Chips stumble but AI software shines on Palantir-Nvidia deal — long PLTR

Verdict: the thesis is real, but this trade as written has never made money — stand aside

The verdict here is unambiguous: this pair trade is not worth taking today, and the realized record argues against taking it at all. The strongest point for the idea is Palantir's fundamental acceleration — revenue up 18.6% sequentially to $1.94B, net margin at 54.9%, and free cash flow more than doubling to $2.09B in the quarter ended June 30, 2026 — plus the concrete Nvidia Sovereign-AI catalyst that drove a 9% pop per the July 1 Yahoo Finance report. The strongest point against is the completed backtest of this exact rule set: one trade over 60 months produced a -75.0% return with a 90.2% maximum drawdown and a 0% win rate, and the most recent 12 months still lost 20.9%. Right now the signal state is stand-aside: PLTR's 5-day rate of change is -5.2% versus the required reading above 5, its ADX of 4.2 sits far below the 25 threshold, the PANW leg's 5-day rate of change is -13.6% versus required above 3, and PLTR's price is just $3.71 above its 20-day EMA at $172.81 — a break below confirms the second exit leg. Insider filings for the period ended June 30, 2026 (deadlines passed, so dated disclosure) show net open-market selling at both names — roughly -$136.6M at PLTR and -$26.7M at PANW — and PANW's fundamentals deteriorated sharply in its April 30, 2026 quarter, swinging to a -$177M net loss with gross margin down to 67.6% while shares outstanding jumped 15.7%. The verdict would flip only if a materially reworked rule set — different momentum windows, stop logic, or leg construction — demonstrated positive held-out results, since the published rules as written have never made money in any tested period. **Conviction breakdown** - Thesis support: 70 — PLTR's fundamentals and the Nvidia catalyst genuinely back the software-over-chips story. - Trade readiness: 15 — all entry conditions are far from firing and exit legs are already live. - Risk quality: 20 — insider selling at both names, PANW dilution, and an SMH short leg that is 17.8% Nvidia, the same catalyst powering the long leg. - Backtest evidence: 10 — a single losing trade per window with a 90.2% max drawdown over 60 months. - Fundamentals trend: 45 — PLTR accelerating, but the PANW confirmation leg swung to a loss with 6 points of gross-margin compression.

Trade now: nothing to execute — the spread trigger is on the wrong side

Do nothing today. The strategy is a momentum pair setup — long PLTR (and optionally short SMH or long PANW leg rules) once PLTR's 5-day rate of change turns up above 5% — and right now it is pointing the wrong way. PLTR's 5-day ROC is **-5.2%**, about 10.2 percentage points below the entry threshold, and its 14-day ADX is **4.2**, far below the 25 trend-strength requirement. The PANW confirmation leg is also distant: its 5-day ROC is **-13.6%** versus the required reading above 3%. Only the ADX conditions on PANW (35.0) and SMH (31.7) are currently satisfied. Notably, the exit logic is already live in the opposite direction: PLTR's 5-day ROC of -5.2% is below the 1% exit threshold, and the last close of **$176.52** sits just $3.71 above its 20-day EMA of **$172.81**. In other words, the signal state is "stand aside," not "waiting at the door." If an entry did trigger, the risk plan is mechanical: a fixed stop at roughly **-2.4%** from entry and a take-profit at roughly **+4.8%**, an effective reward-to-risk of about **2:1**, with positions capped at 25% of the book. "Wait" means concretely: watch for PLTR's 5-day ROC to cross above 5 while ADX climbs above 25 and PANW's 5-day ROC recovers above 3 — then, and only then, take the entry with the 2.4% stop attached. One caution on the evidence itself: the completed backtest on this exact rule set traded once per window and lost money in all three windows tested — a **-75.0%** return with a **90.2%** maximum drawdown over 60 months, and **-20.9%** over the most recent 12 months. That history argues for respecting the wait state rather than anticipating the trigger, and for keeping position size at or below the 25% cap if the setup ever fires.

The software-over-chips rotation has real fundamental fuel

The idea's core rotation claim is well supported at the company level. Palantir's fundamentals are accelerating sharply: in the quarter ended June 30, 2026, revenue grew 18.6% sequentially to $1.94B, net income rose 22% to $1.06B, and the net margin expanded to 54.9% from 53.3% the prior quarter. Free cash flow more than doubled quarter-over-quarter to $2.09B, and operating margin reached 47.1%. On a sector-relative basis, Palantir sits in the top decile of Information Technology peers — the 97th percentile for operating margin, the 91st for gross margin (84.7% in the latest quarter), and the 98th for free cash flow. These are exactly the kind of numbers that justify treating PLTR as the software winner in a rotation trade. The catalyst is concrete, not just narrative. Per the July 1, 2026 Yahoo Finance report, Palantir jumped 9% on the NVIDIA Sovereign-AI deal — the very announcement the thesis is built around — while the same CNBC live-update coverage from July 2 noted the Nasdaq falling again as chipmakers struggled even as the Dow hit a record. That is precisely the divergence the pair trade is designed to capture: long the software beneficiary, short the semiconductor complex via SMH. The short leg also has a fundamental anchor of its own through the idea's third name, Palo Alto Networks.…

Scores

• Conviction score breakdown: 32
• Thesis support: 70
• Trade readiness: 15
• Risk quality: 20
• Backtest evidence: 10
• Fundamentals trend: 45

Watch items

• PLTR — PLTR ROC (5)
• PLTR — PLTR ADX (14)
• PANW — PANW ROC (5)
• PLTR — PLTR close vs EMA (20)
• PLTR — PLTR price vs resistance
• PANW — ROC (5) above 5
• PANW — ADX (14) above 25
• PANW — ROC (5) above 3