Mojo 1.0 and open source: a full-circle moment at ModCon 2026
From the first community Mojo talk at Scale By the Bay 2023 to Mojo 1.0 and an open compiler and toolchain at ModCon 2026.
I went to ModCon 2026 to see the next chapter of Modular and Mojo unfold in person. It turned into a full-circle moment for our community.
At Scale By the Bay 2023, I hosted Chris Lattner for a fireside chat about Modular, AI infrastructure, and Mojo. Mojo had only been publicly introduced a few months earlier. Scale By the Bay was the first community conference to put a Mojo talk onstage. We have always tried to recognize consequential systems work early—not once it is obvious, but while its community and technical direction are still taking shape.
Three years later, I watched Mojo reach 1.0 at ModCon. Modular also opened the Mojo compiler and toolchain, together with the language and standard library, under the Apache 2.0 license with LLVM exceptions. This is more than a version number and more than a source release. It gives developers a stable language epoch, the ability to inspect and improve the machinery underneath it, and a serious open foundation for heterogeneous computing.
The ModCon 2026 opening keynote
Watch the ModCon 2026 opening keynote
Chris Lattner and the Modular team lay out their vision for heterogeneous AI compute: Mojo 1.0, broader hardware support, an open MAX platform, Modular Cloud, and the company’s new chapter with Qualcomm.
The hardware question is becoming inseparable from the software question. AI systems increasingly span CPUs, GPUs, specialized accelerators, memory hierarchies, and interconnects. A language and toolchain capable of expressing that diversity without trapping developers in one vendor’s stack is an ambitious project. Mojo’s Python-facing ergonomics, systems-programming model, and MLIR lineage make it one of the most interesting attempts to meet that challenge.
Chris Lattner on Mojo 1.0 and its community
Watch my ModCon conversation with Chris Lattner
I interviewed Chris both then and now. At ModCon, I asked him what had changed since our 2023 conversation. His answer was striking: not much. The intervening years were about careful execution—turning the original vision into a stable language and safeguarding the heart of Mojo while the compiler, toolchain, platform, company, and community grew around it.
That continuity matters. Successful technology projects often accumulate so many compromises on the road to maturity that their founding insight becomes hard to recognize. Mojo arrived at 1.0 with its central ambition intact: combine Python’s approachability with systems-level performance and give developers a path across increasingly heterogeneous hardware.
Our short ModCon conversation—about Mojo 1.0, open source, its developer community, the Qualcomm partnership, and Mojo’s place alongside Python—is the direct sequel to our longer Scale By the Bay 2023 fireside chat.
In 2023, Mojo was a startling proposition: could one language feel natural to Python developers while also reaching down into systems and accelerator programming? At ModCon 2026, the discussion had moved from proposition to platform. The language is 1.0. The compiler and toolchain are open. Developers can now participate at a fundamentally deeper level.
Hardware, capital, and the next compute stack
Watch Chris Lattner moderate the hardware panel
Chris leads a wide-ranging discussion about heterogeneous AI hardware, specialized accelerators, interconnects, software portability, and the constant tradeoff between programmability and peak performance. The panel makes the case for why compiler infrastructure matters so much: hardware diversity is useful only when developers can productively reach it.
Watch the Funding Compute panel
Investors examine where AI infrastructure capital is moving—from inference software and silicon to power, open models, vertical integration, and robotics—and what may define the next generation of compute companies. It is a useful companion to the technical sessions because infrastructure transitions need ecosystems, companies, and patient investment as well as good abstractions.
The community around the machinery
Conferences are not only what happens onstage. This short hallway film captures the people, conversations, and energy around the future of AI software and heterogeneous compute. Open source makes those hallways larger: the community can now follow ideas all the way into the compiler, not stop at the product boundary.
That is why this moment matters to me. We were the first community to host a Mojo talk, when the language was new and its future was still speculative. At ModCon 2026, we saw Mojo become 1.0 and its compiler and toolchain become open source under Apache 2.0. The early conversation has become an invitation to build.
Watch the complete ModCon 2026 playlist, and revisit the 2023 Scale By the Bay fireside chat with Chris Lattner to see how far the project—and its community—have come.