Palantir, the largest position in both our AI and Cybersecurity portfolios, impressed again last night with revenue growth accelerating for the sixth quarter in a row to 36% (from 30%) and a FY25 revenue growth guidance of 31%, both figures coming 5-10% ahead of expectations. The strong top-line performance sparked another significant margin leverage, with the Q4 operating margin reaching 45% vs. 39% expected.
The company, initially recognized for its software products used by governments and the military worldwide, has evolved over the years into a leading AI player. It now offers a comprehensive AI platform that provides supply chain management, operational and maintenance planning, production optimization, and contract/invoice management across a wide array of industries.
This AI platform is rapidly gaining traction among both government and business customers, with Palantir’s impressive growth in its US commercial sector (+64% in Q4) signaling an accelerating trend in AI adoption by businesses. As the technology matures and costs decrease, this adoption is poised to continue, with potential for further upside.
The AI adoption momentum appears sustainable for two main reasons. First, Palantir’s success is largely attributed to its strategic use of “forward deployed engineers” – in-house consultants who help customers identify, experiment with, and implement AI use cases while offering crucial technical expertise. These engineers play an integral role in the “evangelization” of AI, guiding businesses through the adoption process. As other leading AI companies, including OpenAI, follow suit by deploying their own teams of “forward deployed engineers”, this evangelization process will likely accelerate, driving widespread AI app rollouts across industries.
Second, the recent performance improvements brought by DeepSeek’s R1 model – specifically at the inference level for AI applications (though its training-level improvements remain debatable) – are expected to lower the costs of running AI applications. This is expected to further fuel adoption and usage of AI solutions.
Looking ahead, 2025-2026 is set to be a pivotal period for AI at the enterprise level, marking a significant inflection point in enterprise software spending on AI. Until now, data science software vendors like Snowflake and Elastic have faced challenges, as most enterprises preferred to build AI applications in-house – taking advantage of widely available foundational models and datasets – rather than purchasing off-the-shelf solutions. It’s also likely that enterprise CTOs have been reluctant to commit to new AI software due to the influx of products, the time required to evaluate them, their high costs, and the rapid pace of technological advancements that could render early-generation AI solutions obsolete in a short period.
However, as AI technology matures and as software vendors and AI leaders deploy more “forward deployed engineers” to assist with AI adoption, many enterprises could soon begin transitioning from in-house development to adopting off-the-shelf AI products. This shift would be particularly beneficial for the revenue/earnings growth profile of data science software vendors that have built expansive platforms offering a broad range of storage and data preparation capabilities. These platforms notably enable enterprises to customize large language models based on proprietary data, perform advanced data searches, and carry out semantic mapping.
