微软 Azure AI Foundry helps AI application development

At the annual Ignite Conference in Chicago, Microsoft announced the launch of "Azure AI Foundry," designed to solve the pain points of enterprises in AI application development:

1. Easily switch AI models：用户可在不同的大型语言模型（LLM）之间切换，例如从旧版OpenAI模型转移至新版，或选择Meta Platforms (META) 的工具。
2. Free tools：微软表示，此软体将免费提供，目的是吸引企业更广泛地采用微软的云端服务。
3. Improve return on investment：Azure AI Foundry帮助用户减少对新技术的适配成本，避免因技术更新而重建应用，确保业务稳定运行。

Currently, Azure AI has approximately 60,000 customers and offers 1,700 AI model options. Microsoft emphasized that Foundry can solve cumbersome problems that companies have in the technology update process. For example, developers do not want new models to bring instability and can find the model that is most suitable for the task among diverse options.

Chip innovation improves cloud performance

Microsoft launched two chips at the conference to showcase its hardware innovation capabilities:

1. secure microprocessor：保护加密和签章金钥，从2024年起将内建于所有新伺服器中。
2. Data Processing Unit (DPU)：类似于辉达(NVDA)的网路晶片，加快数据传输速度，提升AI任务处理效率。

These chips support Microsoft's needs for cloud computing, especially in the training and deployment of generative AI models. Rani Borkar, Microsoft's vice president of chip design, said the improvements are aimed at optimizing operating efficiency among servers, chips and software layer by layer.

In addition, Microsoft announced that its Maia AI chips will work together with Vitek chips to support Azure OpenAI services. This further shows that Microsoft is actively building its own hardware capabilities to avoid over-reliance on suppliers.

Growth opportunities in the generative AI market

Microsoft continues to invest in generative AI applications to attract more users:

1. 365 Copilot upgrade：新增自动化处理重复任务功能，提升生产力。
2. Teams Copilot enhancements：理解并回答有关投影片、图片和视觉内容的问题。
3. multi-language support：从2024年起，PowerPoint Copilot将支持40种语言翻译。

These features enhance the practicality of Microsoft products and further consolidate its market position in the field of generative AI.