The AI arena has turned into a strategic battleground where tech giants compete for dominance with multibillion-dollar investments. At the heart of this game lies the most critical element: the hardware that powers artificial intelligence models. While Nvidia continues to lead the market, Microsoft has been developing its own solutions. However, the company has encountered an unexpected setback in its plans to launch a next-generation AI chip. The project, codenamed Braga, has been delayed. This postponement marks a significant interruption in Microsoft’s strategy to reduce its dependence on Nvidia and brings competitive industry dynamics back into focus.
Why Is Microsoft’s AI Chip Strategy So Important?
For tech giants like Microsoft, developing their own chips is not just a matter of prestige—it's a commercial necessity. The underlying reasons behind this strategy have the potential to directly shape the future of the market.
A Customized Solution to Challenge Nvidia’s Dominance
Today, most major tech companies are heavily reliant on Nvidia's GPUs. After Google’s TPU (Tensor Processing Unit) and Amazon’s Trainium, Microsoft followed suit by developing its own custom AI chip. This chip is specifically designed to be fully integrated with Microsoft’s services like Copilot and Azure OpenAI, aiming to optimize performance while reducing cost. It's part of a broader race toward more efficient results through specialized hardware. Google’s rapid advancement with its seventh-generation TPU chip is a testament to how fierce this competition has become.
The Goal: Lowering Cloud Computing Costs
The processing power demanded by generative AI services has driven cloud platform costs to astronomical levels. Microsoft is under pressure to cut costs on its Azure platform, and by deploying its own AI hardware in data centers, the company aims to significantly reduce the cost of services like Azure OpenAI. This move not only boosts Microsoft’s profit margins but also paves the way for more competitive pricing for end users.
At the Center of the Delay: The Next-Gen AI Chip "Braga"
Braga stands as one of the most pivotal milestones in Microsoft’s AI hardware roadmap, and while it’s still moving forward, the project is currently facing delays.
Braga’s Predecessor: The First-Gen AI Chip Maia 100
To understand Braga, we must first look at Maia 100, the first-generation AI chip introduced last year. Maia 100 marked Microsoft’s first major step into the AI chip arena and was designed primarily for Azure OpenAI services. Its success gave the company valuable experience in designing its own hardware. However, Maia 100 was never meant to directly compete with Nvidia’s latest Blackwell architecture—rather, it served as a foundational learning tool.
5nm Technology and the Reasons Behind the Delay
Braga, Microsoft’s next-generation AI chip, was planned to be manufactured using TSMC’s advanced 5nm process, promising better performance and efficiency. However, leaked information suggests the project has been postponed by at least six months, with a new launch window pushed to 2026. One possible reason for the delay could be design complexities that caused instability in early chip simulations. Microsoft reportedly sees this delay as an opportunity to ensure the final product is flawless.
Not Just AI: Microsoft’s Bold Vision in Quantum Computing
Microsoft’s hardware ambitions extend far beyond AI chips. The company is also making groundbreaking strides in quantum computing, which it considers a cornerstone of future computing technologies. This aligns with Microsoft's long-term vision and large-scale ambitions.
The Future of Technology: Majorana 1 and Topological Qubits
In a major breakthrough, Microsoft has developed a new material to create Majorana 1, a key component in its quantum computing efforts. Recently, the company announced that it successfully created and verified a topological qubit using Majorana 1. This achievement addresses one of the biggest barriers to practical quantum computers. Microsoft’s ultimate goal is to build a quantum computer with one million qubits on a single chip. A chip with a million qubits would deliver computational power far beyond today’s systems and could usher in a new era in science and technology. This not only highlights Microsoft’s ambition in AI, but also underscores its serious investment in quantum chip technology.
