🇺🇸 SBIR · STTR · ARPA-H · BARDA · MTEC  —  🇨🇦 SR&ED · IRAP · ISC · MITACS

Capital that costs you no equity.

Stargates helps deeptech and health-AI companies across North America win non-dilutive government funding — then build, protect and commercialize what the proposal promised. We've been the applicant, not just the consultant: our team built Bio Conscious Technologies from a Vancouver research spin-out into a platform integrated with Dexcom, Abbott and Senseonics, billable under Medicare, and funded by government without giving up a share.

0🇺🇸 federal agencies run SBIR/STTR — most founders apply to one
0🇺🇸 SBIR Phase II ceiling incl. modifications (SBA, 2026)
0🇨🇦 new SR&ED limit at 35% refundable — doubled in 2026
0equity surrendered — that's the entire point
Selected work

We've been on your side of the application.

Most funding consultants have never run the company they're writing for. Ours did — through platform integrations that gatekeep an entire category, a reimbursement pathway, and a government funding round that cost no equity.

Below: one engagement that has already run its arc, and one in flight right now at the opposite end of the maturity curve. Deeptech looks very different at each end, and pretending otherwise is how consultants get found out.

Every claim is public and checkable. That's deliberate.

Delivered — five-year arc

Vancouver-based clinical AI platform turning continuous glucose data into forecasting and risk stratification for clinicians. The last five years took it from a working product to an integrated, billable, government-funded platform.

Dexcom API Abbott FreeStyle Libre Senseonics Eversense NVIDIA Inception MTEC member Cerner · eClinicalWorks · OSCAR
2021 — 2023

Built the evidence base that reviewers read

Research presented at successive American Diabetes Association Scientific Sessions. Peer-reviewed evidence is what separates a funded application from a rejected one — and it cannot be manufactured at deadline.

2024

First sensor integration, and an AI layer on top

Abbott FreeStyle Libre compatibility shipped, followed by GPT integration in Endobits V5 — moving the platform from prediction to clinician-facing interpretation.

2025

Landed Dexcom

Strategic API integration with the category's dominant sensor platform. One of the hardest business-development doors in digital health, and the one that makes everything downstream possible.

2026

Completed the sensor triad

Senseonics Eversense integration closed the set — Dexcom, Abbott and Senseonics. Coverage across effectively the whole CGM market rather than a single vendor's install base.

Throughout

Made the software billable, not just impressive

Built against existing Medicare remote patient monitoring and chronic care management CPT codes. Reimbursement is the difference between a pilot that ends and a business that compounds.

May 2026

Raised government money, gave up nothing

$1.5M non-dilutive investment from PacifiCan (Government of Canada) for North American commercialization. Zero equity, zero board seats, zero liquidation preference.

In flight — current engagement

Unmanned aircraft that put out fire with low-frequency sound instead of water or retardant. Pressure waves near 17 Hz thin the boundary layer at the base of a flame and separate it from the fuel surface — no tank, no chemical, no reload, and nothing to refill between sorties. The environmental case is as strong as the operational one: conventional suppressants released an estimated 380,000 kg of heavy metals across the western US between 2009 and 2021.

Low-frequency acoustics · 14–21 Hz Unmanned aircraft Wildfire response Pre-revenue deeptech R&D

Why it's a federal candidate: wildfire is one of the few missions funded across USDA, DoD, DHS, NASA and NSF simultaneously — and on the Canadian side through NRC and the regional agencies. A platform that removes chemical retardant from the equation speaks to every one of those mandates at once. The work is agency targeting, readiness evidence and an IP position that holds while the category gets crowded.

SubsystemPublished readiness
Airframe & flight operationsFlight-tested
Acoustic emitterBench prototype
Airborne power & scalingActive R&D — critical path
Beam-forming & standoffActive R&D
Reignition controlActive R&D

This is what an honest readiness assessment looks like, and it is precisely why the programme is fundable. Hellifighter states plainly that nobody has yet extinguished a wildfire with sound, that acoustic suppression interrupts combustion without removing heat, and that sound pressure attenuates outdoors. Reviewers reward that candour — it is the applicants who quietly round their readiness level up who lose the cycle, and sometimes the award.

The real failure modes

Strong science loses to weak paperwork constantly.

Federal reviewers do not just score whether the technology works. They score whether you can build it, whether anyone will buy it, and whether the intellectual property is clean. Most rejections happen in the second and third.

01

The commercialization plan is an afterthought

Technical founders spend 90% of the effort on the research narrative and write the commercialization section the night before. It's a scored section, and reviewers who fund for a living can tell within a paragraph that nobody has spoken to a customer.

We fix it by having sold the product before writing about selling it.

02

The readiness level is overstated

Claiming a higher Technology Readiness Level than you can demonstrate is the fastest route to a rejection — or worse, to an award you can't deliver against, which ends your Phase II before it starts.

We fix it by engineering the evidence, not just describing it.

03

The IP position is undefined

Who owns what, which background IP you're bringing, what the government gets a licence to, and whether a university holds rights through Bayh-Dole. Ambiguity here is a red flag to reviewers and a landmine at acquisition.

We fix it by settling the IP position before submission.

How the money works

Two countries. Two systems. One playbook.

The US pays you to prove something works. Canada pays you back for having tried. They reward different behaviour, run on different clocks, and — structured properly — they stack.

Competitive, mission-driven, and slow. Eleven federal agencies run SBIR/STTR, each with its own priorities and deadlines. The cheques are the largest available anywhere in North America, and you win them by convincing a reviewer the technology serves that agency's mission — not by being clever. Ceilings below are SBA maximums including modifications; agency norms sit lower.

Phase I — Feasibility
$323KSBA ceiling incl. modifications, April 2026

Prove it can work

The cheque isn't the prize. Phase I buys you eligibility for Phase II and a programme officer who knows your name — both worth more than the money.

~6–12 MONTHS
Phase II — Development
$2.15MSBA ceiling incl. modifications, April 2026

Build the thing

Where the real money is, and where companies discover their Phase I promises were easier to write than to engineer. This is the phase we exist for.

~2–3 YEARS
Phase III — Commercialization
UncappedNo SBIR funds; sole-source authority applies

Sell it back to government

Phase III work derived from your SBIR can be awarded sole-source — agencies can buy from you without recompeting. Most founders never learn this exists.

NO PROGRAMME DOLLAR LIMIT

The cross-border move — and the three ways it goes wrong.

A North American structure can claim SR&ED on Canadian R&D while a US entity pursues SBIR — but only if it was designed that way before the first claim. These are the failure points we check on the first call, because all three are cheap to fix in advance and expensive to fix afterwards.

Where the work is performedSR&ED requires R&D carried out in Canada. SBIR requires US place of performance. The same engineer-hours cannot be counted on both sides, and sloppy allocation invites a review from whichever authority looks first.
Who owns the entitySBIR carries ownership and control requirements. The enhanced 35% refundable SR&ED rate turns on CCPC status. A US parent inserted for one programme can quietly disqualify you from the other.
When you set it upBoth systems judge structure as it existed during the work, not as it exists when you file. Restructuring after the fact is expensive at best and retroactively disqualifying at worst.
20
Year SBIR data-rights protection
The part nobody explains

Taking federal money doesn't mean losing your IP.

Under SBIR data rights, the government receives a licence to use what you develop — but for a protection period of 20 years from each award, it cannot disclose your proprietary data to third parties. That protection is one of the strongest features of the programme, and it is routinely misunderstood by founders and their investors alike.

What you must get right is the boundary: which background IP you bring in, what is developed under the award, and what a university partner may hold through Bayh-Dole. Get it wrong and you find out during diligence, when the leverage is entirely on the other side of the table.

We do this work as technical IP strategy in coordination with your patent counsel and contracts attorney. Stargates is not a law firm.

What we do

Win it. Build it. Protect it. Sell it.

Grant writers write and disappear. Dev shops can't read a solicitation. Patent firms arrive too late to change the architecture. We do the four together because federal reviewers score them together — and because Phase II is won by whoever actually delivered Phase I.

01 / CAPITAL

Non-dilutive funding strategy

Which country, which programme, which phase — and whether you're eligible at all. We map your technology to the programmes whose mission it actually serves on both sides of the border, then build the submission around the sections that get scored rather than the ones that are fun to write.

  • 🇺🇸 SBIR/STTR, ARPA-H, BARDA, MTEC, NSF and DoD targeting
  • 🇨🇦 SR&ED claim strategy, IRAP, Innovative Solutions Canada, Mitacs and regional agencies
  • 🇨🇦 AI-specific: AI Compute Access Fund, Scale AI, NGen, DIGITAL and institute affiliation
  • Eligibility, registration and entity structure (SAM.gov, CCPC status, cross-border design)
  • Technical volume, commercialization plan and budget justification
  • Phase sequencing, resubmission planning and sole-source positioning
02 / ENGINEERING

Technology development

The part that decides whether the award becomes a company or a write-off. Architecture, AI and data systems, device and sensor integrations, and the clinical or field evidence that moves you up the readiness ladder — built to be handed over, not to keep you dependent.

  • AI/ML systems, predictive models and data infrastructure
  • Device, sensor and platform API integrations
  • Health IT interoperability (EHR, HL7/FHIR, remote monitoring)
  • Technology Readiness Level advancement with the evidence to prove it
03 / DEFENSIBILITY

IP & data-rights strategy

Run in parallel with the build, not after it. Landscape the field before you commit to an architecture, capture inventions before disclosure makes them unpatentable, and define the background/foreground IP boundary before a federal award makes it someone else's question.

  • Prior-art and patent landscape analysis
  • Freedom-to-operate assessment
  • Background vs. foreground IP definition for federal awards
  • Invention capture and portfolio prioritisation, with your counsel

Technical IP strategy delivered alongside your registered patent attorney or agent. Stargates is not a law firm and does not provide legal advice or file on your behalf.

04 / MARKET

Commercialization & deployment

Reviewers fund companies, not experiments. We build the reimbursement or procurement pathway, the positioning that survives a technical buyer, and the deployment infrastructure that turns a funded prototype into something a customer can actually run.

  • Reimbursement and procurement pathway design
  • Positioning, messaging and technical marketing assets
  • Pilot and design-partner programme design
  • Production deployment, onboarding and team enablement
The honest comparison

Four ways to do this. Including three that aren't us.

We've marked where someone else is the better call, because you'll find out anyway and we'd rather you heard it here.

StargatesGrant writerDev shopStrategy consultancy
Writes the proposalYesYesNoSometimes
Builds what it promisedYesNoYesNo
Handles IP & data rightsYes — before submissionNoNoSeparate practice, separate fee
Has run the applicant companyYesRarelyNoNo
Fee structureFlat, quoted up frontOften contingency — see belowHourly / time & materialsRate card + change orders
Survives to Phase IISame team, same accountabilityGone after submissionRebuilds context each timeNew engagement
Better than us when…You need the award and the thing it fundsYou have engineering and only need the writingThe spec is locked and boringYou need brand cover for a board decision
Fee structure

Flat fees aren't a preference. They're the compliant answer.

Plenty of grant consultants offer to work for a percentage of the award. It sounds founder-friendly until an audit asks how the fee was paid.

No contingency fees

Contingent and success fees paid out of SBIR/STTR award funds are widely treated as unallowable costs. We charge a flat project fee so there is nothing awkward to explain later. Confirm the treatment with your accountant — but ask any contingency-based consultant the same question first.

The price does not move

The number in the proposal is the number on the invoice. If we underestimated, that's ours to absorb. Additions are re-quoted as their own fixed item and approved by you before any work happens.

Milestone-gated payment

You pay against accepted milestones, not elapsed calendar time. If a milestone isn't accepted it isn't billed, and you can end the engagement between milestones.

NDA before the first real call

Send yours or we'll send ours. Your name never appears in our marketing without written permission — during the engagement or after it.

Everything transfers to you

Code, repositories, credentials, documentation and generated IP are assigned to you in the contract from the start — not negotiated at the end when you have the least leverage.

We'll tell you not to apply

If you're ineligible, mis-targeted, or six months too early, that's what the first call concludes. A rejected application costs you a cycle you may not get back — we'd rather lose the project.

Are you actually fundable?

Thirty minutes. We'll tell you which programmes fit on each side of the border, whether your readiness level supports an application this cycle, and what would have to be true to win. No charge, no obligation.

Book the call
Questions

What you were going to ask anyway.

Do I have to be a US company?

For SBIR and STTR, yes — those programmes are for US small businesses, with ownership, control and place-of-performance requirements that catch a lot of foreign-founded companies off guard. But that's only half the map: if you're Canadian, SR&ED, IRAP and Innovative Solutions Canada are open to you today, and SR&ED in particular is an entitlement rather than a contest.

Structuring a compliant US entity alongside a Canadian one is often the first piece of work. It's far better done deliberately than discovered mid-application — we've been through the cross-border version of this ourselves, which is why it's the first thing we check on a call.

We're Canadian. Is SBIR even worth chasing?

Sometimes, and sometimes not — it depends on how much US market you actually need. A Canadian company can build a very respectable non-dilutive stack without touching SBIR: SR&ED as the base layer, IRAP for project costs, Innovative Solutions Canada for a federal customer, Mitacs for research talent, and a regional agency for scale-up.

Chasing SBIR makes sense when a US federal customer or US market entry is genuinely strategic. It costs a US entity, US place of performance, and a structure that doesn't accidentally cost you CCPC status and the enhanced 35% SR&ED rate. We'll model both paths on the first call rather than assuming the American one is better.

What actually changed with SR&ED in 2026?

Bill C-15 received royal assent in March 2026 and it's the largest expansion of the programme in a decade. The annual expenditure limit for the enhanced 35% refundable credit doubled from $3M to $6M; the taxable-capital phase-out range moved from $10M–$50M up to $15M–$75M; capital expenditures became eligible again for the first time since 2014; and the enhanced refundable credit was extended to certain Canadian public corporations.

The practical consequence is that companies previously phased out may now qualify, and R&D budgets built on the old $3M ceiling are leaving money unclaimed. The constraint is almost never eligibility — it's contemporaneous technical documentation, which is the thing nobody keeps and the thing a reviewer asks for first. Confirm your specific position with your accountant; we work on the technical narrative and the documentation discipline that supports it.

Will you work on contingency — a percentage of the award?

No. Contingent and success fees paid from SBIR/STTR award funds are widely treated as unallowable, and a fee arrangement that creates an audit problem is not a saving. We quote a flat fee up front. If a competing consultant offers you a percentage, ask them in writing how they expect it to be paid and from which funds.

If the government funds it, does the government own it?

No. Under SBIR data rights the government receives a licence to use the data developed under the award, but for a 20-year protection period from each award it cannot release your proprietary data to third parties. What matters is drawing the line clearly between background IP you bring in and foreground IP developed under the award — and checking whether any university partner holds rights through Bayh-Dole.

We handle this as technical strategy alongside your patent and contracts counsel; we are not a law firm.

You're a new firm. Why take the risk?

Because the track record is older than the firm and every item on it is publicly verifiable — the sensor integrations, the conference presentations, the reimbursement pathway, the government funding round. Check them before you call us.

And because the engagement is structured so you carry as little risk as possible: the first call and the written proposal are free, the price can't move against you, payment is gated on accepted milestones, and you can stop between them.

How long does this take, realistically?

Solicitation deadlines drive everything, and they are unforgiving. A serious Phase I proposal wants six to eight weeks of runway; registrations alone (SAM.gov and the associated identifiers) can take weeks and are a common reason companies miss a cycle entirely. If you're three weeks from a deadline with no registrations, we'll usually tell you to target the next cycle and use the time to be genuinely ready.

Do you only work in health and life sciences?

No. Our deepest domain scar tissue is in health AI, medical devices and regulated data — that's where we built and shipped. But the machinery is the same across deeptech: agency targeting, readiness evidence, IP position, commercialization. We work across energy, defence, advanced manufacturing and AI infrastructure, and we'll say plainly on the first call where our domain knowledge is thinner.

Can you help after the award, not just before it?

That's the point of us. Winning Phase I is the easy half; delivering it is what makes Phase II possible, and Phase II is where the real money and the sole-source Phase III authority live. We're built to do the engineering, not to hand you a PDF and wish you luck.

What if we don't win?

Most applicants don't, on the first attempt — this is a competitive programme and any consultant implying otherwise is selling something. What a serious application buys you even when it loses is reviewer feedback, a programme officer relationship, and a technical volume that is 80% reusable next cycle. We plan for the resubmission from the start rather than pretending it won't be needed.

Start here

Tell us what you're building.

A few sentences is plenty. We reply within one business day with a call invite or an honest "not yet, and here's why."

Free fundability reviewAgency fit, readiness level, realistic timeline.
Written fixed proposalScope, milestones, timeline and one price.
NDA before we talk detailSend yours, or we'll send ours.
A straight no, if it's a noIncluding "wait for the next cycle."

Prefer email? hello@stargates.ai

We reply within one business day. Your details are never shared or sold.

Got it — thank you.

Your message is in. We'll come back to you within one business day with next steps.