Modular has released version 1.0 of the Mojo programming language. With this release, the company aims to provide a stable foundation for developers who wish to use the language over the long term. The release comes at an interesting time: Modular is now part of Qualcomm, while Mojo is specifically designed to be a relatively hardware-independent programming language for AI systems.
Mojo was introduced in 2023. The language was developed for general-purpose software development but explicitly targets applications where performance and various types of hardware are important. In combination with the MAX framework, Mojo is designed to support software running on a wide range of CPUs, GPUs, and other accelerators.
This addresses a major issue in the AI industry. Much software is tightly coupled to specific hardware platforms. Nvidia has built a large ecosystem around its GPUs with CUDA, while AMD offers an alternative with ROCm. With Mojo, Modular aims to provide a programming layer that is less dependent on a single vendor.
From rapid development to stability
The move to version 1.0 is primarily about stability. During development, parts of Mojo changed regularly. While this is not unusual for a young language, it makes it more difficult to maintain projects over the long term. Modular therefore states that changes within the 1.x series will mainly consist of extensions. Major changes remain possible but must be implemented more carefully from now on.
The 1.0 release also includes various simplifications and improvements. The language has been made more consistent in certain areas, support for development environments has been improved, and Mojo can now detect more memory safety issues. Several new features have also been added, including support for Python-like lambda functions.
Qualcomm changes the landscape
The promise of stability takes on added significance following Qualcomm’s acquisition of Modular. Qualcomm announced the acquisition in June and finalized it on July 29. According to Qualcomm, Mojo, MAX, and Modular Cloud will continue to exist, and Modular will remain focused on an open ecosystem.
At the same time, Qualcomm itself is a chip manufacturer. This raises questions about Mojo’s independence, writes The Register. If the language is to play a role across different hardware vendors, developers and companies will need to trust that Qualcomm will not steer development in a particular direction. Modular states that the acquisition does not change this. The upcoming open-source initiatives will be crucial for assessing whether that promise holds true in practice.
Compiler must become open source
Mojo’s standard library is already open source. The compiler and toolchain are not yet. Modular has promised to release those components in 2026 and reaffirms that commitment with the 1.0 release.
This makes the planned release more significant than just another step in the development process. When Mojo 1.0 was announced, The Register also noted that, precisely because of the Qualcomm acquisition, developers have questions about the project’s future independence.
Modular is hosting ModCon in San Francisco on August 18. There, the company plans to share more about the future of Mojo, MAX, and open source. Whether the compiler will actually be released at that event has not yet been confirmed.
Mojo 1.0 therefore primarily signifies that Modular considers the language a stable foundation. Development is not stopping. The roadmap includes improvements for asynchronous programming, pattern matching, and unions, among other things.
The bigger question is whether Mojo can build a sufficiently large ecosystem outside of Modular and Qualcomm. In addition to a stable language, this will require libraries, tooling, and trust from developers and other hardware vendors. The 1.0 release is therefore less of an endpoint than a new phase. The coming period will determine whether Mojo can grow into a widely used programming layer for AI hardware.