Equity Research ● Microcaps ● May 2026

Robotics supply chain — picks & shovels across motion, sensing, embedded compute, and power.

A research scan of upstream component suppliers sitting beneath the headline themes of industrial automation, autonomous vehicles, humanoid robots, and defence robotics. Twelve names across the US, Canada, UK and EU.

01 / Executive summaryTop 5 conviction names, ranked

The screen targets enabling component suppliers — companies that own a defensible position in precision motion, sensing, embedded intelligence, or power infrastructure — rather than robot OEMs. The "bottleneck" test asked whether a robot manufacturer could reasonably build the part in-house, or whether the supplier holds protected know-how, IP, or fab relationships that create switching cost.

All five bypass the LiDAR-SPAC graveyard (negative gross margins, perpetual cash burn) and the humanoid-robot speculative list. Each owns a distinct upstream chokepoint — NPU IP, force-measurement physics, custom magnetics, edge compute modules, or defence-grade PCB sourcing — that a downstream robot integrator cannot trivially replicate.

02 / Sector tablesFour sub-themes, ranked within each

Notation. Direct microcap = under $500M. Borderline microcap = $500M–$1.2B. Small-cap exception = $1.2–3B, included only when the picks-and-shovels thesis is unambiguous and no smaller substitute exists.

2.1 Precision motion

The Western public-market universe in precision motion is genuinely thin. Allient is the cleanest profitable representative, but with robotics named as just one of several end-markets (industrial, vehicle, aerospace & defence, medical), it lacks the pure-play torque of Tokyo-listed Harmonic Drive Systems.

2.2 Sensing & vision

VPG is the most defensible name in the sensing cohort — strain-gage and load-cell physics with five decades of brand IP, and meaningful semiconductor-test exposure that compounds the robotics thesis with a separate growth driver. Basler offers profitable European exposure to machine vision at a structural discount to Cognex. Aeva and Innoviz remain optionality plays rather than core holdings; Sivers sits on a powerful technology base but with valuation that has front-run revenue scale.

2.3 Embedded intelligence

Ceva is the highest-quality name on the entire list — a software IP licensing model with 88% gross margin, expanding NPU and connectivity wins, and a royalty stream that should ramp through 2026–27 as NeuPro silicon ships. SECO is the most genuine "orphan" microcap on the report; trades at a low European tech multiple despite serving the same edge-AI-at-the-physical-edge thesis. Kontron is the safer, profitable but already-discovered version of the same idea.

2.4 Infrastructure / power

DSCV and NCAB are the cleanest profitable infrastructure plays. Methode and TTG are deep-value turnarounds with elevated execution risk. Kraken Robotics is included for completeness; the moat is genuine but the share price has compounded 40-fold since 2022 and now demands small-cap pricing.

03 / Detailed picksTop five conviction names

3.1 Ceva, Inc. NASDAQ : CEVA

• What they supply. Silicon and software IP for the smart edge — NeuPro neural processing units (NPUs), SensPro AI DSPs, Bluetooth/Wi-Fi/UWB connectivity IP, PentaG 5G modem IP. Licensed to semiconductor companies and OEMs that integrate the cores into their own SoCs.
• Picks-and-shovels fit. Royalty model on 2.1B+ Ceva-powered devices shipped in 2025; design-in across NXP, Microchip, BOS Semiconductors. Ceva is upstream of every robot SoC that needs efficient inference at the edge.
• Why potentially undervalued. 88% gross margin licensing-and-royalty model. 10 NeuPro NPU agreements signed in 2025, with multiple customers expected to have silicon by end of 2026 — meaningful royalty ramp not yet in numbers. Recent run from $20 to $32 still leaves the name well below previous cycle highs.
• Valuation. Gross margin ~88%; operating margin negative GAAP but improving; FY2026 revenue guidance 8–12% growth with significant margin expansion despite FX and supply headwinds.
• Latest revenue. Q4 2025 record $31.3M (+7% YoY); FY2025 $109.6M; FY2025 royalties $46.0M; licensing $63.6M.
• Risks. Royalty timing concentrated in licensee silicon dates; concentrated R&D spend; modest dilution from November 2025 follow-on offering (~$63M net).
• Thesis. The "ARM of NPUs" — a profitable IP business with dominant share in a rapidly expanding edge-AI inference market, where every robot, drone, wearable and industrial sensor needs efficient on-device compute.

3.2 Vishay Precision Group NYSE : VPG

• What they supply. Load cells, strain gages, force sensors, precision resistors, and complete weighing/measurement systems across the BLH Nobel, Celtron, Tedea-Huntleigh, KELK, Sensortronics, DTS and Pacific Instruments brands. Three operating segments: Sensors, Weighing Solutions, Measurement Systems.
• Picks-and-shovels fit. Force-torque sensing is essential for any compliant or contact-aware robot. VPG's sensors feed into industrial, medical, semiconductor test, aerospace, agriculture, and consumer markets — textbook robot-agnostic exposure.
• Why potentially undervalued. Stock has rallied from $18.90 (52-week low) to ~$64, but still trades at modest revenue multiples for a precision-instruments franchise with five decades of brand IP. Semiconductor-test exposure adds a separate growth lever.
• Valuation. P/S ~2.8×; 13.3M shares outstanding; revenue $307M TTM; EPS $0.40.
• Latest revenue. FY2025 ~$307M; Q4 2025 adjusted EPS $0.07 (vs. $0.21 consensus); FY26 guide mid-to-high single-digit growth.
• Risks. Industrial cyclicality; legacy steel-mill exposure (KELK); Q4 2025 EPS miss; small free float.
• Thesis. A profitable, global precision-sensing franchise with structural exposure to robotics, semiconductor manufacturing, and aerospace — currently re-rating but still below long-term peer multiples.

3.3 discoverIE Group LSE : DSCV

• What they supply. Customer-specific magnetic and power components, embedded computing, interface controls, and sensing-and-connectivity electronics for industrial OEMs across renewables, transportation, security, medical, and industrial automation.
• Picks-and-shovels fit. Designs application-specific components that get embedded into customer products — high design-in retention, low single-customer concentration risk.
• Why potentially undervalued. Sell-side targets cluster GBX 750–1000 versus a spot share price of ~615p. Profitable, dividend-paying, and trading well below long-term valuation multiples after a cyclical industrial downturn.
• Valuation. P/E 25.08; trailing 12-month revenue $559M; TTM EPS GBX 26.0; net margin 3.55%.
• Latest revenue. H1 fiscal 2026 record profitability; Q1 FY26 EPS GBX 19.50.
• Risks. Industrial-cycle exposure; M&A integration risk (active acquirer); possible FX drag from European reporting currency.
• Thesis. A profitable, design-in specialist-electronics roll-up serving the same end-markets as global robotics integrators, trading at a meaningful discount to North American comparables.

3.4 SECO S.p.A. Borsa Italiana : IOT

• What they supply. Computer-on-module (COM) hardware, single-board computers, human-machine interfaces, and the Clea IoT software suite for industrial automation, transportation, aerospace & defence, medical, and smart-device applications.
• Picks-and-shovels fit. COMs are the modular compute heart of industrial controllers and autonomous systems. SECO supplies dozens of OEMs across multiple verticals — robot-agnostic by design.
• Why potentially undervalued. True microcap at $335M after stock sold off from €3.61 to €2.34. Genuine "orphan" — limited US analyst coverage, listed only on Borsa Italiana. Q1 2026 revenue guidance ~€50M with modest growth signals stabilisation.
• Valuation. P/E ~21; revenue $204M TTM; EPS €0.08 forward consensus.
• Latest revenue. FY2025 results delivered 23 Mar 2026; Q1 2026 revenue guide near €50M.
• Risks. German economic weakness (key end-market); memory price volatility; geopolitical tensions in Middle East and Ukraine affecting operations; small free float and low US visibility.
• Thesis. A genuinely orphaned European microcap edge-compute supplier with profitable underlying business, trading at trough valuations after a cyclical industrial downturn.

3.5 NCAB Group AB Nasdaq Stockholm : NCAB

• What they supply. Printed circuit boards sourced via an asset-light model from a network of ~31 partner factories — multi-layer, HDI, RF, flex, rigid-flex, heavy-copper boards. Serves aerospace, defence, automotive, datacom, industrial, medical, power/energy, railway.
• Picks-and-shovels fit. Every robot needs PCBs. NCAB's high-mix-low-volume specialisation targets exactly the niche-equipment-maker segment that humanoid and industrial robotics startups occupy.
• Why potentially undervalued. Q1 2026 order intake +27% in SEK / +49% in USD; AI data-centre support orders >$20M; CMMC 2.0 certification opens the US defence market. Asset-light model produces structurally higher returns on capital than vertically-integrated PCB rivals.
• Valuation. P/E ~49 (TTM); EBITA margin 11.9% (Q1 2026); 11.9% Q1 EBITA margin.
• Latest revenue. Q1 2026 SEK 1,073.7M (+12% SEK / +31% USD); FY2025 ~$361M.
• Risks. China sourcing concentration ~75% (declining from 95%); FX headwind affected EBITA margin by ~150 bps in Q1; rising PCB material costs.
• Thesis. A specialist Nordic PCB sourcing platform riding the simultaneous tailwinds of defence reshoring, AI data-centre buildout, and broad industrial automation — with a balance-sheet-light model that compounds.

04 / WatchlistBorderline names above the cap threshold

These names sit in critical robotics-supply-chain bottlenecks but exceed the $1.2B primary cap. Listed for context, not as primary recommendations.

05 / Special situation flagSub-$100M with Fortune-500 contract

No name on the screen currently meets a clean "<$100M market cap WITH a Fortune-500 production contract AND positive gross margin" threshold as of May 2026. The closest candidate is Innoviz Technologies (INVZ), which holds Tier-1 series-production positions with BMW, Volkswagen Group, and Daimler Truck/Torc Robotics. However:

• Market cap fluctuates between ~$190M and ~$400M depending on share price
• Received Nasdaq listing-rule non-compliance letter on 25 March 2026 (closing bid below $1.00 for 30 consecutive sessions)
• Liquidity ~$72M against operating expenses ~$80.6M creates compressed runway

Innoviz is therefore flagged as a Special Situation with elevated execution and dilution risk, not a conviction long. The thesis path requires three sequential dependencies: (a) cure of the bid-price deficiency, (b) revenue acceleration to $80M+ in 2026, and (c) successful InnovizThree launch. Position sizing should reflect the binary nature of the listing-compliance overhang.

06 / 2026 macro outlookReshoring, rates, and the European discount

Industrial robotics demand

Roland Berger characterises 2026 as the first year of renewed growth momentum after two flat years, with potential CAGR of approximately 9% through 2030 driven by catch-up capex. Mordor Intelligence sizes the industrial-robotics market at $54.3B in 2026 growing to $94.4B by 2031 (11.7% CAGR). The IFR projects 2025 installed-robot units increased modestly with stronger acceleration in 2026.

Reshoring & policy tailwind

US announced reshoring jobs reached ~244,000 in 2024 (cumulative >1.7M since 2010). The CHIPS Act, IRA, and Build America provisions continue to drive multi-billion-dollar capex in semiconductors, EV batteries, and grid modernisation — all automation-intensive. Section 301 duties (25% on Chinese machinery imports) remain through 2026, narrowing the offshoring cost gap. This disproportionately supports North American sensor, motion-control, and embedded-computing suppliers.

Interest rate environment

A modestly easing rate trajectory through 2026 lowers the discount rate applied to long-duration capex projects, accelerating automation paybacks for mid-sized manufacturers — historically the segment most rate-sensitive in robot purchasing decisions. This is an underappreciated tailwind for the smaller profitable suppliers (DSCV, VPG, NCAB, ALNT) that depend on broad SME factory automation rather than headline humanoid deployments.

European context

Germany leads the EU industrial-robot market (~$1.0B in 2025, 7.4% growth), followed by the UK and Italy. European listings remain structurally cheaper than US comparables — meaningful for valuation arbitrage in DSCV, Basler, Kontron, and SECO. The risks (German manufacturing weakness, Eurozone fiscal constraint) are largely embedded in price.

Key tail risks

• A deeper Chinese deflationary export wave compressing component pricing
• US tariff escalation disrupting cross-border supply chains for European listees with US revenue exposure
• The humanoid-robot capital-formation cycle stalling, which would depress sentiment-driven valuations across the entire ecosystem

07 / Methodology & caveatsHonest limitations of this report

• Quality gates applied: gross margin >30% (or credible bridge in 2026E); positive operating cash flow OR liquidity runway >24 months; customer diversity across at least two end-markets; explicit avoidance of consumer-robotics brands, single-OEM-dependent suppliers, and going-concern flags.
• Microcap data is messy. Several aggregators (PitchBook, Yahoo Finance, Investing.com, Stocktitan, Morningstar) reported different market-cap figures for the same name on the same day, sometimes off by 5–10%. Where possible the report cites a range.
• Profitability vs. growth tension. The most exciting growth stories (CEVA, AEVA, INVZ, SECO, SIVE) are still GAAP loss-making. The cleanest profitability profiles (DSCV, VPG, NCAB, ALNT) are mature compounders at small-cap rather than microcap valuations. There is no current Western-listed name that combines microcap valuation, >30% gross margins, GAAP profitability, and a humanoid-robot growth narrative — that tension is itself useful intelligence.
• Structural blind spot. The most defensible bottlenecks (Harmonic Drive Systems' strain-wave gears, Keyence's machine vision, FANUC's CNC controls) sit on Asian exchanges outside the requested geographic scope. The Western-listed universe is genuinely thinner than the prompt's framing implies.
• Confidence scores 1 (lowest) to 5 (highest) reflect data quality and thesis robustness, not return probability.
• Liquidity: SECO, TT Electronics, Methode Electronics, and Innoviz can produce material slippage on size; trade with limit orders only.
• Dilution risk: Innoviz has a Nasdaq compliance overhang likely to require corporate action; CEVA completed a ~$63M follow-on in November 2025; Sivers has serial dilution history.
• No position recommendations are made. This is a research scan; sizing, entry, and risk management are the reader's responsibility.