Does Palantir Control Its Own Fate?

Palantir (NASDAQ: PLTR) is often described as an artificial intelligence company, but it does not build chips, operate large-scale data centers, or develop leading-edge AI models. Instead, Palantir sells software platforms that help organizations integrate, organize, and use their data to make operational decisions.

Palantir’s value proposition is straightforward. Large institutions—particularly government agencies and defense organizations—generate vast amounts of fragmented data across multiple systems. This data is often difficult to access and use in real time. Palantir’s software connects these data sources, structures them into a usable format, and presents them through applications that allow users to monitor operations, run scenarios, and make decisions.

To understand where Palantir fits in the artificial intelligence market, it is useful to introduce the concept of the “AI stack.” The stack refers to the layered structure of technologies required to deliver AI capabilities, with each layer dependent on the one below it. At the top is infrastructure, which includes chips and cloud data centers that provide computing power. Next are AI models, which process data and generate outputs. At the bottom are applications, which translate those outputs into usable tools for end users.

Palantir operates at this bottom layer. It does not control the infrastructure or the models. It instead sits above them, integrating data and delivering decision-making software. This positioning has enabled the company’s growth, but it also makes it dependent on layers of the AI stack it does not control.

AI stack structure

As Chart 1 shows, the artificial intelligence stack is organized with infrastructure at the top, followed by models, and applications at the bottom. This order reflects where control resides within the ecosystem. Infrastructure and model providers determine the cost, performance, and availability of AI systems, while application-layer companies operate on top of those inputs.

Palantir delivers value through data integration and decision-making software, but it does so using infrastructure and models it does not control. As a result, its cost structure and performance are determined by external providers.

Chart 1. AI Stack Positioning – Infrastructure, Models, and Applications

Table 1 shows how the layers differ not only in function, but also in control. Infrastructure providers and model developers determine the cost, performance, and scalability of AI systems. Though application-layer companies can achieve high margins through software deployment, they remain dependent on access to compute and models.

Palantir’s reliance on hyperscaler infrastructure for compute and on third-party models allows the company to remain capital-efficient, but it limits its ability to control cost, performance, and product direction. Palantir cannot directly optimize the underlying systems on which its software depends.

The competitive challenge facing Palantir is not simply the presence of additional AI vendors, but a structural shift in how AI solutions are delivered. Hyperscalers such as Microsoft (NASDAQ: MSFT), Amazon (NASDAQ: AMZN), and Google (NASDAQ: GOOG) increasingly operate across all three layers, offering integrated solutions that combine infrastructure, models, and application capabilities. This allows them to integrate capabilities, bundle services for favorable pricing, and compete advantageously with application-layer providers.

Instead of selecting standalone software providers, customers may favor bundled platforms that offer end-to-end capabilities within a single ecosystem. As bundled solutions become more prevalent, sustaining premium pricing for standalone applications may become more challenging.

Commercial revenue raises the competitive stakes

Table 2 shows that commercial revenue has become PLTR’s primary growth driver, expanding at a significantly faster rate.  Government revenue grows at a steady but slower pace.

This shift highlights increasing adoption of Palantir’s platforms by enterprise customers. But this transition also increases exposure to competitive pressures, as commercial markets are more sensitive to pricing and platform integration.

Is the market catching on to the challenge?

As Chart 2 shows, Palantir’s share price trended strongly upward through most of 2025, consistently trading above its 200-day moving average, reflecting sustained investor confidence and excitement about everything AI. Entering early 2026, the stock began to converge downward towards its 200-day moving average, suggesting that investors are beginning to reassess expectations.

Chart 2. Palantir (PLTR) Stock Price versus 200-Day Moving Average (2025–2026)

Investor takeaway

Palantir has established a strong position within the application layer of the artificial intelligence ecosystem, particularly in government and defense markets. Its ability to translate complex data into actionable decisions remains a meaningful differentiator, and continued demand for AI-enabled software supports its growth trajectory.

However, the company’s position within the AI stack defines its limitations as much as its strengths. Palantir does not control the infrastructure or models that underpin its platform, leaving it dependent on external providers for both cost and capability. As hyperscalers expand across the full stack, they are increasingly able to bundle infrastructure, models, and applications into unified offerings, altering the competitive dynamics of the market.

Looking ahead, the key issue is not whether demand for AI software will continue to grow, but whether Palantir can sustain pricing power and differentiation while lacking control over the underlying stack. Moreover, the company’s growth is increasingly tied to commercial markets, where competition is more intense and integration advantages are more difficult to maintain.