Federal Capability Statement · 2026
Nightbox LLC — Federal Brief
A single-construct dual-use gene therapy platform applicable to solid tumor oncology and host-directed medical countermeasures. US small business, AI-augmented R&D, scalable AAV9-deliverable payload.
Company Identifiers
| Legal name | Nightbox LLC |
| EIN | 39-4373044 |
| State of incorporation | Wyoming |
| Headquarters | Wyoming-incorporated · Los Angeles operations |
| Primary NAICS | 541714 — Research and Development in Biotechnology (except Nanobiotechnology) |
| Secondary NAICS | 541715, 325414, 541711 |
| Business size | Small business (1 employee, < $5M revenue) |
| Founder & CEO | Artem Shakin |
| Point of contact | artem@nightboxllc.com |
| Web | https://nightboxllc.com |
| SAM.gov status | Entity registration in progress (case INC-GSAFSD21007772) · UEI pending · CAGE code pending |
| Wikidata | Q139590659 |
| ORCID (Founder) | 0009-0006-0003-6806 |
Core Capability
NKG2D-LIF6: a 2,123 bp chimeric receptor-effector gene therapy construct combining the human NKG2D extracellular domain (binds MICA and MICB stress ligands selectively expressed on solid tumor cells) with elephant LIF6, the re-functionalized pseudogene driving the elephant cancer-resistance phenotype (Vazquez et al. 2018, Cell Reports).
Single-construct gene therapy delivery via AAV9 — bypasses the autologous cell-processing and manufacturing complexity of CAR-T entirely. Manufacturable at clinical-grade scale by existing CMOs without bespoke infrastructure.
Dual-Use Relevance
The NKG2D recognition mechanism is a generalizable host-directed platform for cellular stress states. Lead indication is solid tumor oncology; the recognition-plus-effector architecture is modular and re-targetable for:
- Post-exposure medical countermeasures against engineered biothreats
- Host-directed antiviral response (replacing pathogen-specific approaches)
- Rapid-deployment gene therapy for stockpileable medical countermeasures
- Combat trauma applications (recognition retargeted to damage-associated molecular patterns)
Differentiators
- Manufacturing scalability: single-payload AAV9 vs CAR-T autologous processing — material implications for Strategic National Stockpile
- Clean intellectual property: no prior art on the specific chimera; provisional patent in progress
- AI-augmented R&D: production Claude-based agent stack for construct design, freedom-to-operate analysis, and competitive intelligence — empirical case study in modern federal R&D operating model
- Solo founder, low overhead: Phase 1-enabling work proceeding on bootstrap capital, suited for SBIR Phase I scope
Current Status
| In silico validation | Complete (~99% TGI ceiling on CRC syngeneic models) |
| Freedom-to-operate | Complete; clean IP whitespace identified |
| Provisional patent | In progress |
| In vivo studies (3 syngeneic tumor models) | Q2 2026 |
| Pre-IND meeting target | H2 2026 |
| First-in-human (Phase 1) | 2027 |
Companion Capabilities — Full Portfolio
Beyond NKG2D-LIF6 lead asset, Nightbox operates a parallel AI-augmented R&D infrastructure portfolio, each piece of which is independently relevant to federal counterparties:
| Capability | Description | Federal Relevance |
|---|---|---|
| OMNIKILL (research-stage simulator) | Internal combination-protocol simulation engine. Used to explore multi-mechanism solid-tumor therapy designs over a patient-anatomical computational model. Outputs reflect the simulator's internal state, not preclinical or clinical efficacy. No in vivo, no clinical, no peer-reviewed data exists for any OMNIKILL-derived design. Treat all OMNIKILL numbers as engineering outputs of a research tool, not therapeutic claims. | Methodologically relevant to NIH NCI, ARPA-H, BARDA, and VA Research priority areas as an example of in-silico drug-combination exploration |
| Patient X DNA Engine | 609,424-SNP pharmacogenomics-locked dose prediction engine over an 85-compartment digital twin. Validated against MyHeritage raw genomic data. | NIH All of Us · VA Million Veteran Program · DoD MHS GENESIS |
| Claude Suffler | Real-time conversation intelligence — captures system audio via WASAPI loopback, transcribes locally with faster-whisper, generates context-aware response suggestions through a Claude inference loop. Operates fully offline-capable. | DoD briefing support · Intelligence community after-action capture · State Department interpreter augmentation |
| Nightbox Live Copilot (Bilingual) | Real-time bilingual EN/RU conversation copilot for international negotiations. Built on stealth-mode meeting overlay (invisible in screen share). Pre-loaded persona system for context-locked responses. | State Department · DoD international engagements · Treasury Foreign Asset Control · Intelligence community |
| NightboxOS (development) | Local-first AI inference layer (Ollama, Whisper) with optional cloud LLM access via unified policy. Designed with air-gappability in mind; no formal SCIF, FedRAMP, or NIST 800-171 certifications yet. | Aligned in design with DoD secure-compute requirements; certification track open |
| Pluely Stack (development) | Local-first AI runtime (Ollama + Whisper + bridge proxy). Designed for offline / air-gapped deployment; no outbound traffic required for inference. Not yet certified for any classified environment. | Aligned in design with CUI-handling and tactical-edge inference requirements; certification track open |
Strategic National Security Context
Federal investment in domestic biotech and AI infrastructure is materially below adversary baselines:
- 2,794 active cell-therapy clinical trials in China vs 700 in the United States — a four-to-one gap.
- CARsgen filed the world's first solid-tumor CAR-T New Drug Application in September 2025 in China — not the United States.
- Russia launched the EnteroMix mRNA cancer vaccine in June 2025.
- $138B in 2025 cell-therapy and biologics deals flowed from China to Western pharma — the United States is now a net importer of late-stage assets.
- Hong Kong 18A biotech IPOs raised $6.43B in 2025 (16x the 2024 total).
- Zero US solid-tumor CAR-Ts approved in 8 years of clinical trials.
Nightbox is structured to close this gap: US-incorporated small business, single-construct gene therapy bypassing CAR-T manufacturing bottleneck, AI-augmented R&D operating model that compresses the timeline from concept to pre-IND, dual-use architecture relevant across the medical countermeasure spectrum.
Target Federal Programs
Detailed solicitation map: /federal/sbir-roadmap — named PA/RFA numbers, deadlines, set-asides, program officers.
- NIH NCI SBIR Omnibus PAR-25-279 — Phase I gene therapy oncology, Sept 5 2026 cycle
- NIH NIAID SBIR PA-25-301 — host-directed medical countermeasures track, dual-use angle
- ARPA-H — Cancer Moonshot 2.0 + Resilient Systems portfolios; abstract in prep
- BARDA DRIVe — solid tumor immune modulation BAA-26-100-SOL-00007
- DARPA BTO — Programmable Pharmaceuticals successor; capability brief in prep
- In-Q-Tel — biosecurity dual-use submission via portfolio team
- DTRA · JPEO-CBRND · ASPR — Strategic National Stockpile gene-therapy stockpileability angle
- DIU · AFWERX · NavalX — AI-augmented R&D operating-model demos, OTA pathway
Differentiated References
- Vazquez et al. (2018), Cell Reports 24, 1765 — elephant LIF6 characterization
- Abegglen et al. (2015), JAMA 314, 1850 — comparative cancer resistance
- Sulak et al. (2016), eLife 5, e11994 — TP53 expansion and DNA damage response
- Bauer et al. (1999), Science 285, 727 — NKG2D-MICA characterization
- Internal preprint: nightboxllc.com/preprint
For further information, draft solicitation responses, or pre-IND collaboration discussions, contact artem@nightboxllc.com.