Tuple Technologies at DAC 2026

At #DAC2026 I met with Vamshi Kothur at Tuple Technologies to see how their cloud infrastructure products help IC design teams get their chip designs to market more quickly without the overhead of big IT and CAD teams. Tuple was founded in 2017 and has grown to about 45 people, headquartered in New Jersey. The challenge of using complex EDA design and verification flows is that there are three options for computing: Cloud, On-premise or a Hybrid approach. Recently, the cloud option has become a big attraction, but it makes IT more complex to manage all the data and EDA licenses. You could build up your own dedicated IT and CAD teams for cloud or co-location, but that requires precious time and money.

Cloud providers AWS, Google and Azure are each slightly different and create a learning curve for IT & CAD teams to master. At Tuple the focus is on IC start-ups, because no IT and CAD teams are needed with the Tuple software stack, as it is a subscription as a platform for IT infrastructure, where you can get ready for cloud use in just 20 minutes or less. The technology platform is called Stratos and is packed with enterprise apps for semiconductor and IC design teams.

As an example a Tuple user could choose AWS, select their compute and CPU, memory, users, licenses, then it gets setup with your favorite scheduler tool (SLURM, SGE or LSF). This approach is called Infrastructure As Code, IAC, with benefits, like:

• A software-based service enables design version control, so over time it evolves, and you can roll back in time to what works if system configurations drift. Also, this provides for fast disaster recovery and security. One customer encountered a Ransomware attack, but they could just roll back in time to have a clean start.
• Security is included in the platform.
• A specialized IT & CAD teams would take 3 months to do a similar setup, but now it’s done in 20 minutes. Your make or buy decision is done quickly as Tuple offers an annual subscription and has 24/7 support. Tuple IT works with your IT & CAD teams to get trained too.

Once the IT infrastructure is built using Forge AI with plain-language requests, there are apps to manage productivity with an easy learning curve, like:

• Wiki for specs and documentation
• Ticketing for bugs using an open source option, or your favorite, like Jira<
• Nitro – a license monitor with analytics on usage, integrates with job queuing schedulers in one spot
• AI layer to extract analytic and management reports with natural language

Omni is used for multi-cloud orchestration, so you may be doing some LLM training or inferencing that require too many GPUs that may exceed allocation, and then Omni can check out other Hyperscaler or Neo cloud to provide you with availability with lower costs. You can even use Spot instances (excess compute at Hyperscalers) to lower costs. At a recent Cadence Live presentation, Vamshi Kothur presented on using Cadence Pegasus on the Cloud with Spot instances to improve turnaround (TAT) times with lower costs.

Tuple has a Security App, Shield, that is used in public infrastructure for end to end security for laptops to servers, keeping ransomware, Malware out of your infrastructure. No security experts are required, because all the security is baked in with the app with Managed Detection & Response (MDR).

What about the case where in the middle of the night some bad RTL is checked in, or any part of the Infrastructure is broken but the people who can fix it are unavailable, what should you do? With Tuple Tech’s TroposAlerts, you get alerts sent to responsible team members that something is broken, so that they can identify and remedy the cause. This Incident Management & Escalation app will send alerts, and if there is no response quickly then it escalates higher up in the chain of command. Team members can even be paged on phone, texted, or emailed to get their attention to resolve issues in real time.

Financial types will love the Billing tracker app, as it has a dashboard to help manage costs throughout the entire duration of the project. No more guessing on cloud and tool costs. If you are worried about exceeding a cloud EDA budget, then setup a budget allocation and then it can be tracked and be alerted across users, teams, projects and departments. Tuple’s EDA license tracking and job management integration is handled by the Nitro License Monitor, cutting engineering software costs by up to 25% by using data-driven renewal decisions. Nitro is integrated with job schedulers and queue visibility which provides for a holistic efficient usage of EDA licenses and compute resources.

Startups can get funded, then buy some EDA tools, but they cannot use Calibre inside of a Cadence chamber. Tuple allows Cadence and Siemens tools in the same network, problem solved. Tuple lets you mix and match tools from multiple EDA and IP vendors.

Ashok Mehta from Tuple Tech talked about automatically generating SystemVerilog Assertions (SVA) using , starting from natural language specs and RTL, supporting the major simulators: Siemens – Questa, Synopsys – VCS, Cadence – Xcelium. Claude can be used for SVA generation, however you must live with 30% hallucinations giving incorrect results. Avestra is an 8-agent pipeline that has been trained to eliminate incorrect assertions caused by tautology. Experience gained from Ashok’s 22 patents and authoring of three books has helped Avestra avoid hallucinations, while employing a proprietary RAG database. Avestra finds inconsistencies in your RTL, crosschecks your spec to the RTL, then suggests fixes to your spec or RTL code. When generating SystemVerilog Assertions it will crosscheck assertions between spec and RTL. Avestra even generates a coverage-driven test bench.

When using Avestra you can triage which assertions are most important, focusing on the critical ones first. Initial users of Avestra include a big foundry and a memory design company. With these auto-generated assertions it will find bugs in your design, speeding up your time to close coverage goals.

Summary

Smaller EDA vendors take more risks and often reap more rewards than their larger competitors with new technology, and this is the case with Tuple Tech in 2026. If you’re a startup design company looking to use Cloud, On-prem or Hybrid resources, then it pays to check out what Tuple has to offer. Verification engineers will benefit from automating SystemVerilog Assertions using the AI-powered approach in Avestra.

Related Blogs

• • Tuple Technologies Develops AI-powered Infrastructure at 2026 DAC
  • Rethinking ECAD IT Infrastructure: From Fragmentation to an Engineering Platform
  • CEO Interview with Vamshi Kothur, of Tuple Technologies
  • Tuple Technologies at the 2025 Design Automation Conference #62DAC