CodeWall

Autonomous AI agents for continuous offensive security testing, mapping attack surfaces and chaining real exploits.

Cover Block

From the public record

Field	Value
Name	CodeWall
Tagline	Autonomous AI agents for continuous offensive security testing, mapping attack surfaces and chaining real exploits [CodeWall, retrieved 2026]
Headquarters	London, UK [CodeWall, retrieved 2026]
Founded	2026 [LinkedIn, retrieved 2026]
Stage	Pre-Seed [GlobeNewswire, October 2026]
Business Model	SaaS
Industry	Security
Technology	AI / Machine Learning
Geography	Western Europe
Growth Profile	Venture Scale
Founding Team	Solo Founder, Paul Price [LinkedIn, retrieved 2026]
Funding Label	Pre-seed [GlobeNewswire, October 2026]
Total Disclosed	~$4,000,000 [GlobeNewswire, October 2026]

The Short Version

PUBLIC CodeWall is building autonomous AI agents for continuous offensive security testing, and it merits investor attention now because it emerged from stealth in October 2026 with a disclosed $4 million pre-seed round behind a product thesis that aims to shift enterprise security testing from periodic engagements to ongoing external attack simulation [GlobeNewswire, October 2026] [SiliconANGLE, October 2026] [Unite.AI, October 2026]. The company is based in London and was founded in 2026 by Paul Price, with his LinkedIn profile placing the start of the business in February 2026, which makes the speed from formation to financing one of the more notable features of the story so far [LinkedIn, retrieved 2026] [GlobeNewswire, October 2026].

The product claim is straightforward: CodeWall says its agents approach enterprise systems from the outside, operate without credentials or internal access, map attack surfaces, and chain weaknesses into exploitable paths, with the company positioning this as a continuous capability rather than a standard pen test cadence [CodeWall] [Unite.AI, October 2026] [SiliconANGLE, October 2026]. That is a credible wedge if the automation proves repeatable in production, particularly for regulated buyers that already spend heavily on validation but often receive point-in-time findings instead of persistent coverage [IT Brief, October 2026] [GlobeNewswire, October 2026].

On team, the public record is still thin but relevant in one respect: Bloomberg identifies Price as having served as Global Head of Distribution at Morgan Stanley Investment Management Ltd., while public sources reviewed for this section do not identify additional co-founders [Bloomberg Markets] [Dr. Stefan Sambol - OMMAX | LinkedIn]. The financing was led by Osney Capital, with participation from InReach Ventures and Antler, and the company is described as a SaaS business targeting enterprises, with public materials pointing in particular to banks and financial institutions as early design partners or customers rather than named large-scale deployments [GlobeNewswire, October 2026] [IT Brief, October 2026] [TokenPost, October 2026].

Over the next 12 to 18 months, the central questions are whether CodeWall can turn early visibility into referenceable enterprise adoption, broaden product coverage beyond the initial attack-agent workflow, and substantiate efficacy claims with repeatable customer outcomes rather than launch-period examples and company-led demonstrations [GlobeNewswire, October 2026] [Unite.AI, October 2026] [SiliconANGLE, October 2026]. The public evidence today supports a serious early product concept and credible investor interest; it does not yet support firm conclusions on commercial traction, retention, or category leadership.

Single-source, plausible -- Core financing and launch details are corroborated by multiple public sources, but product performance and customer traction remain partly company-led or single-source.

Taxonomy Snapshot

Axis	Value
Stage	Pre-Seed
Business Model	SaaS
Industry / Vertical	Security
Technology Type	AI / Machine Learning
Geography	Western Europe
Growth Profile	Venture Scale
Founding Team	Solo Founder
Funding	Pre-seed, total disclosed about $4,000,000 [GlobeNewswire, October 2026]

The Company in Brief

PUBLIC

CodeWall surfaced quickly in 2026. The public record shows a London cybersecurity startup founded in 2026, with its company materials describing a business focused on autonomous offensive security testing and external attack simulation [codewall.ai, retrieved 2026]. Its public LinkedIn profile also places the company in London and presents the business under the CodeWall name, which is enough to establish location and operating focus, but not enough to confirm a separate legal entity name from filings in the materials reviewed [LinkedIn, retrieved 2026].

The dated milestones are narrow but clear. Founder Paul Price states on LinkedIn that he founded CodeWall in February 2026, and the company emerged from stealth on October 14, 2026, announcing a $4 million pre-seed round led by Osney Capital [LinkedIn, retrieved 2026] [codewall.ai, retrieved 2026]. That sequence matters because it suggests an unusually compressed path from formation to financing, with the public launch arriving less than a year after founding according to the founder profile and company website materials [LinkedIn, retrieved 2026] [codewall.ai, retrieved 2026].

What is still missing from the public file is as relevant as what is present. The company website and LinkedIn page support the basic outline, founder, geography, and category, but they do not, in the sources provided for this section, establish a registered entity name, incorporation date from a state filing, or a broader founding team beyond Price [codewall.ai, retrieved 2026] [LinkedIn, retrieved 2026]. At this stage, the overview reads as a newly launched London security startup with a defined product posture and fresh financing, but still a thin public corporate footprint.

Unconfirmed -- Confirmed mainly by company website and LinkedIn; no state filing or independent corporate registry evidence was provided for this section.

What They Have Built

Mixed sourcing

What stands out first is not the model layer but the operating posture: CodeWall is presenting itself as an external attacker simulation product, meant to test enterprise systems from the outside rather than from a privileged internal foothold [CodeWall, retrieved 2026] [SiliconANGLE, October 2026]. The company describes autonomous "attack agents" that map an organization’s attack surface, identify exploitable paths, and deliver verified proof-of-concept vulnerabilities on a continuous basis [CodeWall Docs, retrieved 2026] [SiliconANGLE, October 2026]. Independent coverage adds a more specific claim about method, reporting that the agents can operate without credentials, source code, or internal access, and can chain weaknesses across a technology stack [Unite.AI, October 2026].

The product framing is also clear about the commercial wedge. CodeWall is positioning autonomous testing as an ongoing capability that could sit alongside, or in some cases replace, periodic penetration tests commissioned once or twice a year [SiliconANGLE, October 2026]. That is a meaningful promise if the system can reliably move from surface discovery to exploit validation with low false positives, but the public materials do not yet provide benchmark data, named technical integrations, or third-party performance validation [CodeWall, retrieved 2026] [Unite.AI, October 2026].

The publicly announced roadmap remains narrow but relevant. In its launch announcement, the company said it plans broader enterprise-environment coverage and workflows for sustained testing, verification, and remediation, which is consistent with moving from point-in-time offensive testing toward a fuller continuous security program [GlobeNewswire, October 2026]. Public reporting also ties the product narrative to real-world demonstrations involving vulnerabilities identified across platforms associated with McKinsey, BCG, and Bain, including one report that attributed the Bain incident to credentials exposed in a public JavaScript file, though those reports establish attack outcomes more clearly than product repeatability across customer environments [afr.com, 2026-04-20] [IT Brief, October 2026].

Single-source, plausible -- Product positioning is corroborated by company materials and independent press, but capability detail and roadmap evidence rely partly on company statements.

Market Size and Demand

PUBLIC The market matters now because enterprises are being pushed to validate real attack paths more often, while AI adoption is widening the surface area that periodic testing was designed for a slower era [SiliconANGLE, October 2026] [Unite.AI, October 2026].

Public evidence on CodeWall's precise addressable market is thin, so the cleanest way to frame demand is through the problem definition rather than a manufactured TAM. The company's positioning sits at the intersection of offensive security, penetration testing, and continuous exposure validation: autonomous agents that test an enterprise externally, without credentials or internal access, and attempt to chain weaknesses into exploitable paths [Unite.AI, October 2026] [codewall.ai, retrieved 2026]. That is narrower than the full cybersecurity budget, but broader than legacy point-in-time pen testing if buyers accept continuous operation as a distinct control category [SiliconANGLE, October 2026].

The immediate demand driver is frequency. CodeWall's public pitch contrasts continuous testing with the standard enterprise pattern of commissioning penetration tests once or twice a year, which suggests the company is selling against cadence as much as against incumbents [SiliconANGLE, October 2026]. That framing aligns with a broader enterprise problem: external attack surfaces change continuously as web apps, third-party integrations, and AI-facing systems update faster than formal audit cycles can keep up [Unite.AI, October 2026] [IT Brief, October 2026].

A second tailwind is buyer anxiety around AI-connected systems. The public reporting around CodeWall's findings at McKinsey, BCG, and Bain centered on exposures tied to high-value digital environments, including Bain's Pyxis platform, where one report said CodeWall accessed nearly 10,000 AI-driven conversations after identifying credentials in a publicly accessible JavaScript file [afr.com, April 2026] [blog.rankiteo.com, April 2026]. Even allowing for the promotional incentive in offensive security disclosure, the episode is relevant to market formation: it gives CISOs and boards a concrete example of how ordinary web hygiene failures can expose sensitive AI-era workflows, and why buyers may pay for attacker-perspective validation rather than static controls alone [afr.com, April 2026].

The substitute set is also clear from the way the company is described. Today, buyers can meet parts of this need through traditional penetration testing firms, red-team engagements, attack surface management platforms, breach-and-attack simulation tools, and internal AppSec testing programs. CodeWall's bet is that autonomous agents can compress several of those motions into a more continuous service layer, especially for externally visible systems [SiliconANGLE, October 2026] [codewall.ai, retrieved 2026]. The practical implication is that budget may initially come from existing testing and validation spend, not from a net-new line item.

Market frame	What public sources support	Relevance to CodeWall
Continuous offensive testing	CodeWall is positioned as a continuous alternative to periodic penetration tests [SiliconANGLE, October 2026]	Suggests budget displacement from recurring pen-test spend
External attack surface validation	Agents are described as operating without credentials, source code, or internal access [Unite.AI, October 2026]	Points to externally facing enterprise environments as the first wedge
AI-exposed enterprise workflows	Reporting on Bain's Pyxis exposure tied CodeWall's work to AI-driven conversations and sensitive client use cases [afr.com, April 2026] [blog.rankiteo.com, April 2026]	Supports urgency where AI tools touch confidential business processes
Financial-sector security testing	Public launch coverage said early customers include banks and financial institutions [cfotech.co.uk, October 2026] [GlobeNewswire, October 2026]	Indicates initial demand may be strongest in regulated, high-loss sectors

The picture that emerges is less about headline market size and more about budget convergence. If enterprises begin to treat continuous attacker-perspective testing as a standard control, CodeWall participates in a larger shift of spend away from episodic consulting and toward always-on validation. If they do not, this remains a feature set competing inside crowded offensive security procurement.

Regulation and macro conditions cut in CodeWall's favor, though the support is still indirect in the cited record. Financial institutions are a reported early focus, and that buyer group already operates under heavier expectations for control testing, resilience, and documented remediation than most sectors [GlobeNewswire, October 2026] [IT Brief, October 2026]. At the same time, a tighter funding environment in security can help products that claim measurable verification rather than broad platform ambition, but CodeWall still has to prove that autonomous exploitation can be deployed safely, repeatedly, and with enough signal quality to earn recurring enterprise budget [Unite.AI, October 2026].

Single-source, plausible -- Market framing is supported by named public coverage and company materials, but no independent third-party TAM/SAM/SOM study was identified in the source set.

Who Else Is Fighting for This

MIXED CodeWall is entering offensive security from a narrow but timely angle: continuous external attack-path testing with autonomous agents, positioned as a substitute for periodic penetration testing rather than a full replacement for the broader security validation stack [SiliconANGLE, October 2026] [Unite.AI, October 2026].

The immediate competitive map is easier to describe by category than by company, because the cited record here does not name direct startup peers. Incumbents sit in traditional penetration testing and red-team services, where buyer trust, compliance familiarity, and established procurement paths still matter. Adjacent substitutes include exposure management, vulnerability scanning, and broader security validation tools, which can satisfy part of the same budget line if security leaders decide they need coverage breadth more than exploit realism. CodeWall's public pitch is that enterprises, particularly banks and financial institutions, need something closer to an always-on external attacker that can operate without credentials, source code, or internal access [GlobeNewswire, October 2026] [Unite.AI, October 2026].

That framing gives the company a credible wedge, but the edge appears perishable unless the product can show repeatable outcomes that incumbent services and adjacent platforms cannot match. The strongest public evidence in its favor is conceptual rather than commercial: the product claims a capability to chain weaknesses across an organization's stack and deliver verified proof-of-concept vulnerabilities continuously, which is a more decision-ready output than a long periodic testing report if it works as described [Unite.AI, October 2026] [CodeWall Docs]. The funding base also matters at this stage. Osney Capital led the $4 million pre-seed round, with InReach Ventures and Antler participating, which gives the company enough capital to build before revenue proof is visible, but not enough by itself to secure long-term category ownership [GlobeNewswire, October 2026].

The company's most visible exposure is that it does not yet appear to own a distribution channel or a compliance-driven moat in public evidence. Traditional pen-testing firms already have enterprise buying relationships, and adjacent platform vendors can argue that continuous monitoring plus validation is operationally safer than autonomous exploitation if a buyer is cautious about agent-led testing. CodeWall has also publicized findings involving McKinsey, BCG, and Bain, including the Bain Pyxis incident reported in April 2026, but those references establish technical visibility rather than customer lock-in [afr.com, 2026-04-20] [IT Brief, October 2026]. In other words, public proof that the system can find something is not yet the same as public proof that a large enterprise will standardize on it.

The most plausible 18-month scenario is a split market. The winner if autonomous external testing becomes a standard line item in enterprise security budgets is CodeWall itself, because its product narrative is already tightly aligned with that use case and the company emerged early enough to define the category language around "attack agents" [GlobeNewswire, October 2026] [SiliconANGLE, October 2026]. The loser if buyers instead decide they want advisory-heavy engagements and broader validation suites over autonomous exploitation is the standalone continuous offensive-testing category, including CodeWall's current positioning, because the available public record does not yet show a wider platform, a named systems-integrator channel, or an installed customer base strong enough to force category adoption [GlobeNewswire, October 2026] [Unite.AI, October 2026].

Opportunity

PUBLIC The prize here is larger than a better penetration-testing tool: if CodeWall can make autonomous external attack simulation reliable enough for large enterprises, it could become part of the standing security control stack rather than a periodic consulting purchase [SiliconANGLE, October 2026] [Unite.AI, October 2026].

The headline opportunity is to become the default continuous offensive testing layer for regulated enterprises that already spend heavily on security assurance but still rely on intermittent human-led exercises [SiliconANGLE, October 2026]. That outcome is reachable, not merely aspirational, because the public record already shows three ingredients that matter at this stage: the company has a defined product thesis around external, credential-free attack agents [Unite.AI, October 2026]; it has raised a pre-seed round led by a specialist investor in Osney Capital with participation from InReach Ventures and Antler [GlobeNewswire, October 2026]; and it has attached its narrative to concrete vulnerability-finding examples involving McKinsey, BCG, and Bain, even though those references should not be read as confirmed customer relationships [IT Brief, October 2026] [afr.com, April 2026]. In plain terms, CodeWall is trying to shift offensive security from a scheduled service to a recurring software control, and category value tends to expand when a budget line moves from episodic to continuous [SiliconANGLE, October 2026].

A second reason the upside is real is buyer pain. CodeWall is explicitly targeting banks and financial institutions, a segment where external attack paths, supplier sprawl, and AI-enabled software deployment all increase the cost of waiting months between tests [GlobeNewswire, October 2026] [IT Brief, October 2026]. If the product can repeatedly surface verified proof-of-concept vulnerabilities without requiring credentials, source code, or internal access, it offers a wedge that is easier to trial than a full internal platform replacement [Unite.AI, October 2026] [CodeWall Docs].

Scenario	What happens	Catalyst	Why it's plausible
Continuous control for financial services	CodeWall wins repeated deployments at banks and other financial institutions, then expands from point assessments into an always-on control for internet-facing assets	A referenceable financial customer or a product release that broadens enterprise-environment coverage and remediation workflows	The company says early customers include banks and financial institutions, and its roadmap explicitly points to broader coverage plus sustained testing and remediation workflows [cfotech.co.uk, October 2026] [GlobeNewswire, October 2026]
Standard layer for AI-era external assurance	Enterprises adopt autonomous attack agents as a standing check on AI-linked applications and exposed assets, especially after high-visibility failures in public-facing systems	Another public incident, or a demonstration that continuous agents catch exploitable paths faster than periodic pen tests	CodeWall has tied its launch to reported findings across major consulting firms, including Bain's Pyxis platform, and positions the product as continuous rather than periodic testing [afr.com, April 2026] [SiliconANGLE, October 2026]
Platformization beyond testing	The company moves from finding attack paths to verifying fixes and orchestrating repeated retesting, which turns a security test into an operating workflow	Product expansion into verification and remediation loops	The launch materials describe planned development for sustained testing, verification, and remediation, which is the product shape required for a durable software platform rather than a single-use scanner [GlobeNewswire, October 2026]

The compounding mechanism, if it appears, is product learning plus workflow entrenchment. Every successful run that identifies a real exploit path should improve how the agent maps exposed systems, prioritizes chains of weaknesses, and packages findings in a way security teams can act on, although the public sources do not yet quantify those improvements [CodeWall Docs] [Unite.AI, October 2026]. If the company also adds verification and remediation workflows as planned, the buyer relationship can deepen from discovery into recurring proof, retest, and governance, which tends to improve retention because the system becomes part of how security teams evidence control effectiveness over time [GlobeNewswire, October 2026].

There is also a distribution flywheel available here, albeit still early. Public examples involving recognizable enterprise targets create attention; attention can open trials with security-conscious buyers; those trials can generate more attack-path data and remediation patterns; and a broader pattern library can improve future detections and customer trust [IT Brief, October 2026] [afr.com, April 2026]. None of that is proven yet in reported commercial metrics, but the shape of the loop is visible in how the company is positioning itself around continuous adversarial testing rather than one-off findings [SiliconANGLE, October 2026].

The size of the win depends on whether this becomes a feature, a product, or a control category. Public sources provided for this report do not include a named market size estimate or a directly cited public comparable valuation, so the cleanest way to frame upside is by business-model translation rather than by unsupported TAM arithmetic. If CodeWall were to become a standard continuous offensive-testing platform for regulated enterprises under the first or third scenario above, the company could plausibly support a venture-scale outcome because recurring security controls with enterprise workflows generally command much higher strategic value than consulting-style testing engagements (scenario, not a forecast) [SiliconANGLE, October 2026] [GlobeNewswire, October 2026]. The burden of proof now sits on conversion from technical narrative to repeatable deployment, but the upside case is straightforward: a credible shift from annual pen test spend to always-on attack validation software creates room for a much larger company than the current $4 million pre-seed financing would imply [GlobeNewswire, October 2026].

Single-source, plausible -- Based primarily on company materials and single-source media reports, with partial corroboration across GlobeNewswire, SiliconANGLE, Unite.AI, IT Brief, and afr.com.

Sources