The short version, and a caveat about Europe
Two stories are running side by side. Palantir keeps winning analyst support and announcing AI partnerships. Meanwhile one of Wall Street's lead aerospace and defence analysts is warning that good earnings may not rescue the sector's share prices.12 Our reading is that the market is paying up for AI adoption stories and is more cautious on traditional government-linked contractors.3
For a European business reader, we have to be plain about the limits of the evidence. Every company, analyst and event in the material behind this piece is American. The one item with a sovereignty angle is Regna AI, described as a venture to help governments, businesses and institutions deploy AI models in "sovereign environments." The record does not say where those customers are, and it names no European company or regulator.4 We will not invent a European link. What follows is the global picture, with the questions it raises for European holders of US technology and defence shares.
Palantir: strong opinions, a falling share price
The bullish calls are specific. Ryan Canady upgraded Palantir to Buy after concluding that customer-built AI applications can remain complementary to Palantir's platforms rather than displace its core offering.1 Bernard Zambonin remains constructive. He emphasises customer expansion and net dollar retention as central to the bull case, while acknowledging that the elevated forward valuation makes continued execution essential.1 In plain terms, net dollar retention measures whether existing customers spend more over time. The dossier does not name the firms behind either call.
The share price tells a more cautious story. On Friday, 16 October 2026, Palantir fell more than 7% to close at $186.03, down $14.47. It was the fifth lower close in a row, after a 52-week high of $212.94 earlier that week. The drop came during a broader technology sell-off.1 A stock that falls five sessions straight right after a record, while analysts are upgrading it, is showing something investors should note. Good news about the business is already reflected in a price that is hard to live up to. Zambonin's own caveat about valuation says as much.
The newsflow behind the optimism is real. In mid-October, Ondas and Palantir announced Regna AI, a standalone private company, with Palantir expected to provide model-training infrastructure and engineering.4 Customers Bank announced a collaboration with Palantir on an AI-oriented banking operating model. Its own forward-looking statements point to implementation risks, reliability concerns and contractual flexibility.5 Surf Air Mobility signed its fifth commercial OperatorOS customer, with software powered by Palantir.6 These are announcements of intent. The dossier contains no revenue figures for any of them.
A note on the Goldman Sachs upgrade
Our summary of the market narrative mentions a Goldman Sachs upgrade of Palantir.3 The dossier gives no analyst name, date or price target for it, so we cannot say more. What we can verify is Goldman's own financial record. Its earnings per share were $22.87 in fiscal 2023, $40.54 in 2024 and $51.32 in 2025.7 In the second quarter of 2026 they were $20.98, against $10.91 a year earlier, almost double.7 That says Goldman itself is doing well. It says nothing about whether its view on Palantir is right.
Defence: a sector under a cloud
J.P. Morgan's Seth Seifman, the bank's lead aerospace and defence analyst, expects a challenging third-quarter earnings season. He cites slowing airline traffic growth, uncertainty over US military spending and lingering supply-chain constraints.2 He also expects most companies to deliver solid results and express confidence in underlying demand, but says good earnings may not reverse the sector's stock-market weakness.2
The Leidos downgrade shows the mood. Seifman's reasoning is that Leidos' relatively low valuation is insufficient to justify an Overweight rating when other aerospace and defence stocks also offer substantial potential returns.2 He is not saying Leidos is a bad business. He is saying cheapness alone is not a reason to favour it. The same research note expects Howmet's 2026 earnings estimates to rise, with management confident about growth in aerospace and industrial gas turbines.2 The pessimism is therefore selective rather than sector-wide. Our narrative summary also records doubts about Boeing's 2027 cash flow, though the dossier gives no detail beyond that.3 In our entity map, BigBear.ai competes with Leidos, a reminder that AI-focused challengers are entering the government contracting space.3
Politics adds a dated risk. Congressional midterm elections on 3 November 2026 could change control of Congress and constrain military spending growth, and some uncertainty may fade after that date.8 For European readers, US defence budgets matter to those who hold American contractors, as many global index funds do. The dossier does not give any index weightings, so we will not guess how much of a typical portfolio this represents.
The speculative edge of the AI trade
The same market also rewards bets with little current revenue behind them. D-Wave Quantum fell about 10% to $17.64 after a 20% jump the previous day, and it is down 25% over the past month.9 Melissa Tucker's analysis puts it at a stretched 171 times its annual sales, with projected 2026 revenue near $35 million. She estimates fair value at $9.50 a share, implying roughly 40% downside, and cites a delayed technology roadmap.9 Our own arithmetic on the $17.64 close gives a larger gap, which suggests her figure was set against a different price. Treat the 40% as approximate. Marc Gerstein, by contrast, sees D-Wave's dual approach, spanning both annealing and gate-model quantum computing, as a competitive advantage over single-method peers.9 The two views are in direct conflict.
Elsewhere the optimism is more grounded in actual business. Nicholas Amicucci argues that Bloom Energy's ability to provide reliable, dispatchable power to a volatile demand profile differentiates it from competitors.10 On NuScale Power, Bank of America's Rinny Singh reiterated a buy rating with a $12 target, roughly 24% upside, even though the shares have fallen nearly 40% this year.11
How much to trust these sources
Our measurements of source reliability matter here. The news service behind the Palantir, Leidos and D-Wave articles had 57% of 881 checked claims hold up.129 The Bloom Energy article's source scored 45% of 520.10 The NuScale article, from Motley Fool, scored 22% of 18 checked claims, a small sample but a low result.11 We therefore treat the analyst names and direct claims as reported rather than independently confirmed, and we quote no figure from them that we could not trace to a document. The only canonical, fidelity-checked numbers in this piece are the Goldman figures from SEC filings.7
What to watch
- 3 November: the US midterms, the dated event most likely to change the outlook for defence spending.8
- Palantir's next sessions: whether the price recovers from $186.03 or the five-day slide extends, and whether the customer expansion and net dollar retention that analysts cite show up in reported numbers.1
- Defence results: whether Seifman's prediction holds, that solid earnings do not lift the sector.2
- Regna AI: any sign it names European or other non-US government customers, which would give the sovereign-AI theme the European dimension this dossier lacks.4
This is analysis of published reporting, not investment advice.
