Google Tops Forrester's First Ranking of AI Foundation Models
Forrester Research Inc. recently published its inaugural Wave ranking of artificial intelligence foundation models, and the results are quite surprising. According to the research firm, Google LLC emerged as the leader in this category by a significant margin. Other companies like Databricks Inc., Nvidia Corp., IBM Corp., and OpenAI LLC were recognized as "strong performers" in the ranking.
Despite facing some initial challenges with its consumer products, Google managed to secure the top position in the ranking. The report highlights Google's strong breadth and quality of enterprise offerings, especially its flagship large language model known as Gemini. According to the study's lead authors, Gemini is uniquely differentiated in the market, excelling in multimodality and context length capabilities.
Key Findings from the Report
Google's success in the AI market can be attributed to its enormous infrastructure capacity, a talented team of AI researchers, and a growing customer base in Google Cloud. The report evaluated AI foundation models across various categories such as data quality, multilingual capabilities, governance, scalability, innovation, and partner ecosystem.
Databricks received recognition for its hybrid strategy, offering customers pre-trained models and support for training their own models. Nvidia's Nemotron family of models stood out for its multilingual capabilities, although the company was advised to enhance its enterprise tooling. IBM's Granite models were praised for catering to enterprise needs with high quality and robust governance features.
OpenAI, considered a pioneer in the field, was recommended to improve its tooling for better enterprise adoption. While Amazon Web Services, Microsoft, Cohere, Anthropic, and Mistral were acknowledged for their model quality, there were suggestions for enhancing tooling and ecosystem support for enterprise users.
Future Trends in AI Foundation Models
Looking ahead, experts predict a drop in the cost of pretraining models, making them more accessible for various use cases. The competition in context window sizes and efforts to lower model training costs are expected to intensify. Multimodality, exemplified by Google's PaLM-E model, is anticipated to become more prevalent in the market, offering interactive and context-rich user experiences.
In conclusion, the Forrester report sheds light on the evolving landscape of AI foundation models and the strategic priorities for leading companies in the field. As the AI market continues to expand, companies will need to innovate and adapt to meet the evolving needs of customers and enterprises.
