# Super Intelligence Language in SBIR/STTR Grant Proposals: Tech Founder Guide

Executive Order 14434 directs federal agencies to substitute Super Intelligence for artificial intelligence across non-statutory solicitations. For deep-tech founders preparing SBIR and STTR Phase I proposals, Section 3(a) anchors agency terminology to technical definitions in 15 U.S.C. 9401(3). Winning applications match solicitation phrasing while articulating rigorous technical feasibility, verifiable milestones, and commercial potential without speculative hype.

Source: https://fast.io/resources/super-intelligence-language-sbir-proposals/
Author: [Tom Langridge](https://fast.io/authors/tom-langridge/)
Last reviewed: 2026-10-09

## What Executive Order 14434 Means for SBIR and STTR Applicants

Under Executive Order 14434, new Small Business Innovation Research (SBIR) and Small Business Technology Transfer (STTR) solicitations may say "Super Intelligence" where they used to say AI, while evaluators continue to score proposals on technical feasibility and commercial potential. The two mistakes to avoid are filling a proposal with speculative claims and ignoring the solicitation's own wording.

The administrative mandate took effect on September 29, 2026, when the White House issued [Executive Order 14434](https://www.whitehouse.gov/presidential-actions/2026/09/inaugurating-the-era-of-super-intelligence/), titled "Inaugurating the Era of Super Intelligence." Under Section 2(a) of the order, executive departments and agencies are directed, to the maximum extent permitted by law, to use "Super Intelligence" and the abbreviation "SI" in place of "Artificial Intelligence" and "AI" across official correspondence, public communications, websites, reports, policy documents, and other non-statutory documents within the executive branch. Section 1 establishes an administrative policy that executive agencies will not acknowledge earlier terms in applicable settings.

For small business founders and proposal coordinators, the bare abbreviation SI introduces immediate communication hazards. SI also has other meanings, such as the International System of Units. To prevent ambiguity during technical evaluations and administrative triage, grant writers should spell out "Super Intelligence (SI)" on first appearance across project abstracts, specific aims, and technical narrative volumes.

America's Seed Fund is coordinated by the Small Business Administration across 11 participating federal agencies that fund innovations through the SBIR and STTR programs. These agencies obligate significant non-dilutive capital to small businesses each year, requiring applicants to balance administrative responsiveness with technical accuracy.

The legal standard governing this vocabulary is set forth in Section 3(a) of Executive Order 14434:

"For purposes of this order, and except where otherwise provided by law, the terms 'Super Intelligence' and 'SI' mean the technologies and systems encompassed by the term 'artificial intelligence' as defined in section 9401(3) of title 15, United States Code."

Under Section 3(a) of Executive Order 14434, federal Super Intelligence encompasses the technologies defined as artificial intelligence in 15 U.S.C. 9401(3). That statutory provision, [15 U.S.C. 9401(3)](https://www.law.cornell.edu/uscode/text/15/9401), was enacted by Congress in the National Artificial Intelligence Initiative Act of 2020:

"The term 'artificial intelligence' means a machine-based system that can, for a given set of human-defined objectives, make predictions, recommendations or decisions influencing real or virtual environments."

Because Section 3(a) incorporates 15 U.S.C. 9401(3) directly, federal Super Intelligence and artificial intelligence name the exact same systems in federal usage today. The statutory definition focuses entirely on functional capabilities, such as predictive models, automated decision support, computer vision, and natural language processing. It establishes no cognitive autonomy benchmark or superhuman threshold.

This statutory grounding separates federal policy from commercial computing concepts:

- Federal Super Intelligence (two capitalized words, abbreviated as SI) is an administrative label for existing operational software and computational pipelines meeting 15 U.S.C. 9401(3).

- Research superintelligence (one compound word, commonly abbreviated as ASI for artificial superintelligence) is a hypothetical future system exceeding human expert capabilities across virtually all domains, which does not exist today. Commercial research laboratories, including Meta Superintelligence Labs, Microsoft's MAI Superintelligence Team, and Safe Superintelligence Inc., identify artificial superintelligence as an aspirational research target.

Under Section 3(b), the Assistant to the President for Science and Technology submits proposed legislative language for a federal definition of Super Intelligence to the President within 60 days of the order, which falls on November 28, 2026. That is an advisory proposal to the President, not an enacted statutory definition. As of October 2026, no statutory definition of Super Intelligence exists. What the proposal says, and whether anything will be enacted by Congress, remains unknown.

Small business leaders must also observe the legal boundaries established in Section 4(c) of Executive Order 14434. The text provides that the order creates no right or benefit, substantive or procedural, enforceable at law or in equity by any party against the United States, its agencies, officers, or any other person. The order directs executive agencies in their internal administrative practices, not private technology startups or commercial contractors.

## How SBIR Solicitations from DoD, NSF, NIH, and DOE Use Super Intelligence Phrasing

Under Section 2(a), new agency solicitations may use Super Intelligence in place of AI. Because each participating agency operates under distinct mission priorities, submission portals, and scoring rubrics, founders must understand how agency program managers translate Executive Order 14434 into funding topics.

### Department of Defense (DoD)

The Department of Defense awards SBIR and STTR contracts through the Defense SBIR/STTR Innovation Portal (DSIP) across components including DARPA, AFWERX (Department of the Air Force), the Army Applied SBIR Program, the Department of the Navy, and SpaceWERX. 

New DoD topics on autonomous systems, sensor fusion, electronic warfare and mission planning may use Super Intelligence terminology. DoD evaluators look for concrete operational utility: low size, weight, and power (SWaP) edge compute, real-time inference latency, cyber hardening, and adversarial robustness. Claiming that a system possesses general cognitive autonomy will trigger skepticism from defense technical evaluators. Founders should mirror "Super Intelligence (SI)" in executive summaries and topic responsiveness sections while grounding technical proposals in verifiable mathematical models, quantized neural networks, and explicit hardware execution targets.

### National Science Foundation (NSF)

The National Science Foundation runs its SBIR and STTR programs through the Directorate for Technology, Innovation and Partnerships (TIP). NSF proposals are submitted through Research.gov and evaluated on Intellectual Merit and Broader Impacts, along with any criteria the solicitation adds.

NSF already uses SI on its own pages, including its America's Seed Fund topic for SI systems and SI-based hardware. NSF SBIR panels look for a sound technical approach and a credible path to market.

### National Institutes of Health (NIH)

The National Institutes of Health awards SBIR and STTR grants through Grants.gov and eRA Commons across institutes such as the National Library of Medicine (NLM), the National Cancer Institute (NCI), and the National Human Genome Research Institute (NHGRI). Solicitations appear as Notices of Funding Opportunities (NOFOs) published in the NIH Guide for Grants and Contracts.

NIH applications are reviewed by Scientific Review Groups (SRGs), commonly known as study sections, on five core criteria: Significance, Investigators, Innovation, Approach, and Environment. Biomedical reviewers demand rigorous experimental controls, clinical validation data, patient privacy protections, and diagnostic interpretability. If a proposal claims speculative superintelligence, reviewers will downgrade the scientific premise. Founders should mirror the agency's Super Intelligence terminology in the Specific Aims and Project Summary, while specifying exact algorithmic pipelines, training datasets, and validation protocols in the Research Strategy.

### Department of Energy (DOE)

The Department of Energy Office of Science and Advanced Research Projects Agency - Energy (ARPA-E) award SBIR/STTR funding through the Portfolio Analysis and Management System (PAMS) and Grants.gov. Topics focus on scientific machine learning, physics-informed neural networks (PINNs), materials discovery, fusion energy controls, and electrical grid optimization.

DOE reviewers evaluate Phase I proposals on the strength of the scientific and technical approach, the ability to carry out the project, and the likely impact, including commercial potential. Founders should highlight distributed training efficiency, GPU memory management, and empirical validation against physical simulations while linking the computational pipeline to 15 U.S.C. 9401(3).

The following table summarizes how major federal agencies implement Super Intelligence terminology across small business funding solicitations:

| Federal Department | Primary Submission Portal | Representative AI Topics | Primary Evaluation Focus | Recommended Terminology Action |
| :--- | :--- | :--- | :--- | :--- |
| Department of Defense (DoD) | DSIP | Edge computing, autonomy, sensor fusion | Mission utility, SWaP constraints, operational feasibility | Mirror SI in topic compliance; specify exact model weights and hardware bounds |
| National Science Foundation (NSF) | Research.gov | Foundational algorithms, automated reasoning | Intellectual Merit, Broader Impacts, theoretical novelty | Mirror SI in Project Summary; provide mathematical proofs in Project Description |
| National Institutes of Health (NIH) | eRA Commons / Grants.gov | Clinical predictive models, diagnostic imaging | Significance, Approach, clinical safety, interpretability | Mirror SI in Specific Aims; ground Approach in validated biomedical pipelines |
| Department of Energy (DOE) | PAMS / Grants.gov | Scientific computing, grid optimization, PINNs | Technical approach, ability to carry out the project, impact | Reference SI in project goals; detail compute budgets and simulation baselines |

## Four-Point Roadmap for Articulating Technical Feasibility and Milestones

Generic grant writing guides fail to address the specific Phase I feasibility and commercialization rubric required by SBIR program managers when evaluating deep-tech artificial intelligence innovations. In Phase I, agency reviewers are not looking for marketing brochures or speculative vision statements. Evaluators seek a rigorous, de-risked work plan that proves technical feasibility within constrained budgets and tight timelines.

Founders preparing small business proposals can follow a four-point roadmap to structure technical narratives that satisfy federal review panels:

### 1. Deconstruct Agency Topic Descriptions

Every SBIR solicitation topic represents an unmet agency operational need or technical gap. Founders must parse the topic description to identify explicit technical metrics, target operational environments, and required deliverables. 

Topic descriptions distinguish between direct, narrow technical solicitations and open-topic mechanisms. In narrow agency topics (common in DoD components and NASA), program managers list precise target metrics, such as inference speed, target operating systems, or specific input data formats. In open topics (such as AFWERX open topics or NSF broad areas), the applicant must articulate the defense or commercial problem with equal precision.

Founders should construct a bidirectional requirements matrix matching every topic requirement to an explicit section of the proposal narrative. If the topic uses Super Intelligence, mirroring that term in the executive summary, project abstract and technical objectives shows responsiveness; the order does not require it of applicants.

### 2. Articulate Technical Innovation Without Hype

Peer reviewers discount proposals that rely on commercial marketing superlatives. Phrases like "disruptive cognitive reasoning" or "unprecedented autonomous intelligence" signal scientific immaturity and a lack of technical grounding.

To establish technical credibility:
- Define the algorithmic baseline: Identify current state-of-the-art approaches and quantify their failure modes, such as latency bottlenecks, computational overhead, or hallucinations.
- Specify model architectures: Detail model types, parameter scales, neural network layers, loss functions, optimization algorithms, and inference pipelines.
- Describe training and validation datasets: Identify data provenance, annotation procedures, synthetic data generation methods, and data cleaning protocols.
- Quantify compute requirements: Specify training compute budgets, hardware acceleration requirements, memory footprints, and target runtime environments.
- Note the definition where useful: if the topic uses SI, one sentence noting that the work falls within the definition of artificial intelligence in 15 U.S.C. 9401(3), which the order uses for SI, removes any doubt.

### 3. Structure Quantitative Phase I Milestones

Phase I SBIR/STTR awards exist solely to establish the technical feasibility of an innovative concept. Reviewers want to see a clear, chronological work plan divided into discrete tasks with quantitative go/no-go milestones.

A strong Phase I work plan avoids vague goals like "investigate algorithmic approaches" or "explore machine learning models." Instead, each task must yield measurable empirical results:
- Task 1: Dataset acquisition and preprocessing. Milestone: Clean, curated validation dataset containing balanced test partitions with verified ground truth.
- Task 2: Architecture design and baseline model training. Milestone: Model convergence achieving target loss criteria within planned compute allocations.
- Task 3: Performance benchmarking and optimization. Milestone: Documented validation meeting predetermined quantitative feasibility targets, such as bounded inference latency and classification precision exceeding domain benchmark thresholds.
- Task 4: Feasibility reporting and Phase II transition planning. Milestone: Comprehensive Phase I final report establishing feasibility and detailing the Phase II commercial prototype roadmap.

### 4. Build a Credible Commercialization Strategy

Federal agencies fund SBIR/STTR projects not merely to advance basic research, but to stimulate technological innovation that transitions into commercial markets or federal acquisition programs. Even in Phase I, reviewers evaluate commercialization potential.

Founders must articulate:
- Commercial market opportunity: Define the addressable commercial market, customer segments, pricing models, and competitive alternatives.
- Dual-use potential: For defense and energy proposals, explain how the underlying technology addresses commercial industry needs alongside government missions.
- Intellectual property strategy: Detail existing patents, trade secrets, and plans to protect SBIR data rights under federal procurement regulations.
- Phase III transition pathway: Identify prospective follow-on funding sources, including private venture capital, non-SBIR federal procurement contracts, or commercial enterprise licensing.
- Customer discovery and Letters of Support: Include substantive Letters of Support from prospective enterprise customers, prime contractors, or government transition partners validating commercial demand.

The following table contrasts common Phase I narrative pitfalls with winning proposal strategies:

| Proposal Dimension | Conventional Pitfall | Competitive Proposal Strategy |
| :--- | :--- | :--- |
| Vocabulary Alignment | Ignoring agency SI terminology or asserting speculative artificial superintelligence | Mirroring agency SI phrasing in summaries while grounding technical details in 15 U.S.C. 9401(3) |
| Technical Approach | Describing models with vague marketing buzzwords and unproven claims | Detailing mathematical architectures, loss functions, hyperparameters, and compute profiles |
| Work Plan Milestones | Proposing subjective milestones like "study model effectiveness" | Defining quantitative go/no-go thresholds for accuracy, latency, memory, and convergence |
| Commercialization | Treating commercialization as an afterthought or citing generic market research | Presenting dual-use customer discovery, patent strategies, and enterprise Letters of Support |

## Organizing SBIR Proposal Files and Technical Artifacts in Fast.io

Preparing an SBIR or STTR application requires intense collaboration across technical founders, principal investigators, university research administrators, grant consultants, and external advisors. A complete Phase I proposal submission comprises dozens of individual documents: Project Narrative, Budget Justification, Biographical Sketches in SciENcv format, Facilities and Other Resources statements, Data Management and Sharing Plans, Letters of Support, Company Commercialization Reports (CCR), and STTR consortium agreements.

When proposal teams manage these assets across fragmented consumer storage links or email chains, version control errors frequently arise. An external grant writer might revise an outdated technical narrative, or an administrative coordinator might upload an unapproved budget spreadsheet to Grants.gov or DSIP, leading to administrative rejection.

Small business research teams maintain structured, auditable proposal workspaces using dedicated collaboration tools. [Fast.io workspaces](/product/workspaces/) provide centralized cloud environments where distributed proposal teams store, organize, and search grant documentation. Technology startups establish org-owned workspaces for specific solicitations, separating public narrative drafts from proprietary company records.

Key workspace capabilities for SBIR proposal development include:

- Granular access controls: Permissions are set per organization, workspace and share; a share gives an outside party access to only the files or folder placed in it. Founders can grant university subcontractors and grant writers access to technical narrative drafts while restricting confidential company cap tables, founder salaries, and provisional patent filings to authorized executive staff.

- Per-file version history: Every draft uploaded to a workspace maintains a complete, auditable revision history. Proposal managers can track iterative changes across internal review cycles, compare technical edits, and restore earlier text if a collaborator introduces formatting errors.

- Detailed activity log: Fast.io keeps a detailed activity log of file uploads, changes, and permission and membership changes, so founders can see who changed which file and when. This log provides founders with operational visibility over proposal assets and institutional commitment letters.

- Metadata Views: Federal solicitations often span dozens of pages of complex formatting instructions, font requirements, page limits, and mandatory attachments. [Metadata Views](/product/document-data-extraction/) turn dense solicitation documents and compliance files into structured, queryable data. Teams define extraction fields in natural language, and the system extracts application deadlines, page constraints, allowable costs, and required forms into a sortable spreadsheet view.

- Workspace e-signature: STTR intellectual property allocation agreements, university subaward commitment letters, and consultant agreements can be sent, routed, and executed directly within the workspace through [e-signature](/product/e-signature/). The platform supports sequential or parallel signer routing, optional one-time passcodes, and files completed agreements with tamper-evident audit certificates back into the project folder.

Fast.io provides predictable subscription tiers documented on [Fast.io pricing](/pricing/). Proposal teams can evaluate the platform through a 14-day trial on monthly plans, which requires a credit card.

| Plan Tier | Monthly Billing | Workspace Allowance | Included Storage | Monthly AI Work Credits |
| :--- | :--- | :--- | :--- | :--- |
| Starter | $9.99 per month | 5 workspaces | 250 GB | 100,000 credits |
| Business | $49.99 per month | 50 workspaces | 5 TB | 600,000 credits |
| Enterprise | $199.99 per month | 200 workspaces | 25 TB | 3,000,000 credits |

## Founder Pre-Submission Checklist and Post-Award Grant Management

Before submitting an SBIR or STTR application through DSIP, Research.gov, or Grants.gov, founders must execute a disciplined quality control process. Ensuring proposal compliance protects the application from early administrative rejection and positions the business for successful Phase I execution and Phase II transition.

### 1. Audit Solicitation Instructions and Q&A Windows

Examine the governing agency solicitation instructions and topic descriptions. Use the question periods agencies offer, such as the DoD pre-release period and the Topic Q&A in DSIP, to ask topic authors technical questions. If instructions contain ambiguous phrasing regarding Super Intelligence or artificial intelligence terminology, request clarification through official agency channels before the Q&A window closes.

### 2. Verify Vocabulary Alignment Across Proposal Volumes

Review the Project Abstract, Specific Aims, Technical Narrative, and Commercialization Plan to verify consistent terminology. Where the agency solicitation uses Super Intelligence, mirror that phrasing in the executive overview and abstract to confirm responsiveness during initial administrative screening.

In the technical narrative, verify that algorithmic descriptions avoid speculative marketing jargon. Confirm that the technical approach specifies concrete architectures and explicitly links the software pipeline to 15 U.S.C. 9401(3).

### 3. Protect Historical Past Performance and Biosketches

Section 2(b) of Executive Order 14434 states that nothing in that section requires the alteration of previously issued regulations, Presidential actions, contracts, grants, or other historical documents.

When preparing Biographical Sketches, Current and Pending Support disclosures, and past performance summaries, founders must retain original published titles for past papers, conference proceedings, patents, and prior federal contracts. Changing "artificial intelligence" to "Super Intelligence" in a past title means the citation no longer matches the published record.

### 4. Confirm STTR Subcontracting Allocations

For STTR proposals, the small business must partner formally with a research institution: a nonprofit college or university, a nonprofit research institution, or a federally funded research and development center. Founders must verify that budget allocations comply with statutory minimums:
- Small business minimum: the small business must perform at least 40 percent of the work.
- Research institution minimum: the research institution must perform at least 30 percent of the work.
- Intellectual property agreement: the STTR program requires a written agreement allocating intellectual property rights between the two parties; check the solicitation for when it is due.

### 5. Establish Post-Award Reporting and Phase II Roadmaps

When an SBIR award is granted, project management shifts from proposal writing to technical execution. Awardees must submit regular progress updates through the Research Performance Progress Report (RPPR) or agency-specific milestone deliverables.

Describe project accomplishments and technical milestones using the vocabulary of the funded proposal. Program managers evaluate progress against the specific aims outlined in the original award. Maintaining consistent terminology avoids administrative confusion during technical reviews.

The following checklist assists founders and proposal teams during final review:

| Proposal Volume | Primary Review Objective | Founder Action Item | Governing Authority |
| :--- | :--- | :--- | :--- |
| Project Abstract | Programmatic responsiveness | Mirror agency SI terminology if present; spell out term on first mention | EO 14434 Sec. 2(a); Agency Topic |
| Technical Narrative | Algorithmic rigor and feasibility | Detail model architectures, loss functions, compute budgets, and 15 U.S.C. 9401(3) | 15 U.S.C. 9401(3); Peer Review Rubrics |
| Work Plan Milestones | Objective progress verification | Establish quantitative go/no-go thresholds for speed, precision, and efficiency | Phase I Solicitations; Agency Guidelines |
| Biographical Sketches | Investigator credentials and track record | Preserve original, verbatim publication titles and historical award names | EO 14434 Sec. 2(b); SciENcv Standards |
| Commercialization Plan | Market demand and follow-on viability | Include enterprise Letters of Support, dual-use applications, and Phase III paths | SBA Policy Directive; Agency Rubrics |

## Frequently asked questions

### How should startups describe AI innovations in SBIR proposals after EO 14434?

Startups should adopt Super Intelligence terminology in project abstracts, specific aims, and executive summaries when responding to agency solicitations that use that phrasing. In technical narratives, founders should give precise details, including model architectures, training datasets, compute allocations and quantitative benchmarks; that detail, not a statutory citation, is what shows technical rigor.

### Does using 'Super Intelligence' impact SBIR Phase I technical scoring?

Using Super Intelligence confirms administrative responsiveness when an agency solicitation adopts that vocabulary under Section 2(a) of Executive Order 14434. However, technical merit scoring depends entirely on algorithmic feasibility, scientific methodology, quantitative milestones, and commercialization potential. Speculative claims of artificial superintelligence can harm scoring during peer review.

### How do SBIR solicitations from DoD and NSF define Super Intelligence?

Check each solicitation's definitions. Where a solicitation uses SI without defining it, Section 3(a) of Executive Order 14434 defines SI by reference to the definition of artificial intelligence in 15 U.S.C. 9401(3). Under this statutory standard, Super Intelligence encompasses existing functional artificial intelligence systems capable of making predictions, recommendations, or decisions for human-defined objectives.

### Does Executive Order 14434 apply retroactively to active SBIR or STTR awards?

No. Section 2(b) of Executive Order 14434 states that nothing in that section requires the alteration of previously issued regulations, Presidential actions, contracts, grants, or other historical documents. Active Phase I and Phase II awards remain governed by their original terms and scopes of work without requiring contract modifications.

### What is the difference between federal Super Intelligence and research superintelligence in grant applications?

Federal Super Intelligence (two words, capitalized, abbreviated SI) is an administrative label for current operational machine learning systems defined in 15 U.S.C. 9401(3). Research superintelligence (one compound word, abbreviated ASI) refers to a hypothetical future system exceeding human expert capabilities across all fields, which does not exist today. Grant applications should focus on operational systems rather than speculative lab concepts.

### Will the November 28, 2026 proposal immediately change SBIR solicitation rules?

No. Section 3(b) of Executive Order 14434 directs the Assistant to the President for Science and Technology to submit proposed legislative language for a federal definition of Super Intelligence to the President by November 28, 2026. This submission represents an internal executive proposal rather than an enacted statute. Any formal change to statutory definitions requires an Act of Congress.

### How should founders handle historical publication titles in SBIR Biographical Sketches?

Founders must maintain original, verbatim titles for all past publications, patents, conference presentations, and prior grant awards in biographical sketches. The order does not ask anyone to rename past work, and an altered title no longer matches the published record.

### Can SBIR Phase I research funds be used to train frontier foundation models?

SBIR Phase I awards are designed to establish technical feasibility for specific commercial or mission applications, typically with budgets focused on targeted proofs of concept. Training large-scale frontier foundation models exceeds typical Phase I budget constraints. Founders should focus Phase I proposals on fine-tuning, domain adaptation, novel model architectures, or specialized algorithmic pipelines.

## Sources

- [The White House: Executive Order 14434](https://www.whitehouse.gov/presidential-actions/2026/09/inaugurating-the-era-of-super-intelligence/): Under Section 3(a) of Executive Order 14434, federal Super Intelligence encompasses the technologies defined as artificial intelligence in 15 U.S.C. 9401(3).
- [SBIR.gov: About America's Seed Fund](https://www.sbir.gov/about): America's Seed Fund is coordinated by the Small Business Administration across 11 participating federal agencies that fund innovations through the SBIR and STTR programs.

## About Fast.io

Fast.io provides shared workspaces where people and AI agents work on the same files, with built-in semantic search and citation-backed chat over what they hold. Agents reach it through a remote MCP server, a REST API at https://api.fast.io/current/, and a command line client published on npm as @vividengine/fastio-cli. MCP setup is at https://mcp.fast.io/docs: Claude and most MCP clients connect to https://mcp.fast.io/mcp/tools, ChatGPT to https://mcp.fast.io/mcp/operations, and coding agents to https://mcp.fast.io/mcp/code.