EQTY Lab

• A silicon or cloud vendor, including Intel or NVIDIA themselves, could build confidential and verifiable compute with provenance certificates natively into their own platforms, absorbing the layer EQTY sells, because those same vendors built the underlying method with EQTY.
• The named engagements, spanning customer stories with Masdar and Hala Systems, a partnership with ADQ and a research collaboration with Stability AI and Princeton, all appear on EQTY's own site with no disclosed revenue or customer count, so a procurement team cannot verify the scale of adoption from independent reporting.
• Beyond the $5M RSAC finalist award, no venture round or named investor appears against a hardware and cryptography build, so if no independent revenue figure or funding round appears within a year, the resourcing-to-ambition gap would widen.
• The product depends on specific confidential-computing hardware, Intel Trust Domain Extensions with confidential VMs and NVIDIA H100 and H200 GPUs (EQTY describes extension to NVIDIA's forthcoming Blackwell architecture), so a buyer without that silicon or on a cloud that does not offer it faces a deployment barrier.
• The differentiator is a cryptographic method built with Intel and NVIDIA, so if either partner internalizes verifiable compute or shifts its confidential-computing roadmap, EQTY's technical edge would erode faster than a data or install-base moat would.

EQTY Lab treats the origin and handling of an organization's AI as the thing to be governed, and sells cryptographic tools to prove it. The company targets regulated industries, naming finance, healthcare, and the public sector, where auditors and regulators demand evidence for how AI systems were trained and run. Its pitch is that a buyer can present a tamper-proof record of the lineage of AI data, models, and agents rather than a vendor's assurance.

The problem has independent corroboration beyond the company's own framing. The New Stack, reporting on EQTY's approach, cited MIT's Data Provenance Initiative audit of more than 1,800 datasets finding that companies cannot reliably track where their training data comes from. Data-sovereignty laws and new AI regulation give regulated buyers a concrete reason to document their AI now.

EQTY frames the gap as an AI supply-chain risk. It describes threats to AI systems from data poisoning to information extraction, and argues that autonomous agents acting with less human supervision widen the exposure. That framing places the product against a compliance and assurance obligation rather than a discretionary purchase. [s1, s8, s4]

EQTY Lab sells the AI Integrity Suite, a set of tools that run from the silicon up. At the base is Verifiable Compute, a hardware-based framework the company built with Intel and NVIDIA that notarizes AI training and inference inside secure chips and writes a tamper-proof record of the data and code to a shared ledger. Above it sit Governance Studio, which expresses and audits governance policies, Lineage Explorer, which catalogs an AI system's history, and AI Guardian, which governs AI agents at runtime.

The mechanism is provenance recorded cryptographically at each step. The New Stack described how Verifiable Compute isolates sensitive AI operations and links inputs, compute, and outputs into an auditable record, then stores the signature on a distributed ledger such as Hedera so the history cannot be edited afterward. The suite connects into the tools customers already run, integrating across Databricks Delta Lake, Unity Catalog, and MLflow and into ServiceNow's governance, risk, and compliance platform.

External validation comes from independent technical accounts rather than open benchmarks. NSFOCUS described a full-stack architecture spanning the hardware layer, cryptographic middleware, and application-layer governance tools, and The New Stack walked through the notary and ledger design. No open API portal or independent performance benchmark appeared in the cited record, so a buyer can read the described architecture but cannot yet test the claims directly. [s4, s11, s3, s7]

EQTY Lab sits between two groups of rivals its buyers already weigh. AI-governance platforms inventory AI systems and map controls to regulation, and data-governance incumbents sell lineage and cataloging into the same teams, the two groups the cited record describes and the competitor list names. EQTY's claim to difference is that it proves lineage with cryptography rooted in hardware rather than recording it in software a customer must be trusted to keep accurate.

The distinctive bet is depth of the trust mechanism. Where most governance tools document what an organization asserts about its AI, EQTY produces a record that a third party can verify without trusting the vendor, a harder engineering path that NSFOCUS credited as full-stack vertical integration. That depth is the company's argument for why a buyer should place it above a generalist governance suite.

The durability question is whether the method stays proprietary. The cryptographic approach depends on confidential-computing features from Intel and NVIDIA, so the same silicon and cloud vendors that make the technology possible are also the parties most able to offer it themselves. EQTY's defense is being early and vertically integrated, not a data advantage a rival cannot obtain. [s7, s11, s4]

EQTY Lab's clearest proof is building its core technology with major vendors rather than a long customer list. It built Verifiable Compute with Intel and NVIDIA, a partnership independent outlets including The New Stack and NSFOCUS describe in technical detail, which is stronger evidence than the anonymous logos many governance startups show. The 2025 RSA Conference Innovation Sandbox named EQTY a top 10 finalist, an independent recognition of the approach.

Named customer proof is broader than a single logo. The homepage carries customer stories with Masdar, which deploys EQTY's verifiable retrieval to authenticate AI insights, and Hala Systems, which uses EQTY's agents to support European prosecutors pursuing justice for Ukraine, plus a partnership with ADQ and a neuroscience research project with Stability AI and Princeton. Those references span the public sector, energy, and research without sizing any of it.

Commercial scale is otherwise hard to verify from outside. Every named reference appears on EQTY's own site, the company publishes no customer count or revenue figure, and the strongest outside markers are the RSAC selection and the partner engineering rather than a buyer speaking on the record about production use. The proof of demand trails the strength of the technology. [s13, s11, s6, s5]

EQTY Lab's credibility centers on its founder. Jonathan Dotan is the founder and chief executive, and NSFOCUS describes him as the founding director of the Starling Lab at Stanford University and the University of Southern California, a research group focused on data integrity and human rights. He proposed the verifiable-computing concept the product is built around.

That standing gives the company a credible provenance for its technical claims. The founder's academic base and the company's stated academic advisors tie the work to established research rather than marketing alone, and the RSAC finalist slot put the founder on a recognized industry stage to present the agent-governance product.

What the public record does not show is a deep bench beyond the founder or a named security-industry exit. The company's headcount and the backgrounds of the wider team are not documented in the fetched sources, so the team signal comes from one strongly credentialed founder rather than a broadly evidenced group. [s10, s7, s4]

EQTY Lab's own security posture is not documented in public. No company attestation was found by a probe of the homepage footer, the trust, security, compliance, trust-center, and certifications paths, and the trust and security subdomains as of 2026-07-03. The compliance language on the site describes what the product does for customers, enforcing policies and collecting evidence across AI compliance frameworks, not a certification EQTY itself holds.

The product's job raises the diligence bar rather than lowering it. Because Verifiable Compute processes sensitive AI operations inside confidential-computing environments and writes records to a shared ledger, a security review will ask where that processing occurs, what is written to the ledger, and how tenant data is isolated. The hardware dependency on Intel and NVIDIA confidential computing is itself a procurement question for buyers whose infrastructure does not offer those features.

The absence of a published attestation is a readiness gap for regulated buyers who require one. For a company selling assurance to finance, healthcare, and the public sector, obtaining and publishing its own SOC 2 or equivalent is the step most likely to surface in procurement. [s12, s1, s3, s4]

Add analyzed competitors to compare them side by side with EQTY Lab.

EQTY Lab sells to the organizations that must answer to regulators and auditors for the AI they run. The company names finance, healthcare, and the public sector as its markets, and the RSA Conference described its applications spanning the public sector, life sciences, and finance. These are buyers with a documentation obligation rather than a discretionary interest in AI oversight.

The segment depends on a provenance gap that outside research confirms. The New Stack cited MIT's Data Provenance Initiative audit of more than 1,800 datasets finding that companies cannot reliably track where their AI training data comes from, and data-sovereignty laws and new AI regulation give regulated buyers a concrete reason to record their AI now. EQTY's mission language frames the goal as making every AI system transparent and secure, which it calls AI Integrity.

The motion targets high-assurance accounts rather than a self-serve tier. The named work supports European prosecutors and enterprise clients in regulated fields, so the addressable set is the large regulated or governmental buyer that runs strict security and legal review, not a broad developer base.

The AI Integrity Suite runs from the silicon up rather than as a single application. Verifiable Compute, built with Intel and NVIDIA, notarizes AI training and inference inside secure chips and writes a tamper-proof record of every data object and code to a shared ledger. Governance Studio expresses and audits governance policies, Lineage Explorer catalogs an AI system's history, and AI Guardian governs AI agents at runtime.

The distinctive capability is provenance a third party can verify without trusting the vendor. The New Stack described how Verifiable Compute links the inputs, compute, and outputs of each session into an auditable record and stores the signature on a distributed ledger such as Hedera, so the history cannot be altered afterward. The suite integrates into the tools customers already run, connecting across Databricks Delta Lake, Unity Catalog, and MLflow and into ServiceNow's governance, risk, and compliance platform.

External validation comes from independent technical accounts rather than open benchmarks. NSFOCUS credited a full-stack architecture spanning the hardware layer, cryptographic middleware, and application-layer governance tools, and The New Stack walked through the notary and ledger design. No independent performance benchmark appeared in the cited record, and while an open-source Lineage Explorer offers an entry path, the full suite's depth is described and independently reported rather than benchmark-tested.

EQTY Lab's strongest proof is engineering alongside major vendors rather than a broad customer roster. It built Verifiable Compute with Intel and NVIDIA, a partnership The New Stack and NSFOCUS describe in technical detail, which is stronger evidence than the anonymous logos many governance startups display. The 2025 RSA Conference Innovation Sandbox named EQTY a top 10 finalist.

Named customer proof is broader than a single logo. The homepage carries customer stories with Masdar, which deploys EQTY's verifiable retrieval to authenticate AI insights, and Hala Systems, which uses EQTY's agents to support European prosecutors pursuing justice for Ukraine, plus a partnership with ADQ and a neuroscience research project with Stability AI and Princeton. The references span the public sector, energy, and research without sizing any of it.

Commercial scale stays hard to verify from outside. Every named reference appears on EQTY's own pages or through partners rather than a buyer speaking independently about production use, and the company publishes no customer count or revenue figure. The RSAC selection and the partner engineering are the strongest outside markers.

EQTY Lab does not publish prices. The cited pages show a Start building call to action and no public price or self-serve signup, the pattern of a vendor selling negotiated contracts to regulated accounts through direct engagement.

The implied unit of value is the AI estate under assurance rather than a seat. A product that certifies the lineage of an organization's AI data, models, and agents and enforces policies across it prices naturally against the breadth and risk of what the buyer must govern, though EQTY does not state its model publicly.

The hidden model shapes the entry point. A buyer cannot size a first purchase without contacting the company, and the hardware dependency means the cost conversation also includes the confidential-computing environment the product needs, which a buyer without that infrastructure would have to account for.

EQTY Lab delivers a hardware-rooted service that runs inside confidential-computing environments. Verifiable Compute uses trusted execution features on Intel processors and NVIDIA GPUs to isolate AI operations, and it records provenance to a distributed ledger such as Hedera, so the operating model couples a customer's compute to specific silicon and to a ledger the customer must accept.

The hardware dependency is the central operational question. Because the product relies on Intel Trust Domain Extensions and NVIDIA confidential-computing GPUs, a buyer whose infrastructure or cloud does not offer those features cannot run it as designed, which makes the available hardware a gating item in any deployment.

The diligence surface is what the system processes and records. Verifiable Compute handles sensitive AI operations and writes records to a shared ledger, so a security review will ask where the processing occurs, what is written to the ledger versus kept private, and how tenant data is isolated. The fetched record describes the architecture at a press and technical-article level rather than with published service commitments.

EQTY Lab publishes no security attestation of its own. No company certification was found by a probe on 2026-07-03 of the homepage footer, the trust, security, compliance, and trust-center paths, and the trust and security subdomains. The compliance and certified language on the site refers to what the product does for customers, enforcing policies and collecting evidence across AI compliance frameworks, which is distinct from a SOC 2 or ISO report that EQTY itself holds.

The trust story leans on the founder's research standing instead of a certificate. Jonathan Dotan is the founding director of the Starling Lab at Stanford and USC, a group focused on data integrity, and EQTY cites academic advisors, which lends the technical claims credibility a certification page would otherwise carry.

The missing attestation is the readiness gap most likely to surface in procurement. A company selling verifiable assurance into finance, healthcare, and the public sector will face buyers who require the vendor's own SOC 2 or equivalent, so obtaining and publishing one is the clearest step to close the gap.

EQTY Lab positions itself as a trust layer across an enterprise's AI rather than a feature inside one vendor's stack. It integrates into Databricks and ServiceNow so governance runs where teams already work, and it builds on Intel and NVIDIA hardware and the Hedera ledger, so its reach depends on partnerships with the platforms beneath it.

Those same partners are also the likeliest substitutes. Intel and NVIDIA provide the confidential-computing foundation EQTY relies on, and cloud providers are adding confidential computing of their own, so a deepening native capability inside those platforms could absorb the layer EQTY sells. The partnership that validates the technology also exposes it.

The ecosystem reach is integration depth rather than a builder network. The suite connects to the enterprise data and governance stack, but no developer marketplace or third-party integration catalog appeared in the fetched pages, so the platform claim rests on native integrations and hardware partnerships rather than an outside ecosystem effect.

EQTY Lab's credibility centers on its founder. Jonathan Dotan is the founder and chief executive, and NSFOCUS describes him as the founding director of the Starling Lab at Stanford University and the University of Southern California, a research group focused on data integrity and human rights. NSFOCUS credits him with proposing the verifiable-computing concept behind the product.

That standing ties the company's technical claims to established research. The founder's academic base and the company's stated advisors give the work a provenance beyond marketing, and the RSAC finalist slot placed the founder on a recognized industry stage to present the agent-governance product.

The public record does not show a deep bench beyond the founder. The company's headcount and the backgrounds of the wider team are not documented in the fetched sources, and no named security-industry exit appears, so the team signal comes from one strongly credentialed founder rather than a broadly evidenced group.