Nov 27, 2023

monsitj/iStock via Getty Images Morgan Stanley said that recent announcements from OpenAI and Microsoft (NASDAQ: MSFT ) signal that competition to own more of the emerging AI stack is heating up. An AI stack a collection of technologies, frameworks and tools used to develop and deploy AI applications. An AI stack usually includes multiple layers or components. On the back of Microsoft's Ignite conference and OpenAI's inaugural DevDay, the analysts said, that it was clear to them that the race to secure strategic positioning within the emerging AI application stack between model providers, the cloud hyperscalers and the broader infrastructure software ecosystem is heating up. They expect this theme will become more evident coming out of Amazon.com's ( AMZN ) Amazon Web Services, or AWS, re:Invent conference this week. The analysts added that much like the early innings of past compute cycles (virtualization, cloud computing, containers/Microservices), Generative AI is in a period of rapid experimentation and it will likely take time for the market to settle on a standard reference architecture for Gen AI applications. However, the analysts noted that, what is distinctive about this cycle is that the pace of innovation at the infrastructure layer has been rapid and the focus on abstraction — automating the workflows and tasks to build consequential applications is happening much earlier than they would otherwise expect. In this "trial and error" stage of the market, they expect a wave of competing approaches to proliferate as product velocity across the software ecosystem accelerates. As per this backdrop, the analysts think declaring clear "winners" or "losers" will be challenging for investors, but what seems more clear is that competition at the infrastructure layers will likely intensify in the near term. Given the recent announcements from Microsoft and OpenAI, the analysts believe that basic Retrieval Augmented Generation, or RAG, use cases could become more commoditized. This potentially limits the opportunity for pure play vector database players and database incumbents who have recently introduced vector support. While Microsoft and OpenAI are looking to subsume additional layers of the AI Stack outside of the foundational models, there will continue to be a need for an independent AI Stack which comprises of open-source software and select cloud infrastructure software providers, according to the analysts. Microsoft ( MSFT ): Microsoft has an Overweight rating and is the firm's top pick. The tech giant remains best positioned in Software to monetize Generative AI across infrastructure and applications, as per the analysts, as the company's ability to function as the 'picks and shovels' of AI through Azure AI Services and the associated data, orchestration, development services paired with its development of out-of-the-box Copilot applications differentiates the company from the pack. The analysts added that, with Microsoft looking to own more parts of the emerging 'AI Stack' and their chief information officer, or CIO, survey signaling a modest acceleration in IT spend into 2024 (supported by CIO's interest in Generative AI and Large Language Models with 66% of CIOs indicating these technologies are directly impacting IT investment priorities verus 56% of CIOs in Q2 and 45% in Q1) and hyperscalers remaining the favored vendors for CIOs in helping them to use these innovative technologies, the analysts remain convinced in their top pick and above-consensus estimates. MongoDB ( MDB ) has an overweight rating at Morgan Stanley. The analysts said that while they expect more competition for retrieval augmented generation use cases which may limit the vector search opportunity, MongoDB continues to be among the best positioned cloud infrastructure names in the emerging Generative AI stack, in their view. This reflects — emerging AI initiatives steer a renewed focus on data strategy and potentially standardization on a modern data platform; accelerated code generation and developer productivity should drive more applications on to MongoDB given the company's strong traction with modern application developers; and incremental monetization opportunities emerging in broader semantic search/keyword search, despite potentially more competition in RAG use cases. The analysts added that they see strong potential for positive estimate revisions in 2024 and beyond, on the back of an improving software spending environment and a new product cycle which comprise more than just vector search (such as stream processing, relational migrator, vector search). This should buttress a sustainable 25% growth profile. Snowflake ( SNOW ) also has an Overweight rating at the firm. The analysts continue to believe that structured/semi-structured data represents an integral part of the enterprise knowledge base, which will prove highly complimentary to the largely unstructured datasets that large language models, or LLMs, are trained on, placing Snowflake in a position to enrich LLM-based applications with up-to-date data. Snowflake's combination of Streamlit, Container Services, Snowpark, and Neeva could help the company emerge as a platform for customers to build Gen AI apps using this high-value data in the company. The analysts think that the breadth of its portfolio is important as the company has now reached a stage where developers (rather than business analysts) can program application in the relevant languages in Snowpark, using the required compute infrastructure (such as GPUs) through Container Services, with orchestration tools from Neeva, to make applications available internally and/or via Streamlit. In addition, the analysts noted that Snowflake's ( SNOW ) core positioning and the recent demonstration of AI product velocity gives us confidence that the company can capitalize on the growing secular trend. More near-term, they believe that Snowflake may be well positioned against mixed investor sentiment, given — the analysts are picking up an increasing focus on data strategy and cloud migrations as organizations prepare for AI initiatives, potentially offsetting the more pronounced focus on Snowflake optimizations evident during H1; and its favorable exposure to North America and Large Enterprises which appear to show healthier spending trends in H2 on a relative basis. Microsoft ( MSFT ) has a Hold rating at Seeking Alpha's Quant Rating system, which consistently beats the market. Meanwhile, the Seeking Alpha authors' average rating is more positive with a Buy and so is the average Wall Street analysts' rating, with a Strong Buy .