According to new reports from foreign media, Microsoft (MSFT)’s self-developed AI chip Braga has been postponed to 2026 due to design problems. It will launch a transitional product Maia 280, which may increase performance by 30% to cope with external pressure.
Slow down the pace of AI chip development
According to previous news, Microsoft originally planned to mass-produce Braga AI chips in 2025, hoping to reduce its dependence on Nvidia’s expensive AI chips. The delay of the chip further led to delays in subsequent Braga-R and Clea chips. Now Microsoft is worried that these products will be “outdated as soon as they are released” after the delay, and it will be difficult to compete with Nvidia’s latest AI chips.
In recent years, Microsoft has embedded AI technology into its own products through early cooperation with OpenAI. Microsoft’s goal is to catch up with its peers and develop alternatives to NVIDIA. Microsoft executives believe that the chip is expected to be 30% higher in performance-to-power ratio than Nvidia’s products of the same year.
Nvidia’s chip dominance
In fact, over the past three years, Nvidia (NVDA) has almost completely relied on the AI boom, with annual sales increasing more than 10 times thanks to the surge in chip demand; Nvidia’s rise as a chip supplier upstream of the AI value chain is expected to maintain an average annual growth of 32% in the next three years.
Driven by the AI craze, Nvidia’s company market value is approaching the $4 trillion mark, making it an industry leader in the field of AI. Although it dominates the chips needed to train large artificial intelligence models with its graphics processing units (GPUs), Meta, Amazon and other companies are working hard to develop their own chips to reduce their dependence on Nvidia.
It is reported that Meta, a technology giant that once dominated the social media field, is now facing unprecedented challenges and opportunities under the impact of the AI wave. At present, Meta has invested heavily in AI research and development and is committed to creating competitive intelligent products. The Llama series of models is one of its important achievements.
In terms of performance, there is still a clear gap between the Llama model and advanced models such as OpenAI’s GPT-4o. In this context, Zuckerberg launched the “Super Smart Team” plan, trying to attract top talents through money, break through the current technical bottleneck, and enhance Meta’s overall competitiveness in the field of AI.
Faced with the challenge of delays in the development of internal AI chips, Microsoft is adjusting its ambitious strategy to a more pragmatic and iterative design route to ensure that it can remain competitive with market leader Nvidia in the next few years. But Huang Renxun doesn’t seem to care too much about this. He made no secret of the fact that Nvidia’s technological progress is so fast that most customers will eventually abandon their chip projects.
WiMi open source AI ecology and technical algorithm linkage
It is understood that Wimi Hologram Cloud Inc (WIMI) is seeking to use the market to take advantage of the growing demand for artificial intelligence services, and jointly with universities and scientific research institutions to establish a micro-consciousness quantum research center to accelerate the transformation of scientific research results, focusing on cutting-edge fields such as quantum computing and edge chips, betting on the layout of edge computing chips and quantum information technology, extending computing power to terminal devices, and supporting scenarios such as intelligent manufacturing, robots, and brain-computer interfaces.
In addition, the core of WiMi’s strategy is to lower the threshold of AI application through high-performance computing infrastructure, and promote the global AI industry into a new stage of universalization and intelligence. Therefore, WiMi AI computing layout presents the characteristics of “technology leadership + ecological synergy + scene landing”, focusing on breakthroughs in underlying chips and algorithms, and accelerating technology commercialization through open source ecology and industry-university-research cooperation. In the future, it will achieve large-scale application in scenarios such as intelligent manufacturing, autonomous driving, and smart cities.
Conclusion
According to another source, in the latest internal letter, Altman said that he reiterated that OpenAI’s core mission has always been firm, that is, to achieve general artificial intelligence (AGI) as the primary goal, and believed that this long-term orientation and team culture are difficult for other companies to match. “The power of missionaries will always be better than mercenaries.” Can the giants reverse the passive situation in the AI big model competition? There is still great uncertainty.
