unified Architecture** - Anand explains why the industry is moving to split pre-fill/decode systems and how Tensordyne aims to unify them.

+ Full timestamped outline available in the app

Show Notes

Tensordyne co-founder and CPO R K Anand joins Austin to discuss the company's strategy for disrupting AI inference. RK explains how Tensordyne combines power-efficient logarithmic math with a battle-hardened networking fabric from partner HPE Juniper. The result is a high-density, air-cooled system designed to efficiently run massive Mixture-of-Experts models in existing data centers.

Key Takeaways:
- The core innovation isn't just log math, it's the patented method for accumulation. This turns expensive multiplications into cheap additions, freeing die space for a massive on-chip SRAM cache.

- Networking is a partnership, not a project. Tensordyne leverages HPE Juniper's 7th-gen router fabric, skipping development cycles to get a 1-2 microsecond latency solution ideal for random MoE traffic.

- The power and density claims are radical. By combining log math silicon with an air-cooled fabric, Tensordyne packs 72 chips into a 13U chassis at just 30 kW — a quarter of the space and power of an NVL72.

- One go-to-market advantage is air cooling. The 30 kW, 19-inch rack system can be deployed in existing 'brownfield' enterprise and telco data centers that cannot support liquid cooling.

- Partnerships de-risk the aggressive timeline. Broadcom provides access to TSMC 3nm and HBM, while strategic investor HPE Juniper provides the carrier-grade fabric with 'five nines' reliability.

Chapters:
0:00 Introducing Tensordyne
5:32 The Juniper vs. Cisco Playbook
11:29 Origin Story: Automotive Power Constraints
15:37 The Secret Sauce of Log Math
18:02 Pivoting to the Data Center
22:08 Leveraging a Router Backplane for AI
27:22 Why Router Fabrics Suit MoE Models
34:12 The Three Phases of Inference Hardware
37:40 How One Chip Handles Pre-fill & Decode
40:34 The 'Too Good to Be True' System Specs
43:31 Go-to-Market: The Air-Cooled Advantage
48:21 De-risking with Strategic Partnerships
52:37 Solving the Software Problem with AI

Follow Chipstrat:
Newsletter: https://www.chipstrat.com
X: https://x.com/chipstrat

Follow Vik:
Newsletter: https://www.viksnewsletter.com/
X: https://x.com/vikramskr

Follow Semi Doped:
Get more of Austin and Vik daily, free: https://daily.semidoped.com/

- The software moat is eroding. Tensordyne argues that modern agentic AI workflows can now automate the generation of optimized software kernels, solving the classic adoption problem for new hardware.

Semi Doped

More from this podcast

Semi Doped →