Palantir stock finished Friday at $209.05, up 5.17%. The sovereign AI opportunity offers a new reason to look beyond its U.S. business. Yet Palantir’s own accounts show how much work an overseas expansion still has to do: customers outside the United States supplied less than one-fifth of second-quarter revenue.
The useful question after the rally is whether deployments that keep data and computing under a customer’s control can broaden Palantir’s growth. Its October 1 partnership with Armada makes that proposition more concrete. It does not yet put a dollar value on the opportunity.
The price is the October 9 regular-session close at 4 p.m. EDT, in U.S. dollars, from Yahoo Finance via TECHi’s Palantir quote page. It rose $10.27 from Thursday’s $198.78 close; the provider does not specify its delay interval. U.S. cash equities are closed over the weekend. This is a dated closing observation, not a live weekend price.
| Customer location | Q2 2025, US$ million | Q2 2026, US$ million | Year-over-year growth |
|---|---|---|---|
| United States | 732.592 | 1,573.047 | 114.7% |
| Rest of world | 271.105 | 362.417 | 33.7% |
| Total | 1,003.697 | 1,935.464 | 92.8% |
The overseas business is growing—and losing share
Palantir’s second-quarter 2026 Form 10-Q reports $362.417 million of revenue from customers outside the U.S., against $271.105 million a year earlier. That is 33.7% growth, calculated by TECHi. U.S. revenue increased much faster, from $732.592 million to $1.573047 billion, or 114.7%.
The result is easy to miss in a headline about rapid company-wide growth. The overseas share fell from 27.0% to 18.7% even though overseas sales rose. Approximately 90.2% of the year-over-year increase in total revenue came from U.S. customers.
Those are customer-location figures, not a count of where servers sit. Palantir assigns geography using a customer’s headquarters or an agency’s location at the time of sale. A sovereign installation is therefore not automatically overseas revenue, and an overseas customer is not automatically buying a sovereign AI deployment.
Armada supplies a deployment route, not a revenue forecast
In the October 1 announcement, Armada described a combination of its modular data centers and infrastructure-management platform with Palantir’s software. The proposed appeal is that customers can retain control over their physical infrastructure, models and data, including configurations designed to operate without an external cloud connection.
That could matter to buyers who cannot send sensitive work to a conventional shared cloud. The partnership also addresses a practical obstacle: software is of limited use if a customer cannot get suitable computing capacity deployed where it needs it.
But the announcement provides no contract value, committed customer volume or incremental Palantir revenue. A faster route to deployment is commercially useful only if it produces paying installations and durable software spending. It would be premature to treat the partnership itself as a booked sales win.
What would count as financial progress?
For shareholders, the strongest evidence would combine three developments: named customer deployments, a disclosed commercial commitment and revenue growth that persists after installation.
A launch announcement answers whether the product can be offered. A customer contract answers whether someone will pay. Reported revenue answers how much of that commitment has reached the income statement. These are different steps, and the stock can move before all three are visible.
Palantir’s filing adds another complication. Many contracts allow customers to terminate for convenience. Its remaining performance obligations cover noncancelable contracted revenue, with certain exclusions, rather than every potential dollar in its sales pipeline. At June 30, those obligations were $4.9 billion; the company expected about 43% to become revenue in the following 12 months.
That means neither an announced partnership nor a maximum contract ceiling can be dropped straight into a quarterly earnings estimate. Investors following the next report through TECHi’s PLTR earnings page should look for actual conversion, alongside the geographic revenue disclosure.
The strongest argument for the bulls
The overseas share does not need to rise for Palantir to keep growing quickly. A business expanding by one-third is healthy in isolation; its share can fall simply because the U.S. business is exceptional. Sovereign deployments can also serve American customers, so a geographic mix shift is an incomplete measure of success.
That is the credible counterargument to treating 18.7% as a weakness. It is a starting point for assessing diversification, not a verdict on product demand.
The risk is paying today for a much broader international franchise before its economics are disclosed. TECHi’s earlier October Palantir forecast examined how share-price gains had outpaced earnings-estimate increases. Friday’s rally makes the distinction between opportunity and recognized sales worth keeping in view.
A sustained acceleration outside the U.S., supported by paying deployments and clear contract terms, would strengthen the expansion case. More infrastructure partnerships without that evidence would leave the newest part of the investment story largely unmeasured.