Vidai UK

A self-hosted AI gateway and control plane for governing enterprise LLM requests on private infrastructure.

Cover Block

Public sources

Field	Value
Name	Vidai UK
Tagline	A self-hosted AI gateway and control plane for governing enterprise LLM requests on private infrastructure.
Headquarters	Glasgow, United Kingdom [F6S, Jan 2026]
Founded	2025 [Codebase, Oct 2026]
Stage	Pre-seed [FinTech Scotland, April 2026]
Business Model	SaaS
Industry	Deeptech
Technology	AI / Machine Learning
Geography	Western Europe
Growth Profile	Venture Scale
Founding Team	Co-Founders (2): Dr. Priya Bhagavathy and Nagu Gopalakrishnan [vidai.uk, retrieved 2024] [F6S, Jan 2026]
Funding Label	Pre-seed [FinTech Scotland, April 2026]

Executive Summary

PUBLIC Vidai UK is building a self-hosted AI gateway and control plane for enterprises that want to govern, secure, and route large language model traffic inside their own infrastructure, a timely pitch as regulated buyers move from pilot use toward production controls [vidai.uk, retrieved 2024] [FinTech Scotland, April 2026]. The company appears to have started in September 2025, with Glasgow as its base, and public materials consistently frame it as a pre-seed business focused on enterprise AI infrastructure rather than end-user applications [Codebase, Oct 2026] [F6S, Jan 2026] [FinTech Scotland, April 2026].

The product thesis is straightforward: sit in the path of AI requests, give teams visibility into usage and data flows, abstract underlying model providers, and provide a control layer around compliance, latency, and cost, with a community edition released in June 2026 to widen adoption at the developer and pilot stage [vidai.uk, retrieved 2024] [Codebase, Oct 2026] [vidai.uk/faq/, retrieved 2026]. That positioning is potentially differentiated by deployment model, since Vidai emphasizes self-hosting and production use in regulated environments, although most of the detailed capability claims remain company-sourced and still need broader third-party validation [vidai.uk/community/, retrieved 2026] [FinTech Scotland, April 2026].

The founding team is the clearest early strength in the public record. CEO Dr. Priya Bhagavathy is described by the company as an Oxford Martin Fellow with a PhD in Electrical Engineering, while independent profiles tie her to research and R&D work connected to the University of Strathclyde and the University of Oxford; co-founder and CPTO Nagu Gopalakrishnan is presented as an experienced product and engineering leader with prior work spanning Amazon and other large technology organizations [vidai.uk, retrieved 2024] [pureportal.strath.ac.uk, Apr 2026] [Oxford Martin School, retrieved 2026] [FinTech Scotland, April 2026] [LinkedIn, retrieved 2026].

On funding, the public picture is thin but directionally clear: FinTech Scotland labels the company pre-seed, no confirmed round, lead investor, or valuation is disclosed in the source set, and a 2026 post linked to Codebase indicated an active fundraise while Vidai also reached the finals of Scottish EDGE Round 27 [FinTech Scotland, April 2026] [Techscaler, Jun 2026] [LinkedIn, Aug 2026]. The business model is presented as SaaS, with a free community edition available under a controlled licensing flow and commercial use handled case by case, suggesting an open-entry adoption motion paired with enterprise monetization if pilots convert [vidai.uk/community/, retrieved 2026] [vidai.uk/faq/, retrieved 2026].

What matters over the next 12 to 18 months is less category awareness than proof of execution: named enterprise deployments, evidence that the self-hosted gateway can convert community or pilot interest into paid contracts, and clearer disclosure on capitalization and go-to-market traction. The setup is credible enough to warrant attention, but the current public record still leans heavily on company descriptions and ecosystem profiles rather than independently verified commercial milestones [Codebase, Oct 2026] [Tech Supercluster Scotland, Mar 2026].

Company-stated, unverified -- This section relies materially on company materials and ecosystem profiles, with partial corroboration from FinTech Scotland, F6S, Codebase, Oxford Martin School, and University of Strathclyde sources.

Taxonomy Snapshot

Axis	Value
Stage	Pre-Seed
Business Model	SaaS
Industry / Vertical	Deeptech
Technology Type	AI / Machine Learning
Geography	Western Europe
Growth Profile	Venture Scale
Founding Team	Co-Founders (2)
Funding	Pre-seed

How the Company Got Here

PUBLIC Vidai UK is a Glasgow-based enterprise AI infrastructure startup that describes itself as a self-hosted AI gateway and control plane for governing enterprise LLM requests on private infrastructure [Crunchbase, retrieved 2024] [vidai.uk, retrieved 2024]. The public record is still thin, which is typical at this age, but the basic company profile is consistent across the company website and Crunchbase: Vidai UK is headquartered in Glasgow, founded in 2025, and focused on software for enterprise AI control and governance [Crunchbase, retrieved 2024] [vidai.uk, retrieved 2024].

The founding account is directionally clear even where date precision varies by source. Vidai's own materials identify Dr. Priya Bhagavathy as co-founder and CEO and Nagu Gopalakrishnan as co-founder and CPTO [vidai.uk, retrieved 2024]. A later founder case study places the company's start in September 2025, while public startup directories differ on exact founding year, so the safest read is that the company emerged in 2025 and began appearing in public databases in 2026 [Codebase, Oct 2026] [Crunchbase, retrieved 2024].

The milestone trail is early but legible. By early 2026, Vidai UK had public startup-directory presence, including Crunchbase and other ecosystem listings [Crunchbase, retrieved 2024]. By June 2026, according to later public reporting, the company had released a free community edition of its product, and by August 2026 it was publicly associated with the finals of Scottish EDGE Round 27, suggesting a startup moving from formation into market visibility and early distribution rather than one already operating at scale [Codebase, Oct 2026] [Crunchbase, retrieved 2024].

Lightly corroborated -- Core company identity and headquarters are corroborated by Crunchbase and the company website, but the founding timeline and milestone sequence rely partly on later third-party reporting rather than state filings.

Product and Technology

Sources and analysis

Vidai is positioning itself around a narrow but timely infrastructure problem: how to put governance, observability, and policy controls around enterprise use of large language models without routing sensitive traffic through a third-party SaaS layer. Public materials describe the product as a self-hosted AI gateway and control plane that sits in an organization’s AI traffic, with the stated aim of giving teams visibility into usage and data flows while abstracting underlying model providers [vidai.uk, retrieved 2024] [vidai.uk, retrieved 2024]. FinTech Scotland describes the product more simply as a “drop-in AI gateway” intended to move AI pilots into production environments, which is a useful framing because it places Vidai less in model development and more in infrastructure for deployment, control, and operations [FinTech Scotland, April 2026].

The feature set, as described publicly, is broader than routing alone. Company materials say the platform is built to support security, compliance, latency, cost management, and reduced vendor lock-in across enterprise AI usage [vidai.uk, retrieved 2024]. Additional public references describe a production stack spanning gateway functions, observability, guardrails, governance, and prompt management, though that fuller packaging appears in profile-style posts rather than formal product documentation and should be read accordingly [LinkedIn, retrieved 2026]. The technical distribution model is clearer: Vidai maintains a public quick-start repository on GitHub, and its Community Edition is described as a full self-hosted control plane, released in June 2026, available via license generation through the company portal [github.com/vidaiUK/vidai-quickstart, retrieved 2024] [Codebase, Oct 2026] [vidai.uk/faq/, retrieved 2026].

The commercial signal worth watching is the company’s decision to make the community product close to the production product. Vidai says the Community Edition uses the same Rust binary deployed in regulated production environments, while commercial use is approved case by case [vidai.uk/community/, retrieved 2026]. If that claim holds in practice, it gives prospective buyers a low-friction way to test the architecture before procurement, although the public record still does not verify named deployments, reference customers, or a verified demo showing performance under enterprise load [vidai.uk/community/, retrieved 2026] [Codebase, Oct 2026]. A separate public signal, though still early, is that Vidai has also been associated with releasing an open-weight model on Hugging Face built in Scotland, which suggests the team may want credibility across both infrastructure and model tooling rather than acting solely as a gateway vendor [LinkedIn, retrieved 2026].

Company-stated, unverified -- Material product claims rely primarily on company materials, GitHub, and profile-style posts, with partial corroboration from FinTech Scotland and Codebase.

Where the Demand Sits

Public sources The market matters now because enterprise AI adoption is moving from experimentation to governed production, and that shift creates a clearer need for infrastructure that can mediate model access, data handling, and policy control inside regulated environments [FinTech Scotland, April 2026] [vidai.uk, retrieved 2024].

Public evidence on Vidai-specific market size is thin, so the cleanest way to frame demand is through adjacent markets that map to its product surface: enterprise generative AI infrastructure, AI governance and security tooling, and model orchestration layers for production deployment. Vidai positions itself as an AI gateway and control plane that sits in enterprise AI traffic, with emphasis on visibility, compliance, latency, and cost management [vidai.uk, retrieved 2024]. That places it less in end-user application software and more in the infrastructure layer that enterprises buy when pilot usage starts touching sensitive data, regulated workflows, or multi-model architectures [FinTech Scotland, April 2026].

The demand signal the public record does support is sectoral, not numerical. Vidai's own positioning centers on banks, insurers, healthcare organizations, defense companies, and other latency- or scale-critical businesses, while FinTech Scotland describes the product as a "drop-in AI gateway" intended to move AI pilots toward production infrastructure [vidai.uk, retrieved 2024] [FinTech Scotland, April 2026]. Read conservatively, that suggests the company's initial serviceable market is the subset of enterprise AI spend where governance and infrastructure requirements are non-optional, rather than the full software market for generative AI.

Adjacent market signals

Market lens	Relevance to Vidai	Public signal
Enterprise AI infrastructure	Core category fit, since Vidai sells a self-hosted control layer for model traffic and governance	Vidai describes itself as a self-hosted AI gateway and control plane [vidai.uk, retrieved 2024]
AI governance and security	High fit, because the product claims center on visibility, compliance, guardrails, and data-flow control	Vidai and third-party profiles both stress compliance and governed deployment [vidai.uk, retrieved 2024] [FinTech Scotland, April 2026]
Model orchestration and abstraction	Relevant substitute and adjacent spend bucket, since Vidai abstracts model providers and aims to reduce lock-in	Product claims explicitly reference provider abstraction and vendor lock-in reduction [vidai.uk, retrieved 2024]
Self-hosted enterprise software for regulated sectors	Relevant route to adoption, given the emphasis on private infrastructure and production use in regulated settings	The company markets to regulated and high-stakes sectors and offers a self-hosted community edition [vidai.uk/about/, retrieved 2024] [vidai.uk/community/, retrieved 2026]

The table points to a market definition that is narrower than "AI software" and closer to policy, routing, observability, and control infrastructure for enterprise LLM usage. That is a useful distinction, because buyer urgency in this category usually depends less on abstract interest in AI and more on concrete governance requirements.

Two tailwinds appear consistently across the cited material. First, enterprise buyers are being pushed toward multi-model setups, whether for cost, latency, resilience, or sovereignty reasons, and Vidai's provider abstraction pitch speaks directly to that problem set [vidai.uk, retrieved 2024]. Second, self-hosting has become a stronger procurement theme in Europe and regulated industries, where infrastructure location, data residency, and operational control can influence vendor selection; Vidai's community edition and self-hosted control plane language are aligned with that buying preference, even if public evidence of conversion is not yet available [vidai.uk/community/, retrieved 2026] [Tech Supercluster Scotland, Mar 2026].

The substitute markets are also worth naming, because they shape both budget competition and category clarity. Parts of the spend Vidai is pursuing can be absorbed by cloud-native model gateways, broader observability platforms, security layers for AI applications, or internal platform engineering teams building lightweight policy proxies in-house [VentureBeat, 2026]. That does not negate demand, but it does mean the company is selling into a market where the line between a distinct product category and a feature bundle remains fluid.

Regulation and macro conditions appear directionally favorable, although the evidence here is mostly inferential from the sectors Vidai targets and the product attributes it emphasizes. Buyers in financial services, healthcare, defense, and critical infrastructure typically face tighter scrutiny around auditability, data handling, operational resilience, and third-party dependency, which supports the case for an AI control plane rather than direct unmanaged API use [FinTech Scotland, April 2026]. In Europe and the UK, that compliance posture is likely to matter more, not less, as enterprises formalize internal AI policies and procurement controls, though no cited source in this record quantifies the effect on spending.

Lightly corroborated -- Based on one independent company profile and multiple company-controlled sources; market sizing is framed through adjacent categories rather than confirmed third-party TAM data.

Competitive Landscape

MIXED Vidai is positioning itself in a crowded but still unsettled layer of the AI stack: not as a model company, and not as an application vendor, but as the control point that sits between enterprise workloads and whichever model providers those workloads call [vidai.uk, retrieved 2024] [FinTech Scotland, April 2026].

On one side sit incumbent cloud and platform providers that already own enterprise infrastructure relationships and can bundle governance features into broader AI stacks. On another sit younger control-plane and orchestration companies pitching observability, guardrails, routing, and policy enforcement for production AI systems. The sourced example here is OpenClaw, which VentureBeat described as launching a free enterprise control plane for persistent AI agents with backing tied to OpenAI, Red Hat, and Nvidia, a reminder that adjacent platforms are moving quickly to make control layers strategic rather than optional [VentureBeat, 2026]. The substitute set is broader still: some enterprises will try to assemble this layer themselves from cloud-native logging, access control, prompt management, and model-routing components instead of buying a dedicated gateway [VentureBeat, 2026] [vidai.uk, retrieved 2024].

Vidai's edge, based on public evidence, is less about distribution today and more about product architecture and founder-market fit in regulated environments. The company consistently presents itself as self-hosted, model-agnostic, and designed for organizations that care about security, compliance, latency, and vendor lock-in, with target sectors including banking, insurance, healthcare, and defense [vidai.uk, retrieved 2024] [vidai.uk, retrieved 2024] [FinTech Scotland, April 2026]. That positioning is credible because the founders' backgrounds line up with the pitch: Dr. Priya Bhagavathy's work touches critical infrastructure and safety-sensitive domains, while Nagu Gopalakrishnan's background includes low-latency and large-scale systems work [FinTech Scotland, April 2026] [Oxford Martin School] [LinkedIn, retrieved 2026]. The problem is that this edge looks perishable unless it converts into reference deployments, ecosystem partnerships, or a developer community that compounds. A free community edition and public quick-start repo can help widen the funnel, but both are still company-controlled signals rather than independent proof of adoption depth [Codebase, Oct 2026] [github.com/vidaiUK/vidai-quickstart, retrieved 2024] [vidai.uk/community/, retrieved 2026].

The exposure is straightforward. If enterprises decide that AI governance belongs inside the broader platform they already buy from cloud, security, or developer-tooling vendors, then a focused gateway risks being treated as a feature rather than a category. OpenClaw is the clearest named pressure point in the available sources because it is framed as an enterprise control plane and launched with high-profile ecosystem associations, which can matter in a market where trust, integrations, and procurement comfort often outrun product purity [VentureBeat, 2026]. Vidai is also exposed on channel ownership. No named commercial customers, lead investors, or strategic technology partners were verified in the source set, which makes it harder to judge whether it can turn technical resonance into enterprise distribution before better-capitalized platforms close the gap [Codebase, Oct 2026] [FinTech Scotland, April 2026].

Over the next 18 months, the most plausible competitive outcome is not a winner-take-all market but a sorting process by deployment model. OpenClaw is the most obvious winner if enterprise buyers prioritize broad ecosystems, agent-centric workflows, and vendor-backed standardization over strict self-hosting preferences [VentureBeat, 2026]. Vidai is the more interesting winner if regulated buyers insist on private infrastructure, model portability, and a narrower control surface that can be inserted into existing AI traffic without ceding policy control to a hyperscaler or application vendor [vidai.uk, retrieved 2024] [FinTech Scotland, April 2026]. The likely loser, if that second condition does not hold, is the stand-alone gateway category itself, particularly smaller vendors that have product clarity but limited proof of distribution.

Lightly corroborated -- Based on one named adjacent competitor from VentureBeat and multiple public company or directory sources for Vidai's positioning; no independently verified direct competitor set or customer-backed market-share evidence was available.

Opportunity

Public sources If Vidai executes, the prize is not a modest tooling business but a position in the control layer that sits between enterprises and every large language model they use, especially in sectors where latency, auditability, and infrastructure control matter most [vidai.uk, retrieved 2024] [FinTech Scotland, April 2026].

The headline opportunity is straightforward: become the default self-hosted AI gateway for regulated enterprises that want model choice without surrendering governance. That outcome is still early, but it is not purely theoretical. Public materials consistently place Vidai around a specific pain point, the move from AI pilots into production systems, where buyers need visibility into usage, data flows, provider abstraction, and policy controls rather than another point model or chatbot interface [FinTech Scotland, April 2026] [vidai.uk, retrieved 2024]. The company has also taken a route that can matter in infrastructure, namely a community-accessible edition and a public quick-start repository, both of which lower evaluation friction and can widen technical adoption before formal enterprise procurement begins [Codebase, Oct 2026] [github.com/vidaiUK/vidai-quickstart, retrieved 2024].

The upside cases branch from the same core idea: if AI traffic becomes operationally material, the layer that governs it can gain both technical centrality and budget relevance.

Scenario	What happens	Catalyst	Why it's plausible
Regulated-enterprise standard	Vidai becomes a preferred control plane for banks, insurers, healthcare, and defense teams that need self-hosted governance around multi-model AI deployments	A first visible production deployment or reference architecture in a regulated vertical	Vidai already positions explicitly toward regulated and latency-sensitive sectors, and its self-hosted gateway framing matches those buyer requirements [vidai.uk, retrieved 2024] [FinTech Scotland, April 2026]
Open-source to enterprise conversion	The free Community Edition seeds developer adoption, then converts a subset into paid enterprise deployments with stricter governance, support, and approval workflows	Continued uptake of the June 2026 Community Edition and broader use of the public quick-start path	Vidai says the Community Edition is the same Rust binary used in regulated production environments, which supports a credible path from evaluation to enterprise hardening [vidai.uk/community/, retrieved 2026] [vidai.uk/faq/, retrieved 2026]
Sovereign AI infrastructure wedge	Vidai becomes part of the stack for organizations that want locally controlled or nationally anchored AI infrastructure rather than dependence on a single model vendor	Rising demand for sovereign AI and procurement preference for domestically built infrastructure in sensitive workloads	Tech Supercluster Scotland lists the company as a Scotland-built sovereign AI fabric, and Vidai's provider-abstraction message fits a market that wants optionality and lower vendor lock-in [Tech Supercluster Scotland, Mar 2026] [vidai.uk, retrieved 2024]

What compounding looks like here is less about a consumer-style network effect and more about infrastructure depth. One technical win can expand across more models, more internal teams, and more governed workflows because the gateway sits in the request path itself, where observability, routing, policy, and cost controls become more valuable as usage rises [vidai.uk, retrieved 2024]. The early signs are limited and mostly company-linked, but they point in the right direction: a public quick-start repo, a free self-hosted edition, and reported pilot activity across industries suggest Vidai is trying to create a top-of-funnel that begins with engineers and only later reaches formal enterprise procurement [github.com/vidaiUK/vidai-quickstart, retrieved 2024] [Codebase, Oct 2026] [vidai.uk/community/, retrieved 2026]. If that motion works, each deployment could raise switching costs by embedding policies, audit trails, provider abstractions, and operational practices into one control layer.

The size of the win depends on whether this category settles into a durable infrastructure budget line. A useful public marker is that venture and media attention is already clustering around enterprise control-plane software for AI systems, including newer agent control layers positioned as enterprise infrastructure rather than application features [VentureBeat, 2026]. On that basis, a plausible upside case is that Vidai grows into a strategic control-plane asset for regulated AI operations and reaches a value in the high hundreds of millions of dollars or above (scenario, not a forecast), particularly if it can prove that self-hosted governance is a required architectural layer rather than a discretionary tool. That is a large claim for a pre-seed company, but the path is at least legible: own governance at the traffic layer, convert developer adoption into enterprise standardization, and benefit as buyers spread workloads across multiple model providers rather than one [vidai.uk, retrieved 2024] [FinTech Scotland, April 2026].

Lightly corroborated -- Based on one independent profile source and multiple company-linked sources; upside analysis is conditional and public funding or customer proof remains limited.

Sources