Sourcing: theme pointer per @wliang / @crux_capital_ — attributed, not independently verified; tape and financial claims below are our own (summaries + SEC EDGAR).

Thesis: Vishay Precision Group is the profitable precision-sensor arms-dealer for the humanoid build-out — force sensors for grip, strain gauges for joints, motion/weighing sensors for balance: the "boring infrastructure every robot needs." It's the picks-and-shovels expression of physical-ai-robotics ("own the layer that sells to every robot, not the moonshot"), and unlike the pre-rev humanoid de-SPACs it actually makes money. The catch: it's already had its re-rate.

Verdict: Trend-hold-eligible, conviction MEDIUM. VPG is a real, cash-generative arms-dealer in a confirmed uptrend (strong-up, golden cross, above all SMAs) — so it clears the durable-trend bar the humanoid moonshots fail. But it is extended (+249%/3m, ~156% above its SMA200) and the re-rate has run far ahead of the fundamentals (+17.6% FY2025 revenue grew +0.2%; +17.6% is Q1 2026's year-over-year rate — correction 2026-08-03), pricing in the humanoid optionality hard. Correction to the source premise: @wliang's flag called VPG "hasn't really run yet" — the tape says otherwise (it's roughly tripled in three months and sits −4% off its 52wk high). So this is buy strength on a pullback, not an undiscovered name — accumulate toward the SMA20, don't chase $142 after a triple.

The Story Right Now

The appeal is structural: precision sensors are a real, profitable, multi-decade industrial franchise (test & measurement, weighing, force) that also happens to be exactly what every humanoid needs at scale. That's the arms-dealer thesis in one name: you don't bet on which humanoid wins, you sell shovels to all of them.

The tape backs the trend but warns on the entry. Settled Mon 2026-06-29: price $142.09, RSI 63.3, trend strong-up, golden cross active (price above SMA20 $132.63 / SMA50 $101.40 / SMA200 $55.38). Returns: +5.6% over 7d, +12.6% over 30d, +249% over 3m, only −4.25% off the 52wk high; VWAP sits way back at $70.37. That +249%/3m on +17.6% YoY +0.2% filed annual revenue growth is the whole story — the stock re-rated ~14× faster than the business grew (correction 2026-08-03: the +17.6% was Q1 2026's quarterly YoY; the filed FY2025 grew +0.22% over FY2024, so the true re-rate-to-growth gap is far wider than the ~14× this line claimed — the thesis's own caution understated its case), which is the market pricing the humanoid-optionality narrative, not current earnings.

Setup

• Entry zone: Buy strength on a pullback toward the rising SMA20 (~$133) or a sideways flag — not a chase at $142 after a triple. RSI 63 isn't overbought yet, but ~156% above the SMA200 is extended; let it come in.
• Stop: Below the SMA50 (~$101) — a clean loss there breaks the trend structure. SMA200 (~$55) is the deeper line and far below.
• Target: Trend continuation; trail it, no fixed number. The re-rate means downside on a narrative-cool is real, so manage with the stop.
• Conviction: MEDIUM. Sizing: starter only, add on the pullback — do not build a full position into a +249%/3m extension.

Bull case

• Profitable picks-and-shovels. FY2025 revenue $307.2M, net income $5.3M, EPS $0.40, operating cash flow +$14.4M, $82.5M cash — a real, cash-generative business, not a pre-rev lottery. Rare in the humanoid basket.
• Sells to every robot. Force/strain/motion sensors are end-market-agnostic infrastructure; humanoid unit-scaling is a volume tailwind on top of the existing industrial/T&M base — optionality you're paid to wait on (the company earns money today regardless).
• Confirmed uptrend. Strong-up, golden cross, above all three SMAs — clears the trend-hold bar the moonshots (CCXI/SERV/RR) fail outright.
• Two independent analysts now cover it (@wliang, @crux_capital_) — the arms-dealer call is gaining coverage, not just our read.

Bear case

• The re-rate is the risk. +249%/3m vs +0.2% filed annual revenue growth (Q1 2026 ran +17.6%) = a multiple expansion pricing humanoid hype; if the humanoid narrative cools, a sensor company growing single-digits-at-best on the year re-rates back down hard. No margin of safety at −4% off highs.
• Thin margins. ~1.7% net margin ($5.3M on $307.2M) — a low-margin industrial supplier, not a high-incremental-margin compounder; humanoid revenue is still a small slice (it's mostly legacy industrial/T&M today).
• Insider trading reads bearish (leading-indicator) and the quarterly cadence is choppy — Q1 2026 printed a small EPS loss (−$0.02) and roughly breakeven operating cash flow.
• Extended entry. ~156% above the SMA200; chasing here is the opposite of the trend-hold "buy the confirmed dip" discipline.

Catalysts

• Next quarterly report — the key read on whether humanoid/sensor demand is showing up in the revenue line or whether the +17.6% quarterly rate (against a flat fiscal year) justifies the re-rated multiple.
• Humanoid-maker design-win / order disclosures — concrete sensor-content-per-robot or named-customer numbers would convert narrative optionality into modeled revenue.
• Insider-selling reversal — a flip from the current bearish insider read would de-risk the thesis.
• Trend integrity — holding the SMA20/50 keeps the structure; a clean loss of the SMA50 (~$101) is the stand-down signal.

Financials

Source: SEC EDGAR XBRL (FY2025 10-K + 2026Q1). A profitable, modestly-growing, cash-generative precision-sensor manufacturer — the financials confirm "real business," and also confirm the stock's move is a narrative re-rate well ahead of the fundamentals.

Sources

• Price truth: summaries (humanoid-robotics, settled Mon 2026-06-29) — $142.09, RSI 63.3, strong-up, golden cross; SMA20 $132.63 / SMA50 $101.40 / SMA200 $55.38; 7d +5.6%, 30d +12.6%, 3m +249%, −4.25% off 52wk high; VWAP $70.37.
• Financials: the desk's own tooling (SEC EDGAR XBRL) + leading VPG (composite Neutral; revenue growth Bullish +17.6% — a quarterly YoY rate; earnings Neutral 67% beat; insider Bearish). Theme parent: physical-ai-robotics perspective. Community context: Dumb Money #humanoid scrape (2026-06-29-dumbmoney-humanoid-quantum-scrape) — notably the community chases $BOT/Figure/Agility and never mentions VPG, the arms-dealer edge.