Your weekly sync on the pulse of the simulation industry
Industry News
A team from 2getthere B.V. and SimPlan AG has used Siemens’ Plant Simulation software to model India’s first Personal Rapid Transit (PRT) system, planned for Mumbai’s Bandra Kurla Complex (BKC). The project involves a network of autonomous pods designed to move passengers across the business district. Albert Hennche (2getthere) and Niklas Rommelspacher (SimPlan AG) will present the work at the Plant Simulation Conference on October 12, 2026.
Using discrete-event simulation, the team built a digital twin of the full BKC network to test elements that paper-based designs alone could not, including:
- Passenger throughput
- Potential bottlenecks along the guideway
- Fleet sizing and dispatch strategy
The model identified four interconnected complexity areas — vehicle dynamics, shuttle communication, station and merge operations, and fleet and energy management — helping the team replace assumptions with measurable data on throughput, wait times, and utilisation before construction begins. The approach highlights the role of simulation in planning and assessing complex urban transport systems.
Product News
- Next-gen static analysis solver
- Enhanced electric machine NVH workflows
Physics AI for pickup aerodynamics
Surface pressure prediction
Automotive supplier Marelli uses Ansys FreeFlow with Ansys Mechanical to analyze oil flow and heat transfer in electric motors. The simulation approach aims to predict cooling performance under varying conditions, support thermal design, and validate results against physical test data.
Click to read more..
SimScale has opened its Engineering AI agent, earlier tested with enterprise clients, to its global community of engineers, designers, and students. It aims to autonomously run simulation workflows from CAD clean-up, setup, and validated reporting, helping attract and upskill hardware engineering talent.
Click to read more..
AVL has outlined a gas-flow co-simulation method linking CRUISE M and FIRE M, combining one-dimensional surrogate modeling with three-dimensional CFD. Demonstrated on cylinder-specific EGR distribution, the approach aims to improve predictivity for select gas-path components while managing simulation time via staged operation.
Click to read more..
NVIDIA has announced that SpaceXAI aims to deploy its new Vera CPU, built to help AI agents orchestrate tools, execute code and process data, for its next-generation agentic AI applications. SpaceXAI is also expanding Grok’s infrastructure on NVIDIA Vera Rubin and plans to extend it into an orbiting Starmind satellite.
Click to read more..
Industry Events
Simulating fluid-particle systems in fluidized bed reactors requires modeling complex hydrodynamics, heat transfer, and reaction chemistry together. This five-day virtual training builds proficiency in Barracuda Virtual Reactor, covering project setup, solver operation, and Tecplot post-processing, with optional hands-on breakout sessions.
📆 September 21-25, 2026
Designing high-power PCBs requires careful management of current paths, thermal performance, and insulation. This webinar explores practical layout techniques for reducing voltage drop and heating, addressing creepage and clearance, identifying thermal hotspots, and improving interconnect reliability.
📆 September 23, 2026
🕣 20:30 – 21:30 IST
Engineering teams are moving past AI copilots toward autonomous agents that run entire simulation workflows end-to-end. Agentic Unlocks, hosted by SimScale, is a new forum where engineering leaders and simulation experts share real production deployments of agentic AI in engineering and simulation. Join the forum to explore:
- Live lightning demos of agents running real engineering workflows
- Lessons from teams scaling agentic AI in production
📆 September 24, 2026
🕒 15:00 – 19:00 CET
This webinar explores the latest capabilities in Gexcon’s EFFECTS and RISKCURVES v13.1, including new ignition-source modelling for location-dependent risk assessments, extended battery hazard modelling for BESS thermal runaway, and buried gas pipeline rupture modelling via crater formation. The session also covers expanded language support and practical applications in consequence and risk analysis.
📆 September 24, 2026
🕙 10:00 – 11:00 CEST
In FOCUS
Company in Focus
Who they are
Neysa is an Indian AI infrastructure company, founded in 2023 and headquartered in Mumbai. It operates Velocis, a full-stack AI Acceleration Cloud System giving enterprises GPU compute, platform tools, and support to train, fine-tune, and deploy AI in production. It is backed by over $1 billion in funding.
Solutions & Technology
Velocis unifies the AI lifecycle:
- GPU-as-a-Service: on-demand NVIDIA H100, H200, L40S, L4, and AMD MI300X GPUs on VM, container, Kubernetes, or bare metal
- AI PaaS: managed, pre-integrated Jupyter, PyTorch, Hugging Face, and Kubeflow environments
- Inference-as-a-Service: dedicated, single-tenant endpoints on vLLM and SGLang
- Orchestration & MLOps: Slurm and Kubernetes-based training with CI/CD and model registry
- Unified Monitoring: a Command Center for real-time GPU and fleet telemetry
- Catalog: private repositories of models and tools within a customer’s own VPC
- Security & Control: RBAC, audit logs, Zero Trust, and the Aegis LLM Shield
- Marketplace: curated AI-native apps and agents, coming soon
Applications:
Neysa serves banking, insurance, retail, manufacturing, research institutions, and AI-native startups, with stated use cases spanning fraud detection, underwriting, voice AI, and large-scale training.
Did you know?
Before founding Neysa, co-founder and CEO Sharad Sanghi scaled Netmagic into India’s largest data-center company, later acquired by NTT.”
Solutions Focus:
Neysa positions Velocis around production-grade performance, lower unit economics than general-purpose clouds, and enterprise security with India-resident data, backed by a 24×7 NOC and dedicated engineers.
Technology Focus
Predicting exactly how a ship will sit in the water as it gathers speed has long been one of naval engineering’s quieter challenges. Navier AI examined this question through the Wigley hull, a scale-model shape favoured in marine research because its form is defined by a mathematical equation rather than physical measurement, reducing uncertainty tied to the hull’s own geometry.
The study aims to determine whether flow simulations can reliably predict a hull’s sinkage and trim across a range of speeds, and to assess the numerical uncertainty behind its resistance calculations, rather than adjusting results to fit known data afterward. Using OpenFOAM, an open-source simulation platform, engineers modelled the hull moving freely through water while tracking:
- Sinkage across a sweep of speeds, validated directly against earlier towing-tank measurements
- Computed trim trends, reported but not checked against experimental data
- Resistance, assessed through grid, time-step, and iterative uncertainty studies
- The simulated wave tank itself, independently verified before trusting hull results
Sinkage predictions closely matched the towing-tank data, and an early discrepancy was traced to mesh alignment near the waterline rather than the physics itself, then corrected at its source. The case study underscores a simple discipline: credible simulation depends on fixing errors where they originate, not smoothing them over afterward.