Independent research labCoverage updated 2026-07-12
ConstraintAlpha maps the companies, components and capacity constraints behind technological growth.
Research and tools for investors following aerospace, AI infrastructure, robotics, advanced manufacturing and strategic materials.
Evidence-backed profiles · Theme-specific scores · Risks and invalidation
Current constraints radar
Selected live coverage areas. Counts and review dates come from the maintained research dataset.
Generation, transformers, switchgear, cooling, grid interconnection, and electrical construction required by AI-scale data centers.
Optical links, RF power, interconnects, radiation-tolerant electronics, ground infrastructure, and satellite systems.
Reducers, actuators, sensors, batteries, semiconductors, OEMs, and automation integrators across the humanoid supply chain.
HBM supply, 2.5D/3D integration, substrates, bonding, inspection, test, and packaging materials for AI compute.
Rare-earth separation, permanent magnets, specialty materials, precision motion, qualified fabrication, metrology and test.
Company intelligence Atlas
Each company is mapped to a specific constraint, scored in that theme, and documented with the strongest supporting evidence and the clearest risk.
VRT · NYSE
Controls / providesUPS, power distribution and liquid cooling
InterpretationVertiv integrates UPS, thermal management, power distribution, and service for high-density data centers, directly linking rack power growth to equipment content.
Principal riskCrowded expectations, customer concentration, component availability, and execution on rapid growth can create downside even if demand remains strong.
Open profile in AtlasRKLB · NASDAQ
Controls / provideslaunch / spacecraft manufacturing / satellite components
InterpretationInvestment Edge: Rocket Lab is a multi-hop space infrastructure platform rather than a single launch story, spanning launch cadence, spacecraft manufacturing, reaction wheels, star trackers, radios, solar, ground software and defense satellite work. Serenity lens: the bottleneck is credible but partially crowded because several layers are visible to the market already.
Principal riskNeutron execution, cash burn, valuation, and government contract concentration can overwhelm the component bottleneck thesis if launch or space-systems margins do not scale.
Open profile in Atlas6324.T · TSE
Controls / providesStrain-wave precision reducer
InterpretationIn the actuator and precision-motion lens, verified technical bottleneck candidate: HDS's strain-wave reducers address the compactness, torque and positional-accuracy requirements that matter in robot joints. The investable question is not whether the component is relevant; it is whether humanoid demand becomes material relative to HDS's established industrial-automation and semiconductor-equipment base. This membership is scored for joint performance, repeatability, and motion-component substitutability.
Principal riskThe business is exposed to manufacturers' capex cycles, particularly industrial robots and semiconductor/LCD equipment. Theme revenue remains undisclosed, and lower-cost Chinese reducer suppliers can pressure price or share as designs scale. For this subtheme, qualification, backlash, torque density, lifetime, and architecture substitution are the principal invalidation checks.
Open profile in AtlasBottleneck map
Two current research paths show how headline demand connects to less-visible public exposure.
Research layer
Public sector briefs introduce the system map. Verified readers can continue into complete company evidence, scores and risk notes.
AI data-center growth is ultimately an electrical-infrastructure build. New compute capacity needs generation, utility interconnection, substations, large power transformers, switchgear, busway, backup power, cooling, and skilled…
Space infrastructure depends on a specialized supply chain extending from materials and electronic components to satellite platforms, launch-related hardware, optical and radio-frequency links, and ground networks. Components…
AI accelerator output is constrained by a coupled memory-and-packaging system. HBM stacks require advanced DRAM processes and through-silicon-via assembly; logic and memory then need high-density 2.5D or 3D integration, large…
Methodology
Facts, interpretation and uncertainty remain visibly separate. Scores prioritize further research; they do not predict returns.
Read the methodology in contextIdentify the technological demand driver.
Map the physical and industrial system.
Locate the binding constraint.
Identify qualified suppliers.
Assess public-market exposure.
Document evidence and risks.
Update the thesis when evidence changes.
Public research process
Coverage data is generated from canonical company records and theme-specific memberships. Missing history stays visible as a product gap.
company-theme memberships across 10 live research views
canonical company identities; scores remain theme-specific
companies with at least one verified theme membership
No update-history field exists in the current production model. No change records are fabricated here.
Sector exploration
Each sector page combines an open research brief with a gated, filterable company directory.
Power generation, grid equipment, and electrical infrastructure for AI-scale demand.
26 mapped companies Explore sectorCritical components and subsystems powering general-purpose robots.
144 mapped companies Explore sectorPackaging capacity, high-bandwidth memory, and the materials enabling AI compute.
26 mapped companies Explore sectorFuel conversion, enrichment, fabrication, and hard-to-expand nuclear capacity.
29 mapped companies Explore sectorThe physical inputs, qualified processes, and measurement systems that make structural technology scalable.
33 mapped companies Explore sectorCompute, laboratory automation, and enabling infrastructure for AI-native biotech.
27 mapped companies Explore sectorComponents, networks, and ground systems enabling orbital infrastructure.
50 mapped companies Explore sectorBegin the research