AdaCore

How do you prove your code is correct before it ever runs on hardware? Part 3 of our Mars Rover series with Electronic Design is now live. The series follows how the SPARK programming language was used to control a model Mars Rover robot. In this installment, we show how Ada contracts in SPARK strengthen the reliability of this cyber-physical system, catching errors before the code ever runs. Read the full article: https://lnkd.in/ex7PmRP7 #SPARK #Ada #EmbeddedSystems #CyberPhysicalSystems

The EU Cyber Resilience Act requires manufacturers to know their software supply chain down to the component level. Your compiler is part of that chain. Most teams still treat the compiler as free infrastructure. In practice, that quietly shifts real risk onto engineering: no SBOM for the toolchain, no one monitoring it for vulnerabilities, and tool qualification work repeated on every project. This article makes the business case for treating your compiler as a managed, supported component of your product, with a contractual counterparty behind it. ➡️ Worth a read if you own risk on a long-lived, regulated, or security-sensitive program. https://bit.ly/4vCNQ1Y

AI can now find the defects hiding in your automotive C and C++ codebase. The question is whether you can verify the code you write. Our new CTO guide covers four approaches to memory safety in embedded systems, and why formal methods offer the practical payoff: speed is worthless in a safety-critical system without a way to prove correctness. https://lnkd.in/e836Kc-B AdaCore #automotive

Registration is open for our Public Ada Training sessions in the US. New to Ada? This five-day course takes you from the fundamentals through to advanced language concepts, led live by AdaCore's expert instructors. Over the week you'll cover: 🔹 Fundamental data types: numeric, enumerated, arrays, records 🔹 Expressions, statements, and subprograms 🔹 Modular program construction with packages and data abstraction 🔹 Exceptions and exception handling The format combines instructor-led lectures, interactive quizzes, and hands-on lab exercises, so you leave with practical experience, not just theory. It's designed for software professionals with little to no Ada experience. If you can code in any compilable language, you're ready. 📅 Next US session: September 14 to 18, 2026 (remote) 👉 More details and to register: https://bit.ly/3R64BEx Questions about pricing or group registration? Reach us at sales@adacore.com #Ada #AdaProgramming #SoftwareTraining #EmbeddedSoftware #SafetyCritical #HighIntegritySoftware

Memory safety is now a board-level concern for automotive software. With regulators writing secure-by-design expectations into the rules, what options do CTOs still building on C and C++ have? Read on and find out ⬇️

SPARK catches embedded software bugs before your hardware does. Every embedded C developer knows the pattern. The code compiles cleanly; you flash the board, and then something fails in the field. SPARK, the formally verifiable subset of Ada, is built for exactly this environment: constrained, safety-critical systems where a runtime crash is not an option. Our latest blog post walks through three concrete examples where SPARK eliminates entire classes of bugs that C leaves open, proving correctness at compile time. Read it here ➡️ https://bit.ly/4aEXHNc #SPARK #EmbeddedSystems #FormalVerification #SafetyCritical #Ada

High-integrity C development meets agentic AI. In this new video, Mark Hermeling demonstrates a collection of AdaCore tools for high-integrity software development in C, working on host or in a cross development environment. You'll see GNAT Pro for C, GNATcoverage delivering code coverage analysis up to MC/DC level, and CodeSonar static analysis, all working together and driven from an agentic LLM. It's a practical look at how AI can orchestrate proven verification tools: https://bit.ly/4wu53LH

Our latest AdaCore Digest is out today: a technical update that shares roadmap developments and news about what our engineering team is focusing on. In this issue: 💡 The use of artificial intelligence in developing high-integrity software 💡 26.2 updates 💡 Multi-language software development Read it here ➡️ https://bit.ly/3RkMuuv

Last week at SDV USA, Filip Gajowniczek caught up with the conference organizers to discuss the emerging trends in automotive software development. Software-defined vehicles live or die on software quality. As vehicles become platforms for continuously updated software, the cost of a defect no longer ends at the test stage, but extends to the road. That is where AdaCore comes in. We help automotive teams: 🔹 Build with memory-safe languages like Ada, SPARK, and Rust, eliminating whole classes of vulnerabilities by design 🔹 Find defects early with static analysis through CodeSonar, before they reach integration 🔹 Meet ISO 26262 requirements with qualified tools and proven workflows Teams at NVIDIA, Zenseact, Autoliv, eLeapPower, and Stoneridge are already building safer automotive software this way. 🎥 Watch the full interview: https://lnkd.in/dwfyRRaB If your roadmap includes more software and less tolerance for failure, let's talk. #SoftwareDefinedVehicles #SDV #AutomotiveSoftware #FunctionalSafety #ISO26262 #MemorySafety

Large language models have made remarkable strides over the past few years, to the point that entire aspects of software engineering that once required careful, manual human attention are on the cusp of automation. Given that large parts of software development are going to be automated, where is the future of static analysis headed? Read the article 👉 https://lnkd.in/eAcpjxt7