Executive summary
A factory does not earn more because an AI model gives an impressive answer. It earns more when that answer helps avoid downtime, fulfil an order or use inventory better. The gap between a capable model and a useful business decision is where we see Palantir's opportunity.
We hold Palantir. What interests us most is how its software can become embedded in daily operations. Its Q2 2026 results give that idea substance: revenue growth was accompanied by a 47% GAAP operating margin, while existing customers accounted for much of the increase in global commercial revenue. The investment case still depends on useful deployments, competition and the price paid for the shares.
From company data to decisions
Imagine a manufacturer facing a delayed component. Before changing production, it needs to connect supplier information, inventory, customer commitments and the cost of alternative plans. An answer is more useful when the people making the decision can see how it relates to the rest of the business and act within the company's controls.
Palantir's Ontology (https://www.palantir.com/docs/foundry/ontology/overview) represents those real-world relationships in software. It connects data and models to objects such as equipment and orders, with permissions and actions that let people work within the organisation's controls.
The manufacturing example is illustrative. The commercial idea is straightforward: a useful recommendation becomes more valuable when the organisation can act on it safely. The work involves connecting a model's output to the data, relationships and decisions that shape an operating business.
Our reading is that this creates a potential source of staying power. As teams build workflows around shared data and business logic, replacing the system can become a larger operational project. That advantage must be earned through useful deployments, not assumed from the AI label.
Customer expansion gives the opportunity substance
For Q2 2026, Palantir reported (https://investors.palantir.com/reports-2026.html):
- Revenue of $1.94 billion, up 93% year on year.
- U.S. commercial revenue of $764 million, up 149% year on year.
- A 47% GAAP operating margin.
One detail stands out: $407 million of the $495 million year-on-year increase in Q2 global commercial revenue came from customers already present at the end of 2025. That points to deeper adoption within existing organisations, alongside new business.
This is the part of the result that connects most directly to the operating thesis. The opportunity extends beyond persuading a company to try an AI product. Continued expansion within customers gives us evidence to examine when assessing whether the software is becoming more useful to those organisations.
First-half 2026 operating cash flow (https://investors.palantir.com/reports-2026.html) reached $2.12 billion, compared with $850 million a year earlier. Billing and collection timing affect that figure, and it includes a $467 million non-cash stock-compensation add-back. We therefore watch cash generation alongside dilution, not in isolation.
Productive applications complete the AI investment case
NVIDIA supplies accelerated computing. Microsoft, Amazon and Alphabet provide cloud infrastructure and AI development platforms. Palantir focuses on connecting data, models and decisions inside organisations.
These roles overlap. Customers have alternatives, including building on their existing cloud platforms. Palantir's opportunity is to make the path from fragmented systems to dependable operations valuable enough that customers choose it and expand their use.
For our broader AI thesis, this adds a useful perspective: infrastructure investment ultimately needs productive applications. Software that helps businesses put that capacity to work is another part of the value chain worth understanding. The question is how effectively the software helps people carry out the decisions that matter to their business.
What we will watch next
Palantir's investor calendar (https://investors.palantir.com/reports-2026.html) lists Q3 results for 2 November 2026. We will focus on three questions:
- Customer expansion: Are customers expanding into more workflows?
- Operating results: Is that growth translating into durable operating profit and cash?
- Per-share progress: Is progress reaching shareholders on a per-share basis?
There are real risks. Deployments can be complex, competitors are capable, and a share price built on high expectations can fall even while the business grows. Position size and the price paid still matter.
The opportunity we are following is the move from experimenting with AI to relying on it in everyday operations. Palantir's recent results give that idea substance; continued execution will determine how far it can go.
You can explore our portfolio and investment approach on Razvan Ceapa's eToro profile, razvancd (https://www.etoro.com/people/razvancd).
Copy Trading is not investment advice | Capital at risk | Past performance does not guarantee future results