Maps of futures, drawn for builders.

1. Off-target safety & in-vivo specificity
2. Delivery beyond liver and blood
3. One-shot cures at <$5K, not $500K

CRISPR Tx · Verve · Vertex · IIT-Bombay / inStem groups

1. Lyophilised, non-cold-chain formulations
2. Therapeutic indications (oncology, autoimmune)
3. Affordable LNP IP and manufacturing

Moderna · BioNTech · Gennova · Bharat Biotech

1. Closing the loop from data to clinical action
2. FDA-grade signal from consumer hardware
3. Skin-tone-equitable optical sensing

1. Biologic CDMO capacity beyond Hyderabad
2. API self-sufficiency post-China
3. Quality control at scale

1. Workflow integration with PACS/HIS
2. Medico-legal liability for autonomous reads
3. Rural deployment without on-site radiologists

Qure.ai · Synapsica · 5C Network

1. Affordable NGS in tier-2 cities
2. Pan-Asian somatic variant databases
3. Trial-matching for rare variants

1. Viral-vector capacity
2. Closed-system cell-therapy automation
3. Cold chain for living drugs

1. Scale-up beyond pilot fermenters
2. Titre/yield vs petrochemicals
3. Regulatory path for engineered organisms

1. Clinical-safety boundaries
2. Vernacular & cultural localisation
3. Payer model for prevention, not crisis

1. Multi-organ coupling
2. FDA Modernization Act 2.0 acceptance pathways
3. Standardisation across vendors

1. Standardised aging clocks
2. Reimbursable indications
3. RCTs at population scale

1. Strain consistency batch-to-batch
2. Mechanism-based trials beyond FMT roulette
3. Cold-chain economics

1. Real-time genomic pipelines
2. Cross-border data-sharing protocols
3. Alert thresholds that actually move policy

1. Multi-scale physics from atoms to tissues
2. Compute (this is the real "AI for science" demand)
3. Validation across cell types

1. Immune compatibility
2. Supply of donor scaffolds
3. Consent ethics at scale

1. Surrogate endpoints regulators will accept
2. Access pricing
3. What we measure is what we get

1. Long-term safety of chronic implants
2. Ethics of cognitive enhancement
3. Consumer vs medical regulatory paths

1. Endpoint biomarkers regulators will accept
2. Long-duration safety on healthy populations
3. Equitable pricing if it actually works

1. 7-day biopsy-to-injection pipeline
2. Insurance for n-of-1 manufacturing
3. Combo protocols with checkpoint inhibitors

1. Lentiviral payload limits
2. Conditional kill-switches
3. Off-target T-cell expansion

1. PERV elimination at scale
2. Donor-pig facility biosecurity
3. Long-term immune monitoring

1. Adult-cell reprogramming without tumorigenesis
2. Spatial cues for organ topology
3. Ethical frame for "elective limb replacement"

1. Distributional ethics if it works
2. Biomarkers regulators will accept
3. Tolerance and withdrawal at scale

1. Ecosystem stability over years
2. Strain IP and provenance
3. Field conditions for live therapeutics

1. Containment and escape risk
2. International wildlife law
3. Success metrics beyond charismatic species

1. From sequence to function in living cells
2. Manufacturing designed candidates at therapeutic scale
3. Binders that actually work in vivo

1. Translating to clinical success rates
2. Causal target discovery (beyond correlative)
3. IP rights for AI-discovered molecules

1. Long-term electrode stability
2. Consumer ethics & informed consent
3. Signal-to-action latency

1. Cellular models that match in-vivo dynamics
2. Ethics review for 14-day+ synthetic embryos
3. Export-controlled IP and data

1. Multi-scale physics from atoms to behaviour
2. Experimental falsification at scale
3. Data-rights frameworks for digital twins

1. Long-context economics
2. Verifiable outputs (no hallucinated charts)
3. Vertical fine-tuning on tiny labelled data

1. Codebase-aware retrieval
2. Verifiable correctness
3. Security review of generated code

1. Permissions-aware retrieval
2. Multi-tenant cost ceilings
3. Evals that match production reality

1. Tokenisation for code-mixed Hinglish
2. Indic-language benchmark suites
3. Safety evals beyond English

AI4Bharat · Sarvam AI · Krutrim

1. Reliable tool-use across long horizons
2. Cost ceilings on autonomous runs
3. Human-in-the-loop UX that scales

1. Quantisation without accuracy collapse
2. Edge-to-cloud handoff
3. Vertical specialisation for narrow tasks

1. Distribution-faithful synthesis
2. Legal status of synthesis from copyrighted data
3. Benchmark contamination

1. Physical-accuracy benchmarks
2. Sim-to-real transfer
3. Long-horizon coherence

1. Containment and rollback
2. Online evaluations during training
3. Preventing reward-hacking at scale

1. Multi-tenant national-cloud architecture
2. Export controls and chip access
3. Talent retention

1. Closed-loop wet-lab integration
2. Verification of model proposals
3. Ownership of AI-discovered IP

1. Scaling sparse-autoencoder methods
2. Behavioural certification standards
3. Regulator-facing tooling

1. Deployment-time evals
2. Scalable oversight schemes
3. Independent red-team agencies & standards

1. Verifiable local processing
2. Battery & cost budgets
3. UX vocabulary beyond chat

1. Memory architecture beyond LoRA stacking
2. Selective unlearning
3. Verifiable identity over time

1. Operational definition of "deception"
2. Behavioural certificates
3. Oversight scaling to model speed

1. Closing the experiment loop
2. Ownership of discovered IP
3. Peer review for AI-authored science

1. Governance of autonomous spend
2. Long-horizon evaluations
3. Legal status of AI-run organisations

1. Definitions that survive their benchmarks
2. Safety case for deployment
3. Concentration of capability

1. Specification-gaming proofs
2. Interpretability that scales
3. Rollback in deployed systems

1. Reward functions that don't collapse to proxies
2. Safety of unconstrained search
3. Compute economics for sustained runs

1. Whose values they encode
2. Auditability of recommendations
3. Multi-stakeholder governance

1. From sequence to function in living cells
2. Manufacturing designed candidates at therapeutic scale
3. Binders that actually work in vivo

1. Translating to clinical success rates
2. Causal target discovery (beyond correlative)
3. IP rights for AI-discovered molecules

1. Long-term electrode stability
2. Consumer ethics & informed consent
3. Signal-to-action latency

1. Self-driving lab capex
2. Closing the synthesis loop
3. Benchmarks beyond known-class compounds

1. Sim-to-real that survives novelty
2. Safe failure modes
3. Data collection economies of scale

1. Open-world generalisation
2. Legal & insurance frameworks
3. Energy budgets per task

1. Cellular models that match in-vivo dynamics
2. Ethics review for 14-day+ synthetic embryos
3. Export-controlled IP and data

1. Verification of AI-designed circuits
2. Supply-chain security for novel architectures
3. Talent transition for human chip designers

1. Multi-scale physics from atoms to behaviour
2. Experimental falsification at scale
3. Data-rights frameworks for digital twins

1. Counter-drone for civil airspace
2. Rural last-mile economics
3. Long-endurance hydrogen UAVs

Garuda · ideaForge · ePlane · Skye Air

1. Sub-$15K cobots
2. Vision for unstructured environments
3. SME financing

1. Affordable EO data pipelines
2. Spectrum coordination
3. Debris liability frameworks

1. In-situ QA for safety-critical parts
2. Powder economics
3. Design tools that aren't bolted onto CAD

1. Manipulation in cluttered scenes
2. Safety certification near humans
3. Actuator cost curves

1. Highway-only L4 unit economics
2. Teleop handoff protocols
3. Insurance & liability frameworks

1. Ruggedised field deployment
2. Supply of NV-diamond and atom-trap parts
3. Killer apps (subsurface, MEG, navigation)

1. Yield ramp on first nodes
2. Fab-ops talent (not just design)
3. Water and power infrastructure

1. Error correction at scale
2. Usable software stacks
3. Problems that are actually quantum-advantaged

1. Process-temperature stability
2. Integration with electronics
3. Designer-facing tooling

1. Communication latency for remote ops
2. In-situ resource use (water ice)
3. Docking & berthing standards

1. Mil-grade qualification cycles
2. Export controls for dual-use
3. Cyber-hardened firmware

1. Error-corrected gate fidelities
2. Cryogenic infrastructure economics
3. Algorithm portfolios beyond Shor

1. Re-entry thermal management at high cadence
2. Spaceport regulation
3. Per-seat economics

1. Rendezvous & docking with small bodies
2. Return economics
3. Off-world resource law

1. Materials surviving 14 MeV neutrons
2. Tritium fuel cycles
3. Grid interconnect for first-of-kind plants

1. Continuous bulk manufacturing
2. Recycling of composite structures
3. Cost parity with steel

1. Engine reuse beyond ~50 flights
2. Ground-cryogenics throughput
3. Range-safety regulation

1. Regolith dust mitigation
2. Radiation shielding (water/regolith berms)
3. Closed-loop ECLSS over years

1. Capture mechanisms for tumbling targets
2. Commercial model - who pays
3. International debris liability

1. Return-vehicle cadence
2. In-orbit feedstock refining
3. Insurance for in-space accidents

1. Rendezvous with small bodies at industrial cadence
2. Return economics
3. Off-world resource law

1. Re-entry thermal management at high cadence
2. Spaceport regulation
3. Sonic-boom permissions over land

1. Stable cryo supply chains
2. Quantum-classical hybrid SDKs
3. Cryptography and chemistry migration

1. Self-driving lab capex
2. Closing the synthesis loop
3. Benchmarks beyond known-class compounds

1. Sim-to-real that survives novelty
2. Safe failure modes
3. Data collection economies of scale

1. Open-world generalisation
2. Legal & insurance frameworks
3. Energy budgets per task

1. Pyro vs hydrometallurgy economics
2. Reverse-logistics infrastructure
3. Rare-earth separation IP

1. Verification of AI-designed circuits
2. Supply-chain security for novel architectures
3. Talent transition for human chip designers

1. Microwave/laser power-transfer safety
2. Orbit-to-orbit assembly
3. Rectenna land-use and permitting

1. Land-use conflicts
2. Afternoon-peak alignment
3. Module-recycling chain

1. Domestic cell manufacturing (today: imports)
2. Fire-safety codes
3. Second-life cell markets

1. Battery-as-a-service models
2. Rural fast-charging
3. Thermal performance in 45 °C summers

1. Supply-chain Scope-3 fidelity
2. Satellite methane MRV
3. Registries that don't double-count credits

1. Electrolyzer manufacturing scale-up
2. Transport & storage
3. Off-takers willing to pay green premium

1. Cycle-life parity with Li
2. Prussian-white cathode supply chains
3. Cell form-factor standardisation

1. Dendrite suppression at fast charge
2. Manufacturing yield
3. Supply chain for sulfide electrolytes

1. Mounting structures cheap for 1-acre holdings
2. Farmer-finance models
3. Agronomy by crop and latitude

1. Hydrogen DRI economics for steel
2. Electrified kilns for cement
3. Carbon-capture retrofits

1. Solvent-regeneration energy
2. Long-term storage MRV
3. Off-takers paying for CO₂

1. $20/kWh capex targets
2. Round-trip efficiency vs cost
3. Right-sized for grid duty cycles

1. Sorbent durability at scale
2. Cheap-renewables geographies
3. Permanent-storage receipts

1. Field-trial regulatory clarity for edited seeds
2. Farmer advisory at scale
3. Weather-linked insurance products

1. Green premium adoption (govt procurement is the lever)
2. Standards for low-clinker cement
3. Capital intensity

1. Sub-100 kg/hr leak detection
2. Attribution to owners
3. Integration with regulatory action

1. Licensing first-of-kind designs
2. HALEU fuel supply
3. Community acceptance

1. Tritium breeding ratios
2. Net-electric (not just net-energy)
3. Commercial financing of first plants

1. Catalysts surviving intermittent operation
2. Capital intensity per kg
3. Offtake from shipping and aviation

1. Field-trial governance
2. International treaty scaffolding
3. Reversibility studies

1. Inertia services without spinning turbines
2. Markets that price flexibility
3. Transmission build-out at speed

1. Capex per ton at $50
2. Cheap-renewables geographic monopolies
3. Permanence MRV at billion-ton scale

1. Reprocessing safeguards
2. Proliferation-resistant fuel forms
3. Economics against SMR + fusion competition

1. Energy per kg vs field production
2. Crop variety beyond leafy greens
3. Integration with existing supply chains

1. First-of-kind plant financing
2. Tritium fuel-cycle economics
3. Competitive vs renewables + LDES

1. Reverse-logistics economics for low-value materials
2. Design-for-disassembly mandates
3. Trade rules for material flows

1. Governance with no global authority
2. Reversibility & exit strategy
3. Moral-hazard incentives on mitigation

1. Embrittlement-resistant steel
2. Offtake regulation across borders
3. Cost vs ammonia carriers

1. Pyro vs hydrometallurgy economics
2. Reverse-logistics infrastructure
3. Rare-earth separation IP

1. Microwave/laser power-transfer safety
2. Orbit-to-orbit assembly
3. Rectenna land-use and permitting