Tests neural-device chips against programmable electrode loads before a board spin commits the team to the wrong architecture.

1. MintNeuro reports delivered working silicon and integration milestones with two neural-device developers, showing that partner-level silicon integration is already underway.
2. Imperial describes a consequential design choice between generic components not designed for the body and years of custom chip development.
3. The announced funding is reported to support product launches, partner silicon delivery and engineering expansion, creating a near-term scale-up context for independent test services.

Catalyst. MintNeuro's reported working silicon, partner integration milestones and new funding for product launches and partner delivery put neural-interface silicon into a scale-up phase; Imperial also describes the enduring generic-component-versus-custom-chip tradeoff.

The company builds a calibrated electrical bench that emulates customer-defined electrode impedance and stimulation duty-cycle cases for candidate neural front ends. Engineers submit a chip or evaluation board and their operating envelope; the lab replays the same cases across candidates and records measures such as input noise, channel behavior, pulse fidelity and power under those test conditions. Each paid campaign returns a reproducible technical report for architecture and board-design decisions, with explicit limits on what the bench does not prove about tissue response or clinical performance. Recurring programs can move onto licensed bench hardware and a versioned test protocol.

What's different. Neural chip suppliers build and sell silicon; this venture sells an independent, repeatable way for device teams to compare candidate electronics under their own declared electrical loads. Its defensible asset is a calibrated load-emulation method and growing set of test recipes, not ownership of a chip design or a claim of clinical readiness. The service begins with hands-on testing because the market and degree of load standardization are not established.

Jobs to be done

flowchart LR
  Team["Neural-device engineering team"] --> Recipe["Electrode-load and duty-cycle recipe"]
  Recipe --> Bench["Calibrated load-replay bench"]
  Candidates["Candidate front ends or ASICs"] --> Bench
  Bench --> Measures["Comparable noise, pulse-fidelity, and power measurements"]
  Measures --> Decision["Choose the next board architecture"]

Executive takeaways

• The engineering problem is credible: implantable neural devices face tight noise, power, charge-balance, form-factor, and electrode-interface requirements, and partners are integrating neural silicon now [5][6][17][18].
• Evidence supports established neural chip testing, not a new test principle. RC load models, electrode-impedance measurement, chip characterization, and medical-device test services already exist [1][2][3][7][9][12][17][18][22].
• The white space is narrower and conditional: no dedicated neutral cross-vendor neural-chip replay service was identified in the reviewed public pages, but internal labs and existing ASIC or test providers could supply one.
• The service-only beachhead is small in an explicit bottom-up scenario: about $2.5M global annual spend at 100 assumed buyer programs, one campaign per year, and an unverified $25K fee. Broad BCI market estimates are not substitutes for this buyer count [23][24][26].
• Budget for neural R&D is visible through venture-backed development and public grants, but paid demand for independent campaigns is not. Validate it with paid pilots before buying specialized equipment or licensing a protocol [6][20].

Market definition

The relevant market is third-party engineering campaigns that replay customer-declared electrode impedances and stimulation duty cycles against multiple neural sensing or stimulation front ends before a board architecture or custom-silicon decision. It excludes therapeutic-device revenue, clinical trial services, formal certification, chip design revenue, and broad BCI hardware/software sales. TAM, SAM, and SOM below are scenario estimates, not a reported market census; the $703.7M U.S. BCI estimate is a broad category that includes segments such as non-invasive systems and software and is not used as the service TAM [26].

Customer and buyer

The likely user is an ASIC or systems engineer at a preclinical neural-device developer; the CTO or VP Engineering is the likely budget owner, with procurement, quality, and regulatory staff reviewing sample handling and use of results. The strongest trigger is a comparison of candidate front ends before a board freeze, especially when the team is integrating a chip supplier's device-specific platform. MintNeuro's partner work illustrates that integration timing, but no public source quantifies the incidence or cost of late electrical mismatch [5][6][7].

Buying triggers

• A product-architecture or board revision decision where the team must choose between an off-the-shelf front end, a neural-specific chip, and a custom ASIC; TTP's case study describes the need to define architecture and specifications before committing to development. [7][8][9]
• A new device–chip integration programme with customer-specific stimulation and sensing requirements. MintNeuro and Motif describe joint architecture and hardware validation, with a multi-year chip supply agreement. [5][6]
• A grant-backed preclinical programme that needs independent engineering evidence before the next prototype, though grant funding does not show that the recipient has budgeted for an external bench. [4][20]
• A need to document design verification inputs and test results for a medical-device programme; the bench can support engineering decisions but cannot replace sponsor-led system verification, clinical evidence, or regulatory review. [14][15][21]

Willingness to pay

Not validated. Intan lists RHD recording controllers at $9,950 and $14,250, providing an equipment-capex reference, not a price for independent testing; its full system requires additional components. No public campaign rate, purchase order, or tender for cross-vendor neural-chip load replay was found. Treat $15K–$30K per campaign, with $25K used in the sizing scenario, as an operator hypothesis only. A buyer must believe the fee is below the value of engineering time saved or a wrong architecture avoided; grant and venture funding are not evidence of spend on this service. [1][2][6][20]

Category dynamics

Growth signal 17.34% forecast CAGR for the broad U.S. BCI market, 2026–2035; not a forecast for independent test services

Validation signals

• MintNeuro reports that its sensing chip has been integrated into Motif devices used in preclinical work and describes a multi-year commercial supply agreement, demonstrating an active integration workflow but not independent test demand.
• Public neural-interface funding programmes support preclinical R&D; they may help source pilots, while award amounts do not show spend on the proposed test service.
• Published neural-chip studies use RC models and saline testing, and vendor instruments already measure electrode impedance, validating feasibility while weakening the novelty of the base method.
• Independent reporting describes several active BCI architectures and notes different trade-offs in invasiveness and bandwidth; the relevant market is technologically heterogeneous rather than a uniform load class.

Regulatory & technical constraints

• Neural-device classification and premarket pathway depend on intended use and risk; the bench must not imply that its engineering report constitutes FDA clearance, an IDE, or clinical evidence.
• FDA QMSR effective February 2, 2026 incorporates ISO 13485:2016; device customers may flow supplier, record, and quality expectations into an external testing relationship.
• EU MDR places risk management, clinical evaluation, and technical-documentation obligations on the manufacturer; a campaign protocol must be traceable to customer requirements if used in design records.
• RC electrode models enable repeatable electrical tests but do not reproduce tissue electrochemistry, electrode geometry, chronic changes, or in-vivo safety.
• Electrical measurement ranges, fixture parasitics, channel-to-channel error, calibration, and sample-to-sample variation must be characterized before results can be compared across chips.

Competition is mainly substitution rather than a named dedicated replay-bench vendor. Intan sells neural recording and stimulation systems with impedance measurement; Ripple Neuro and Blackrock sell neural recording/stimulation platforms; MintNeuro supplies purpose-built chips and partner integration. imec IC-Link offers chip test and qualification, TTP and CSEM offer custom ASIC or product-development expertise, Cirtec supports neuromodulation development, and Element offers medical-device testing from prototype stages. Microprobes provides a low-complexity impedance tester. The proposed service would need to win on impartial cross-candidate protocol consistency and reusable customer recipes, not on generic ability to connect an RC load [1][2][3][5][7][8][9][10][11][12][19][22].

Why incumbents do not win by default

• Internal engineering lab and customer-built fixture. The incumbent can use the customer's own instruments, evaluation boards, resistor-capacitor loads, and scripts at marginal cash cost, with immediate access to confidential design details. It loses only if independent calibration, comparable data, or setup speed is important enough to justify an external fee.
• Neural chip and instrumentation vendors. Vendors can supply evaluation boards, impedance measurement, application support, or integration help with the device they sell. They do not automatically give a neutral comparison across competing suppliers, although customers may prefer their bundled support over a separate lab.
• ASIC design and test providers. imec, TTP, and CSEM offer chip testing, qualification, architecture, or custom medical ASIC expertise within broader development engagements. These firms can win when the customer wants design responsibility and test in one contract; the opportunity for a neutral lab exists only if customers perceive a conflict or capability gap.
• Medical-device contract test laboratories. Element provides protocol development, prototype/feasibility testing, and safety services that can absorb verification budgets. Public service pages reviewed do not specify neural front-end comparison under customer-specific electrode loads, but the laboratory has established client access and quality systems.
• Electrode impedance meters and general lab instruments. Microprobes and standard impedance or acquisition instruments make basic checks inexpensive and available in-house. They measure or stimulate an electrode, but do not by themselves create a validated multi-chip replay protocol and an architecture decision report.

Neural-device teams face a documented choice between generic electronics and lengthy custom-chip development, while partner-level neural-silicon integration is underway. The proposed company sells a neutral, fixed-scope bench campaign comparing candidate front ends under one customer's declared electrode-load and stimulation conditions before a board decision. The likely first buyer is a UK or European preclinical neural-device CTO or VP of Engineering with an active prototype and a near-term architecture review. Research supports the engineering context and the feasibility of electrical load testing, but it does not establish late-mismatch frequency, paid demand, or willingness to pay for an independent campaign. The $2.5M global TAM, $0.5M UK/EU SAM, and $0.125M SOM are bottom-up scenarios using unverified buyer counts, one campaign per buyer per year, and a $25K assumed fee; they are not market measurements. Start as a manual service, test a $15K–$30K campaign hypothesis, and buy specialized equipment only against paid pilot demand. Vendor platforms, in-house labs, ASIC providers, and medical-device test labs are credible substitutes, so neutrality and repeatability must change a real architecture decision to matter. Reports must remain bounded bench evidence, not clinical performance, tissue-response, safety, or certification claims. This is a watch-stage opportunity until paid pilots, repeatable results, and repeat purchasing establish a viable service.

Problem

• Neural-device teams must choose or integrate sensing and stimulation electronics despite a reported tradeoff between generic components and lengthy custom-chip development.
• In-house or vendor-specific measurements may not give a comparable view of competing front ends under the same customer-defined electrical conditions before a board revision.
• The available research does not quantify mismatch-driven rework, external test budgets, or how often this decision recurs.

Solution

• Run a calibrated, fixed-scope campaign that replays customer-approved electrode-load and stimulation-duty-cycle cases across candidate front ends.
• Report agreed electrical measures such as input noise, channel behavior, pulse fidelity, and power, with calibration and repeatability limits stated.
• Keep customer recipes and raw results private; label reports as engineering measurements that do not establish tissue response, clinical performance, safety, or regulatory conformity.

Why we win

• A neutral same-protocol comparison could be more useful at a multi-vendor architecture decision than separate vendor demonstrations, if buyers value independence enough to pay.
• The service starts with a specific board-selection workflow rather than competing with neural-chip suppliers on silicon, or with broad laboratories on formal device testing.
• A versioned protocol, calibrated fixtures, and permissioned reusable recipes could improve repeatability; no proprietary dataset or defensible moat is established today.

Milestones

flowchart LR
  Decision["Pre-board architecture decision"] --> Campaign["Paid neutral load-replay campaign"]
  Campaign --> Proof["Customer-agreed repeatable comparison"]
  Proof --> Modules["Reusable calibrated fixture modules"]
  Modules --> License["Conditional bench license and calibration"]

Founding team

Experiment roadmap

Risk assessment

What must be true

• At least three of the first 15 qualified buyer interviews document a recent or imminent front-end decision where comparable load-specific evidence is missing and an economic buyer owns a budget.
• At least two independent buyers sign paid pilot orders at $15K or more by month 6 rather than offering only informal interest or grant support.
• A blinded comparison across three front ends and two customer-approved load profiles meets acceptance criteria agreed before testing and informs at least one real component decision.
• At least two of the first three paid pilot customers place or formally budget a second campaign within 12 months.
• A verified census finds at least 20 active UK/EU programs matching the beachhead, and repeat-campaign delivery can reach direct costs at or below 30% of realized fees.

Open diligence questions

• What recent board or ASIC decision was delayed or changed by a mismatch under the intended electrode loads, and what cost or engineering time did it create?
• Who controls budget for independent front-end testing, and what did that buyer last pay for internal, vendor, or contract-lab work?
• How many UK/EU programs have active neural hardware, a near-term architecture decision, and no adequate internal comparison setup?
• Can electrical load replay change the decision that an in-house, in-vitro, or in-vivo test would otherwise determine?
• What setup hours, measurement uncertainty, and repeatability do customer-approved pilots require across different chip families?
• What sample-security, supplier-quality, calibration, and report limitations must be in place before device teams share unreleased hardware?

Model sanity

• Revenue engine. Base revenue is 3, 4, and 5 accepted $25K campaigns in Y1-Y3, with no license revenue.
• Must go right. At least two pilots must close by M6 and repeat delivery must stay within the 30% direct-cost ceiling to support five paid accounts by M24.
• Model breaks if. If the downside ramp of 1, 2, and 2 campaigns at $20K prevents the conditional $1M seed, modeled Y3 cash is negative $136K.
• Next-round proof. The M25 seed should proceed only after five paying accounts, two pilot repeats, blinded repeatability evidence, and direct delivery cost at or below 30% are documented.

Scenarios

Sensitivity

flowchart LR
  Accounts["Qualified UK/EU device accounts"] --> Campaigns["Paid load-replay campaigns"]
  Campaigns --> Reports["Accepted engineering reports"]
  Reports --> Revenue["Campaign count x $25K fee"]
  Revenue --> GrossProfit["70% target gross margin"]
  GrossProfit --> Cash["Cash after payroll, opex, capex, and funding"]

Flags: No paid demand, campaign quote, or willingness-to-pay is established; all modeled orders and the $25K fee are hypotheses. · The $0.5M UK/EU SAM scenario, eligible account count, one-campaign-per-year cadence, and repeat rate are unverified. · 70% gross margin requires direct delivery cost to fall to $7.5K per campaign; engineer hours and fixture reuse have not been demonstrated. · Indicative LTV/CAC is only 1.3x and CAC payback is 23.2 months; the lumpy project model may make both lifetime and payback overstated. · EBITDA remains negative in all three years, and Y3 revenue growth plus margin produces a negative rule-of-40 direction. · Base cash reaches a $164.3K trough before an uncommitted $1.0M M25 seed; without that financing, base Y3 ending cash is negative $278.2K. · No instrument-license or calibration revenue is included; the business remains a narrow, labor-intensive service unless repeat demand is proven.

• Narrow early market. The sources establish a specific chip-development bottleneck but do not establish the number of neural-device buyers or their test budgets. Mitigation: Start with paid design-partner campaigns and require repeat purchase or a bench-license commitment before expanding the instrument range.
• Test-to-device mismatch. An electrical emulator cannot fully represent tissue response, implanted geometry or clinical performance. Mitigation: Label outputs as bench measurements only, scope recipes with device engineers and compare only against customer-approved operating conditions.
• Bespoke test burden. Device-specific electrode loads and signal modes may make each engagement too custom to scale economically. Mitigation: Track setup hours per campaign, standardize recurring fixture modules and decline unsupported test profiles until they fit the protocol.