Safe Robots
AI-enabled safety and quality engineering for robotics teams deploying fast.
Website: https://www.saferobots.co/
Cover Block
Public sources
| Field | Value |
|---|---|
| Name | Safe Robots |
| Tagline | AI-enabled safety and quality engineering for robotics teams deploying fast. [Safe Robots, retrieved 2026] |
| Headquarters | San Francisco, US [F6S, September 2025] |
| Founded | 2025 [F6S, September 2025] |
| Stage | Pre-Seed [F6S, September 2025] |
| Business Model | SaaS [F6S, September 2025] |
| Industry | Deeptech [F6S, September 2025] |
| Technology | AI / Machine Learning [Safe Robots, retrieved 2026] |
| Geography | North America [F6S, September 2025] |
| Growth Profile | Venture Scale [F6S, September 2025] |
| Founding Team | Co-Founders (2), Nestor Grace and Thejas Kesari [F6S, September 2025] |
| Funding Label | Undisclosed [F6S, September 2025] |
Links
Public sources
Executive Summary
PUBLIC Safe Robots is building software to help robotics teams structure safety and quality engineering work earlier in the development cycle, and it merits attention now because the company is trying to turn a largely document-heavy, expert-driven process into a continuously updated system of record for requirements, hazards, evidence, and compliance [Safe Robots, retrieved 2026] [F6S, September 2025]. Public evidence suggests the company took shape in 2025, joined Berkeley SkyDeck in May 2025, and was pitched at SkyDeck Demo Day on September 30, 2025, with Nestor Grace later listing himself publicly as Co-Founder and CEO beginning in January 2026 [LinkedIn, October 2025] [Berkeley SkyDeck, retrieved 2026] [LinkedIn, October 2026].
The core product, Clear, is described as a "living model" that connects engineering systems such as GitHub, Jira, Drive, and Jama, then layers system modeling, regulatory analysis, scenario mapping, hazard analysis, risk modeling, standards compliance, and traceability on top of that shared model [F6S, September 2025] [Safe Robots, retrieved 2026] [GitHub, retrieved 2026] [OpsHub, retrieved 2026]. The interesting part of the pitch is not the connector layer by itself, but the claim that one model can show how a design change affects requirements, code, tests, and safety evidence, which, if it works in practice, addresses a real pain point for robotics programs operating under customer, insurer, or regulatory scrutiny [Safe Robots, retrieved 2026].
The founding team reads as technically relevant for the problem. Grace previously worked as a Senior System Engineer at Cruise on requirements, safety analysis, testing, verification, validation, and failure-mode monitoring for autonomous robotaxis, while public profiles identify Thejas Kesari as co-founder and describe prior work in safety-critical software and technical product management [LinkedIn, October 2026] [ContactOut, retrieved 2026] [LinkedIn, October 2025].
On capitalization, the public record is thin: Berkeley SkyDeck appears both as accelerator and investor, but no round size, financing date, or lead terms are publicly confirmed, so the company is best treated as pre-seed and lightly financed until management provides primary documents [Berkeley SkyDeck, retrieved 2026] [F6S, September 2025]. The business model is presented as SaaS, and the near-term question for investors is whether Safe Robots can convert a compelling systems-engineering narrative into repeatable adoption inside robotics teams, then expand that motion into adjacent safety-critical markets such as medical and automotive without losing focus [F6S, September 2025]. Over the next 12 to 18 months, the signals that matter are straightforward: evidence of named deployments, proof that Clear becomes embedded in engineering workflows rather than used as a point compliance tool, and clearer separation in the market from similarly named companies such as SafeWorld, where public name collision risk is already visible [The AI Insider, October 2026] [FinSMEs, October 2026] [Companies House, September 2025].
Lightly corroborated -- This section relies on a mix of company materials, F6S, Berkeley SkyDeck, and founder-profile sources, with limited independent reporting on financing or commercial traction.
Taxonomy Snapshot
| Axis | Value |
|---|---|
| Stage | Pre-Seed |
| Business Model | SaaS |
| Industry / Vertical | Deeptech |
| Technology Type | AI / Machine Learning |
| Geography | North America |
| Growth Profile | Venture Scale |
| Founding Team | Co-Founders (2) |
| Funding | Undisclosed |
How the Company Got Here
PUBLIC
Safe Robots appears to be a very recent San Francisco startup focused on safety and quality engineering for robotics teams, with the company describing its offering in plain terms as "AI-enabled safety and quality engineering for robotics teams deploying fast" [Safe Robots]. The public record here is still thin, but the company website presents a consistent product posture around requirements, hazards, evidence, and compliance workflows for robotics deployments [Safe Robots].
Chronology matters because the company is early enough that even its formation date is still best read as directional rather than fully documented. An F6S profile lists Safe Robots as founded in 2025 and based in San Francisco, and names Berkeley SkyDeck as an investor, though it does not disclose a financing amount or round structure [F6S, September 2025]. Berkeley SkyDeck also lists Safe Robots in Demo Day Batch 22 and describes Clear as a product for safety and quality engineering for robotics teams deploying fast [Berkeley SkyDeck, retrieved 2026].
There is one adjacent filing worth separating from the core company record. Companies House shows a UK entity called SAFE ROBOTICS UK LIMITED incorporated on 18 September 2025, but the available public evidence does not connect that entity to the San Francisco startup, its product, or its founders [Companies House, September 2025]. For now, the cleanest read is that Safe Robots is an early-stage US company with a public product website and accelerator affiliation, while legal-entity detail beyond that remains limited in the sources cited here [Safe Robots] [Berkeley SkyDeck, retrieved 2026].
Lightly corroborated -- Relies on company website, F6S profile, Berkeley SkyDeck listing, and an unlinked UK filing that appears separate.
Product and Technology
MIXED
The product story is narrower than the category language suggests, which is useful at this stage. Safe Robots describes Clear as software for teams building safety critical robotics systems, with adjacent applicability to medical and automotive engineering, and frames the product around requirements, hazards, and evidence rather than around autonomous control itself [F6S, September 2025] [Safe Robots]. Berkeley SkyDeck uses similar language, describing Clear as safety and quality engineering software for robotics teams deploying quickly, which provides modest third-party corroboration for the product positioning even if it does not verify technical performance [Berkeley SkyDeck, retrieved 2026].
On the website, the workflow centers on a single "living model" that ties together system modeling, regulatory analysis, scenario mapping, hazard analysis, risk modeling, standards compliance, and traceability [Safe Robots]. The company says this model can show how a change in one component affects related requirements, code, tests, and evidence, and the F6S profile adds that Clear connects with tools including GitHub, Jira, Drive, and Jama [F6S, September 2025]. Public evidence supports the presence of those systems in the intended workflow, but not the depth of each integration or whether the product is already deployed in production environments; the GitHub, Jira, and Jama references establish ecosystem fit more clearly than they establish native product capability [GitHub] [Hacker News, May 2020] [OpsHub].
What stands out is the company's attempt to compress a documentation-heavy safety process into software that updates continuously as engineering work changes. If that claim holds in practice, the differentiator is less the AI layer by itself and more the combination of traceability, scenario generation, and compliance mapping in one operating surface for robotics teams [Safe Robots] [F6S, September 2025]. Public materials do not yet provide a verified demo, benchmark, customer implementation, or technical architecture detail, so the present read has to remain at the level of product intent and workflow design rather than proved system performance [Safe Robots] [F6S, September 2025].
Company-stated, unverified -- This section relies primarily on company-controlled materials from the Safe Robots website and F6S profile, with partial positioning corroboration from Berkeley SkyDeck and ecosystem references that do not independently verify product performance.
Where the Demand Sits
Public sources This market matters now because robotics teams are moving from prototype to field deployment faster than the supporting safety, requirements, and compliance tooling appears to be maturing, at least from the product framing Safe Robots itself has chosen to emphasize [Safe Robots, retrieved 2026].
The public record does not support a direct TAM, SAM, or SOM for Safe Robots specifically, and caution is warranted because the company sits across several adjacent categories at once: robotics software, safety engineering, requirements management, and compliance tooling [F6S, September 2025] [Safe Robots, retrieved 2026]. What can be said from the available evidence is narrower. Safe Robots presents Clear as software for engineering teams in safety critical robotics, with adjacent references to medical and automotive systems, which suggests the initial market definition is not general developer tooling but the subset of teams that need auditable requirements, hazard analysis, and traceability before deployment [F6S, September 2025] [Safe Robots, retrieved 2026].
A direct sizing figure is absent from the source set, so the more defensible framing is by analogous spend domains rather than asserted market share. The closest public analogues in the research are requirements and engineering workflow systems such as Jira, GitHub, and Jama, because Safe Robots says Clear connects to those systems rather than replacing the entire development stack [F6S, September 2025] [GitHub, retrieved 2026] [OpsHub, retrieved 2026]. That matters because it points to a likely wedge: incremental system-of-record value around safety evidence and traceability, sold into teams already using mainstream engineering tools, instead of a full rip-and-replace platform thesis [F6S, September 2025] [Safe Robots, retrieved 2026].
| Market lens | What the public sources support | Why it is relevant |
|---|---|---|
| Safety-critical robotics engineering | Clear is described as serving robotics teams deploying into the field, with emphasis on requirements, hazards, evidence, and liability workflows [Safe Robots, retrieved 2026] | This is the most direct statement of the company's target problem set |
| Adjacent regulated systems | F6S also references medical and automotive systems alongside robotics [F6S, September 2025] | This suggests a broader long-term category adjacency, though not necessarily current go-to-market focus |
| Requirements and traceability software | Clear is described as connecting with GitHub, Jira, Drive, and Jama to maintain a living model of engineering requirements and evidence [F6S, September 2025] | This places the product near established engineering workflow budgets rather than a greenfield line item |
| Compliance and certification support | The website emphasizes regulatory analysis, standards mapping, hazard analysis, and auditable risk registers [Safe Robots, retrieved 2026] | This indicates overlap with compliance tooling and technical consulting budgets |
The useful read from the table is that Safe Robots appears to be chasing a budget intersection, not a single clean category. That can help an early company find demand, but it also means category definition may remain fluid until customer patterns are publicly visible.
Demand drivers are easier to observe than market size. The company's own messaging repeatedly centers on three pressures: deployment into real environments, rising documentation expectations from customers and regulators, and the difficulty of maintaining traceability when requirements, code, tests, and hazards live in separate systems [Safe Robots, retrieved 2026]. Nestor Grace's prior work at Cruise, covering requirements, safety analysis, testing, verification, validation, and failure-mode monitoring for robotaxis, is directionally consistent with a market need emerging from autonomy programs that have already felt the operational cost of fragmented safety workflows [LinkedIn, October 2026]. That does not prove demand at scale, but it does make the problem selection legible.
The adjacent and substitute markets are also important because many buyers may address this problem without purchasing a dedicated new tool. Teams can continue to rely on combinations of Jira, GitHub, Jama, shared drives, and internal process documentation, especially if robotics programs remain small or pre-commercial [F6S, September 2025] [GitHub, retrieved 2026] [OpsHub, retrieved 2026]. At the other end, larger programs may use sector-specific safety engineering methods, in-house compliance teams, or external consultants. In practice, that means Safe Robots is likely competing not only with named software vendors, but also with entrenched workflow habits and services spend.
Regulatory and macro forces look supportive in principle, even if the source set does not provide sector-wide market reports. Safe Robots' website is explicit that deployment location changes which rules apply and that evidence may be requested by a customer, regulator, or insurer, which is a credible summary of how commercial robotics risk gets scrutinized in the field [Safe Robots, retrieved 2026]. The broader implication is straightforward: as robotics moves into warehouses, hospitals, streets, and other public or semi-public settings, the burden of proof around safety and liability tends to increase rather than decrease. For a young software company, that is a sensible tailwind, but the public evidence still stops short of showing how large that budget pool is or how urgently buyers are standardizing around a dedicated software layer.
Company-stated, unverified -- This section relies primarily on company descriptions from Safe Robots' website and F6S profile, with limited independent corroboration from founder background sources and integration references.
Competitive Landscape
Positioning
MIXED Safe Robots appears to be positioning itself less as a robot simulation vendor and more as the system-of-record layer for safety and quality work inside teams shipping safety-critical robotics, with adjacency into medical and automotive engineering workflows [Safe Robots, retrieved 2026] [F6S, September 2025].
| Company | Positioning | Stage / Funding | Notable Differentiator | Source |
|---|---|---|---|---|
| Safe Robots | AI-enabled safety and quality engineering for robotics teams; product called Clear connects engineering tools and maintains a living model of requirements and evidence | Founded 2025; Berkeley SkyDeck listed as investor | Focus on traceability, hazard analysis, regulatory analysis, and continuous compliance inside existing engineering workflows | [Safe Robots, retrieved 2026] [F6S, September 2025] [Berkeley SkyDeck, retrieved 2026] |
| SafeWorld | Robot-safety simulation software | Seed, $12.2M reported in October 2026 | Public market narrative centers on simulation and stress-testing for robot safety | [FinSMEs, October 2026] [The AI Insider, October 2026] |
| SICK | Industrial sensing and safety technology provider | Established operating company | Installed-base advantage in industrial safety hardware and sensing | [SICK] |
The competitive map breaks into three lanes. First are workflow and compliance systems, where Safe Robots is trying to sit, with Clear described as the place where requirements, hazards, scenarios, risks, standards, and evidence stay linked as the product changes [Safe Robots, retrieved 2026] [F6S, September 2025]. Second are simulation-first challengers such as SafeWorld, whose public positioning is around stress-testing robot behavior in simulated environments rather than managing the documentary and regulatory chain behind a safety case [The AI Insider, October 2026] [FinSMEs, October 2026]. Third are adjacent substitutes such as incumbent tool stacks, including Jira, GitHub, and Jama, which teams already use for issue tracking, software version control, and requirements management, even if those tools are not purpose-built to assemble a robotics safety case on their own [GitHub, retrieved 2026] [OpsHub, retrieved 2026].
Safe Robots' clearest edge today is founder-market fit rather than distribution. Nestor Grace's public background at Cruise is directly tied to requirements, safety analysis, testing, verification, validation, and failure-mode monitoring for autonomous robotaxis, which maps closely to the product's stated emphasis on safety-critical engineering rigor [LinkedIn, October 2026]. That matters because the initial product thesis is not merely that compliance artifacts should exist, but that they should be generated and maintained as a living model across engineering systems, a claim the company repeats both on its site and on F6S [Safe Robots, retrieved 2026] [F6S, September 2025]. The edge is meaningful, but perishable. If larger incumbents or better-funded startups package similar traceability and regulatory workflows into tools customers already run, domain insight alone will not protect the position for long.
The company's main exposure is that it does not appear to own a hard distribution channel, a hardware footprint, or a public capital advantage. SICK, as a named incumbent, has brand recognition in industrial safety and a long-standing position around sensing and factory safety infrastructure, which can matter when buyers prefer audited components from known vendors over new software overlays [SICK]. SafeWorld, meanwhile, has the benefit of a disclosed seed financing event and clearer press visibility, which can help with recruiting, partnerships, and category definition even if its product lane is not identical [FinSMEs, October 2026] [The AI Insider, October 2026]. Safe Robots is also vulnerable to being interpreted as an extension of existing systems such as Jama plus Jira plus internal process, unless it can show that its scenario mapping, hazard analysis, and regulatory logic materially reduce engineering time or certification friction [Safe Robots, retrieved 2026] [OpsHub, retrieved 2026].
The most plausible 18-month scenario is a split market rather than a winner-take-all outcome. Safe Robots has a credible path to become the preferred workflow layer for smaller autonomy and robotics teams if buyers decide that safety evidence, traceability, and standards mapping need a dedicated system rather than a stitched-together toolchain; in that case, Safe Robots is the winner if compliance work becomes continuous and software-defined earlier in the development cycle [Safe Robots, retrieved 2026]. The more difficult scenario is one where simulation-led validation captures budget and mindshare first, pushing documentary and traceability tooling into a secondary purchase decision; in that case, SafeWorld is the winner if simulated safety testing becomes the primary buying wedge, and Safe Robots is the relative loser if customers continue to accept Jira, GitHub, and Jama as adequate for the record-keeping layer [The AI Insider, October 2026] [FinSMEs, October 2026] [GitHub, retrieved 2026] [OpsHub, retrieved 2026].
Single unverified source -- This section combines company materials with limited third-party corroboration. SafeWorld's funding and positioning are supported by independent coverage, but Safe Robots' product claims remain primarily company-sourced, and the SICK comparison rests on high-level public category knowledge rather than a direct product-to-product benchmark.
Opportunity
Opportunity
PUBLIC The prize here is large if Safe Robots can turn safety engineering from a fragmented consulting and documentation exercise into the system of record that robotics teams use before customers, regulators, and insurers force the issue [Safe Robots, retrieved 2026].
The headline opportunity is straightforward: become the default software layer for safety and quality engineering in robotics, then widen from robotics into adjacent safety critical development teams. That outcome is not proven, but it is at least reachable on the public record because the company is not pitching a narrow point tool. Its website describes a workflow that starts with system modeling and extends through regulatory analysis, scenario mapping, hazard analysis, risk modeling, standards compliance, and traceability, all tied together in a single "living model" [Safe Robots, retrieved 2026]. The F6S profile points in the same direction, describing Clear as a product for teams building safety critical robotics, medical, and automotive systems, and stating that it connects tools such as GitHub, Jira, Drive, and Jama to maintain engineering requirements and evidence across the stack [F6S, September 2025]. That matters because systems of record tend to compound once they sit between requirements, code, test evidence, and audit trails.
The public evidence also supports a more specific reason this could work. Nestor Grace's prior work at Cruise covered requirements, safety analysis, testing, verification, validation, and failure mode monitoring for autonomous robotaxis, which is closely aligned with the product surface Safe Robots now describes [LinkedIn, October 2026]. Berkeley SkyDeck lists Safe Robots in its Batch 22 cohort and describes Clear as safety and quality engineering for robotics teams deploying fast [Berkeley SkyDeck, retrieved 2026]. For a company this early, founder domain fit plus a product thesis that mirrors real autonomy workflows is a more credible starting point than a generic AI tooling claim.
| Scenario | What happens | Catalyst | Why it's plausible |
|---|---|---|---|
| Robotics system of record | Clear becomes the default workspace for documenting hazards, requirements, traceability, and evidence across robotics programs, first in startups and then in larger autonomy teams | A visible customer or insurer driven buying moment forces teams to replace spreadsheet and document based processes with auditable software workflows [Safe Robots, retrieved 2026] | The product is already framed around auditability, safety cases, and evidence readiness, which aligns with how regulated or liability sensitive buyers make software decisions [Safe Robots, retrieved 2026] |
| Cross vertical expansion | The company uses the same core workflow to move from robotics into medical and automotive engineering teams that face similar requirements, standards, and traceability burdens [F6S, September 2025] | A product release or early design partner proves the workflow travels beyond robotics into another safety critical category [F6S, September 2025] | The F6S profile explicitly places medical and automotive alongside robotics, suggesting the initial architecture was conceived as a broader safety engineering platform rather than a single niche application [F6S, September 2025] |
| Integration layer for compliance evidence | Clear becomes the connective layer between Jira, GitHub, Jama, and document repositories, making it expensive to rip out once embedded in engineering review and certification work [F6S, September 2025] | One or more durable integrations become central to customer workflows, especially where a code change can automatically expose downstream safety or compliance gaps [Safe Robots, retrieved 2026] | Safe Robots describes the product as a "living model" that shows what a change unsettled, and third party sources support the relevance of Jira, GitHub, and Jama in the engineering toolchain [Safe Robots, retrieved 2026] [GitHub, retrieved 2026] [OpsHub, retrieved 2026] |
The compounding logic is less about consumer style network effects and more about workflow entrenchment. If Safe Robots is adopted at the point where a team defines system behavior, maps scenarios, assigns hazards, and ties claims back to code and tests, each new project adds more structured internal knowledge to the same model [Safe Robots, retrieved 2026]. That can improve retention because the software stops being a place to write reports and becomes the place where safety decisions live. The strongest version of that flywheel would look like this: more integrations bring in more engineering context, richer context improves requirement and hazard mapping, better mapping makes audits and customer diligence easier, and successful audits pull the tool deeper into the development process [F6S, September 2025] [Safe Robots, retrieved 2026]. The public record does not yet show customer proof, but the product design is at least consistent with a high switching cost category if adoption starts.
The size of the win is best framed through a nearby comparable rather than a direct valuation claim. SafeWorld, a separate robot safety software company, disclosed a $12.2 million seed round in October 2026 for simulation based robot safety testing, which signals that investors already view robot safety infrastructure as a venture scale category with room for specialist platforms [FinSMEs, October 2026] [The AI Insider, October 2026]. If Safe Robots were to win the "robotics system of record" scenario above and then extend into adjacent regulated engineering teams, the upside could support a venture scale outcome in the hundreds of millions of dollars of enterprise value (scenario, not a forecast). That estimate is directional rather than modeled here because this public record does not establish revenue, pricing, or customer count, but the underlying logic is plain enough: software that becomes part of compliance, certification, and insurer facing evidence chains can command durable budgets once it is embedded.
Lightly corroborated -- This section relies on company website claims, the F6S profile, LinkedIn, Berkeley SkyDeck, and named publisher reporting on a separate comparable. The core upside case is plausible but early, and material evidence such as customers, revenue, or deployments is not yet independently corroborated.
Sources
Public sources
[Safe Robots, retrieved 2026] Safe Robots | https://www.saferobots.co/
[F6S, September 2025] Safe Robots | https://www.f6s.com/company/try-clear?flow=seePage
[LinkedIn, October 2025] Anthony B - Stealth | LinkedIn | https://www.linkedin.com/in/anthony-bo/
[Berkeley SkyDeck, retrieved 2026] Demo Day Batch 22 - Berkeley SKYDECK | https://skydeck.berkeley.edu/batch22/
[LinkedIn, October 2026] Nestor Grace | https://www.linkedin.com/in/nestor-grace
[GitHub, retrieved 2026] Overview · atlassian/github-for-jira · GitHub | https://github.com/atlassian/github-for-jira/security
[OpsHub, retrieved 2026] Jama JIRA Integration | Jama JIRA Connector Tool | Sync Jama JIRA | https://www.opshub.com/jama-software-integration/jama-jira-integration/
[ContactOut, retrieved 2026] Thejas Kesari | ContactOut | https://contactout.com/
[The AI Insider, October 2026] SafeWorld Emerges From Stealth With $12.2M in Funding to Build Robot-Safety Simulation Software | https://theaiinsider.tech/2026/10/07/safeworld-emerges-from-stealth-with-12-2m-in-funding-to-build-robot-safety-simulation-software/
[FinSMEs, October 2026] SafeWorld Raises $12.2M in Seed Funding | https://www.finsmes.com/2026/10/safeworld-raises-12-2m-in-seed-funding.html
[Companies House, September 2025] SAFE ROBOTICS UK LIMITED overview | https://find-and-update.company-information.service.gov.uk/company/16727334
[Hacker News, May 2020] Jira forcing users to give write access to GitHub and drive | Hacker News | https://news.ycombinator.com/item?id=23202723
Articles about Safe Robots
- From Cruise Requirements to Clear, a Founder Pivots to Robot Safety — The early-stage startup, founded by a Cruise systems engineer, is betting that connecting GitHub, Jira, and Jama can automate the path to certification.