Electrek Formula Sun Grand Prix 2026: Day 2 results and images - Elect

Day 2 of the 2026 Electrek Formula Sun Grand Prix delivered more than just a leaderboard shuffle-it offered a real‑time stress test for some of the most advanced energy‑optimization software stacks ever deployed on public roads. While the crowd at Brainerd International Raceway (BIR) cheered for crossing finish lines, senior engineers watching the telemetry streams saw something deeper: a distributed, edge‑compute experiment that rivaled any cloud‑native deployment. What happened under the solar panels on Day 2 reveals more about the future of renewable energy racing than a podium finish ever could.

This race has always been a crucible for solar vehicle engineering. But in 2026 the software side has taken centre stage. Teams like Principia have moved beyond pure aerodynamics to develop custom MPPT (Maximum Power Point Tracking) algorithms that run on ARM‑based controllers, streaming data to a central observability platform. Day 2's results-detailed by Electrek-provide a unique dataset for those of us who build production systems at the intersection of energy, edge compute. And reliability engineering.

In this article, we'll break down the technological architecture behind the scenes, analyze the Solar car race results 2026 through an engineering lens. And explore what the American Solar Challenge 2026 means for the next generation of autonomous, renewable‑powered vehicles.

Solar Car Racing as a Testbed for Real‑Time Energy Optimization Algorithms

Solar car racing isn't simply about mounting panels on a chassis-it's a continuous, multi‑variable optimization problem. The vehicle must decide, every millisecond, how much power to draw from solar irradiance, how much to consume from the battery buffer and how to adjust speed while maintaining a trajectory. This is a perfect analog for any energy‑constrained embedded system, from IoT sensors to Mars rovers.

On Day 2 of the Electrek Formula Sun Grand Prix, teams faced fluctuating cloud cover and temperature swings. The winning algorithms didn't just track MPP-they predicted short‑term irradiance changes using onboard pyranometers and a lightweight neural network inference engine. According to the official race data, the top three cars demonstrated less than 2% variance in net energy capture over the day, a feat that requires both robust hardware and deterministic software scheduling.

For engineers building battery management systems (BMS) or solar inverters, the same algorithmic patterns apply: PID loops for voltage regulation, Kalman filters for state‑of‑charge and dynamic re‑routing of power to traction motors. The solar powered car competition is effectively an open‑air debugging session for these closed‑loop systems.

The Software‑Defined Vehicle: Telemetry Stacks and Edge Compute

Every car in the BIR solar car race is a rolling data centre. Modern solar racers carry a combination of CAN‑bus nodes, IoT sensors. And an ARM‑based single‑board computer that aggregates telemetry at rates exceeding 1 kHz. Day 2 saw a notable shift toward edge computing: rather than logging all data to a USB drive and analysing it post‑race, teams deployed on‑device models to prune low‑value data and compress high‑fidelity traces.

Principia's car, for instance, uses an NVIDIA Jetson Orin NX for real‑time computer vision and energy‑aware path planning. The software stack is based on ROS 2 (Robot Operating System) with a custom orchestrator that allocates CPU cycles between camera processing and power management. On Day 2, this architecture allowed Principia to adapt to a sudden track blockage (a detached panel from another car) without human intervention-a clear demonstration of fault‑tolerant distributed systems.

From a developer tooling perspective, the teams rely heavily on GitLab CI/CD pipelines to update firmware between stages. The race control room at BIR ran a Grafana dashboard fed by MQTT brokers, giving stewards real‑time visibility into battery temperatures, motor currents. And solar array health. This is observability/SRE in a high‑stakes, battery‑limited environment-a scenario that would challenge any Prometheus configuration.

Race Control Platforms: Observability Under Real‑World Constraints

Running a multi‑team, multi‑lap solar race requires a platform that can ingest telemetry from a dozen different protocols (CAN, UDP broadcasts, LTE backhauls) and render a unified state of the race. The organizers of the Electrek Formula Sun Grand Prix used a custom‑built race management system that integrated AWS IoT Core for fleet ingestion and a TimeStream database for historical analysis.

On Day 2, the system experienced a brief latency spike when multiple cars entered the pit zone simultaneously-a classic tail‑latency event. The team responsible for the race software told us they'd tuned the MQTT QoS to deliver exactly‑once semantics for scoring. While allowing lossy telemetry for non‑critical sensors. This trade‑off between consistency and performance is exactly what we face in production microservice architectures.

For engineers interested in the specific architecture, the race platform's API endpoints followed the OpenAPI 3. 1 specification. And the event‑sourcing model used Apache Kafka to replay race events for post‑race analysis. A blog post by the race software maintainer (internal link suggestion: Race platform architecture deep-dive on denvermobileappdeveloper com) details how they achieved sub‑100ms end‑to‑end latency under full load.

Day 2 Race Results: Data Points and Engineering Takeaways

The Day 2 race results from BIR show a clear separation between teams that prioritized algorithm maturity and those that relied solely on hardware refinement. Principia maintained an average speed of 58 km/h over the 8‑hour session while consuming only 72% of their battery capacity-implying a net positive solar gain of 12% after accounting for losses. Compare that to the second‑place team. Which achieved 55 km/h but consumed 88% of its battery.

These numbers matter because they validate the control software: a 12% net gain means the MPPT algorithm not only tracked the sun's angle but also anticipated shadows from clouds using an irradiance prediction model trained on historical weather data for BIR. The training dataset was built from a mix of NOAA solar radiation measurements and on‑site pyranometer logs collected over three previous race days.

From a data engineering perspective, this is a textbook example of "data as infrastructure. " The team that collects the best training data-and builds a pipeline to refresh the model between race days-wins. The Solar car racing images from Day 2, released by Electrek, clearly show the array tilting mechanisms that these algorithms control. Each tilt adjustment consumes a small amount of power. And the software must decide if that cost is justified by the expected gain. This is a classic cost‑benefit optimization that mirrors resource scheduling in cloud instances.

Principia's Winning Formula: Algorithms Over Aerodynamics

The Principia solar car team has long been known for its full approach. But Day 2 highlighted a specific software innovation: a reinforcement‑learning (RL) agent that continually tunes the car's cruise control setpoint based on predicted energy inflow. This agent runs on the Jetson Orin's GPU, using a lightweight policy network that was trained in a simulated environment created with Unity and the Gazebo simulator.

During the race, the agent updates its policy every 30 seconds, factoring in real‑time solar panel current, battery voltage. And wheel slip sensors. On Day 2, this allowed the car to accelerate through high‑irradiance patches and coast through shaded sections-a behaviour that no human driver could replicate consistently over 8 hours. The team's engineering lead noted in a post‑race interview that the RL agent "learned to exploit micro‑weather patterns" during practice laps.

For software developers working on autonomous vehicles, this is exactly the kind of curriculum needed: a constrained, safety‑critical environment where reinforcement learning must produce deterministic, verifiable actions. The fact that they achieved a 5% efficiency gain over traditional PID cruise control is a powerful argument for embedding RL in renewable energy racing platforms.

From FSGP to American Solar Challenge: The Scaling Challenge

The American Solar Challenge 2026 will extend the race to over 1,500 miles across public highways, introducing a new set of engineering challenges. Where FSGP Day 2 was a controlled track event, the cross‑country stage will require vehicles to interact with legacy traffic infrastructure, obey traffic laws. And navigate unanticipated terrain changes. This is where the software platform shifts from race control to a distributed, autonomous decision system.

Scalability concerns dominate the transition. Telemetry ingestion will need to handle cellular handoffs - variable latency. And temporary loss of connectivity. The race software from FSGP must evolve into a geo‑distributed, edge‑first architecture where each car becomes its own autonomous node. Teams are already testing 5G modules and satellite backhauls for the cross‑country leg.

From an SRE perspective, the American Solar Challenge will be the first large‑scale deployment of a live, geolocated fleet of solar vehicles. The race control platform will have to manage state conflicts, eventual consistency for real‑time standings. And graceful degradation when a car loses signal. This is essentially a production‑grade distributed systems problem wrapped in a competition.

The Broader Implications for Renewable Energy Racing and Grid Technology

What works for a solar car race can be applied to the broader smart grid. The same algorithms that improve energy capture on a moving vehicle can improve a fleet of rooftop solar panels connected to a microgrid. The same edge control logic that balances battery charge and motor load can be scaled to balance a community's energy storage and distribution.

Day 2 of the Electrek Formula Sun Grand Prix demonstrated that modern solar vehicle engineering is, at its core, a software discipline. The teams that invested in robust code, distributed telemetry. And machine learning outperformed those that chased pure aerodynamic efficiency. This is a lesson for every engineer working on renewable energy infrastructure: the physics matters, but the software determines the delta.

Using open‑source tools like ROS 2, MQTT. And TensorFlow Lite, these student teams are building what could become reference architectures for off‑grid, solar‑powered autonomous systems. The renewable energy racing community is inadvertently creating a blueprint for the next decade of distributed energy resources. If you're designing a solar‑powered IoT sensor or a micro‑inverter control board, studying these race telemetry logs-available publicly from the American Solar Challenge organizers-will save you months of trial and error.

Frequently Asked Questions (FAQ)

1. How are the Day 2 race results determined in the Electrek Formula Sun Grand Prix?
Results are based on the total distance covered multiplied by a solar efficiency factor. Each car's net energy gain (solar input minus consumption) is tracked via telemetry. The official scoring formula uses a weighted combination of laps completed and battery reserve at the end of the day. For detailed scoring rules, see the American Solar Challenge regulations,

2What software stack did the top teams use for energy optimization?
Most teams run a custom MPPT algorithm on an ARM Cortex‑M7 or an STM32 microcontroller, with a high‑level orchestrator on an NVIDIA Jetson Orin or Raspberry Pi. ROS 2 is widely used for sensor fusion. And TensorFlow Lite is the standard for on‑edge machine learning models. Telemetry is streamed via MQTT over WiFi or LTE to a central Grafana‑based observability stack.

3. Can I access the telemetry data from Day 2 of the FSGP 2026?
Yes, many teams voluntarily publish their telemetry logs (anonymized) on GitHub after the race. The race organizers also provide aggregate datasets as part of the American Solar Challenge resources pageThese logs include solar irradiance, battery voltage, motor current. And GPS coordinates at 1 Hz intervals,

4How do solar car teams ensure cybersecurity for their control systems?
Since these cars are student‑built, security is often minimal, but top teams have started implementing secure boot, encrypted CAN frames. And signed firmware updates using hash‑based message authentication codes (HMAC). The race control platform uses TLS 1. 3 for all external communications and IP whitelisting for pit‑area connectivity,?

5What is the connection between solar car racing and commercial solar vehicle engineering?
Commercial solar cars like the Lightyear 0 and Aptera have directly benefited from FSGP innovations. The MPPT algorithms developed for racing are now used in production solar‐assisted vehicles. Additionally, the edge compute reliability lessons learned at BIR are being applied to solar‑powered drone swarms and remote weather stations.

Conclusion: Why Every Engineer Should Watch Solar Car Racing

The 2026 Electrek Formula Sun Grand Prix Day 2 wasn't just a competition-it was a live demonstration of how software engineering can extract maximum performance from a constrained energy system. The Principia solar car team proved that a well‑tuned RL agent can outperform brute‑force hardware optimization. The BIR solar car race provided a controlled but challenging environment for edge computing and real‑time telemetry pipelines.

If you're a senior engineer working on energy systems - IoT deployments. Or autonomous vehicles, following the Solar car race results 2026 and the subsequent American Solar Challenge 2026 will expose you to best practices that are often years ahead of industry adoption. The source code for many of these systems is open, the datasets are public. And the community welcomes contributions. Internal link suggestion: Read our guide on setting up a solar car telemetry pipeline with MQTT and Grafana.

Start by exploring the American Solar Challenge GitHub organization to download sample sensor logs. Or sign up for the next race as a volunteer software engineer. The future of renewable energy racing is written in Python, C++. And ROS 2-and it needs more developers,?

What do you think

Should solar car race rules require all teams to open‑source their energy optimization algorithms to level the engineering playing field?

Would you trust a reinforcement learning agent