Wells Fargo: The wave of open source models is sweeping the AI world, and Microsoft (MSFT.US) and ServiceNow (NOW.US) will be big winners

The Zhitong Finance App learned that Wells Fargo recently released a research report stating that the continued progress of the open source model is beneficial to large-scale existing platforms such as Microsoft (MSFT.US) and ServiceNow (NOW.US).

A team of analysts led by Michael Turrin said that open source model development is accelerating. Representative projects include Kimi K3 for Moonshot AI (Moonshot AI), GLM 5.2 for intelligent spectral AI (Z.ai), and DeepSeek V4 for DeepSeek (DeepSeek). It is expected that the US market will also heat up, and manufacturers such as Meta (META.US)'s “Thinking Machines” project and Nvidia (NVDA.US) are also worthy of attention.

The Turrin team stated, “More and more market participants are advocating the open source model, as evidenced by Nvidia's open letter (co-signed by Microsoft, Meta, Palantir, ServiceNow, and IBM) on July 24. Although cutting-edge laboratories in the US are more cautious, overall, the acceptance of the open source model is positive for the entire ecosystem, effectively reducing the cost of intelligence and expanding the scope of acquisition and use of AI.”

Analysts further explained that the industry is shifting from “competing for the number of tokens” to “competing for value output”. Their research results show that the open source model is favored by more and more users due to lower costs, weaker locking effects, and stronger security and controllability.

Customers are increasingly willing to move to an “just enough” open source model to complement expensive cutting-edge models and seek a more balanced trade-off between cost and performance. This trend shows that market preferences are gradually being skewed towards output quality and actual value, promoting cost curve optimization, and ultimately benefiting AI consumers.

The Turrin team believes that the evolution of the open source model is a positive sign for large existing software vendors because AI is naturally embedded in the workflow. “The reduction in inference costs means that the threshold for obtaining intelligence in familiar mature applications is lower, and the market risk of 'one model catch' is also reduced. A typical example is Microsoft Copilot, where increased intelligence provides added value to its already strong security and enterprise-grade positioning.”

Furthermore, the reduction in intelligent costs will also stimulate software development activities, boosting companies such as Datadog (DDOG.US), MongoDB (MDB.US), Atlassian (TEAM.US), and GitLab (GTLB.US), but will accelerate the squeezing of the living space of small to medium size players and segments.