The State of
Penetration Testing

Penetration testing is undergoing its most significant transformation in a decade. Regulatory mandates are hardening testing requirements across every major sector. Artificial intelligence is reshaping both the attack surface organizations must defend and the tools testers use to find vulnerabilities. And the traditional once-a-year assessment model is giving way to continuous, integrated security programs.

This inaugural Lion Security Annual Report synthesizes current market data, emerging threat trends, regulatory drivers, and buying pattern shifts to help security leaders make more informed decisions about how - and from whom - they procure penetration testing services.

Sources: Fortune Business Insights, Pentera State of Pentesting 2025, Deepstrike Research 2025. All figures are from independent third-party research and are cited throughout this report.

---

Executive Summary

The global penetration testing market reached an estimated $2.74 billion in 2025 and is on a trajectory toward $6-7 billion by 2033, driven by regulatory pressure, expanding attack surfaces, and a fundamental shift in how organizations think about security assurance. This report examines the forces shaping that growth and what they mean for security leaders evaluating their testing programs.

Five Findings Security Leaders Should Act On

01 — Regulatory mandates have moved from advisory to mandatory.

DORA is fully in force as of January 2025, NIS2 implementation is accelerating, and PCI DSS v4.0's expanded testing requirements became mandatory in March 2025. Organizations in finance, healthcare, and critical infrastructure can no longer treat penetration testing as optional.

02 — The vulnerability landscape is getting more severe.

Critical vulnerabilities in web applications increased 150% in 2024 compared to the prior year. Broken access control remains the #1 finding for the third consecutive OWASP cycle. Cloud misconfiguration underlies 99% of cloud-related breaches.

03 — The PTaaS model is winning.

Over 70% of organizations have adopted some form of Penetration Testing as a Service. PTaaS is growing at 29.1% CAGR - more than double the rate of the overall market - because it delivers real-time visibility, continuous coverage, and platform integrations that point-in-time assessments cannot.

04 — AI is both an emerging threat category and a testing accelerator.

AI and large language model deployments have created a new attack surface that most existing testing programs do not cover. Simultaneously, AI-augmented testing tools are enabling faster reconnaissance and more consistent coverage - when deployed responsibly alongside qualified testers.

05 — The talent shortage is structural and worsening.

Generative AI and large language models will reshape the talent challenge, though the direction is not straightforward. AI-augmented tooling can accelerate reconnaissance, automate repetitive testing tasks, and help smaller teams cover more ground, partially offsetting the availability gap. But the same capabilities are available to attackers. Threat actors are already using AI to discover vulnerabilities faster, generate exploit variants, and scale attacks that previously required significant human expertise. Whatever delivery model an organization chooses, providers who have not integrated AI meaningfully into their testing methodology risk falling behind the threat actors their clients are defending against.

---

The penetration testing market has moved beyond its niche origins as a compliance checkbox and into mainstream security strategy. Sustained double-digit growth, the emergence of dedicated PTaaS platforms, and increasing board-level visibility have transformed the market in a short period of time.

Market Size & Growth Trajectory

Sources: Fortune Business Insights, Grand View Research, Cognitive Market Research. Figures represent blended estimates from multiple analyst forecasts.

The Rise of PTaaS

Penetration Testing as a Service has emerged as the fastest-growing segment in the market, with a projected CAGR of 29.1% through 2029 - more than double the rate of traditional project-based testing. PTaaS platforms grew from approximately $118 million in 2024 to an anticipated $301 million by 2029, as organizations increasingly demand the real-time visibility, platform integrations, and continuous coverage that legacy engagements cannot provide.

Lion Security Perspective

The shift to PTaaS is not just a delivery preference - it reflects a fundamental change in how security teams want to consume testing. Buyers increasingly expect findings to surface in real time, integrate with their existing ticketing and CI/CD workflows, and persist as tracked items rather than static PDFs. Providers who cannot meet these expectations will struggle to compete for the enterprise segment.

North America & Asia-Pacific Lead Growth

North America commands 35-39% of global market share, underpinned by the density of regulated industries, mature cyber-insurance frameworks, and high concentration of enterprise technology firms. Asia-Pacific is the fastest-growing region at an estimated 22.1% CAGR through 2030, driven by rapid cloud adoption, expanding digital infrastructure, and increasingly stringent national cybersecurity frameworks. European growth is accelerating directly as a result of DORA and NIS2 enforcement timelines.

---

The single most significant near-term demand driver for penetration testing is regulatory enforcement. Three major frameworks reached critical milestones in 2025, and their combined effect is compelling organizations that have historically treated testing as discretionary to embed it as a permanent operational requirement.

Sources: European Banking Authority DORA Technical Standards; ENISA NIS2 Guidance; PCI Security Standards Council v4.0 Summary of Changes; U.S. SEC Final Rule on Cybersecurity Risk Management.

Lion Security Perspective

DORA is the most prescriptive penetration testing mandate the financial sector has ever faced. It doesn't just require testing - it specifies threat-led methodology (TLPT), mandates third-party tester accreditation, and requires regulators to be involved in scoping for significant institutions. CISOs in scope should verify their chosen providers meet DORA's tester independence and qualification requirements before engagement.

Budget Impact of Regulatory Drivers

The regulatory wave is translating directly into budget growth. 85% of organizations increased their penetration testing budgets in 2024, and 87% of CISOs plan to maintain or grow investment through 2025. U.S. enterprises spend approximately $187,000 annually on penetration testing, with large enterprises (10,000+ employees) averaging $216,000.

However, budget pressure is also intensifying: 44% of CISOs cited budget constraints as a key limiting factor in 2025, nearly double the 24% who said the same in 2024. This creates a bifurcating market where organizations are spending more overall but are increasingly scrutinising provider value and efficiency. CISOs are looking for providers who can demonstrate clear ROI through remediation tracking, trend data, and continuous coverage - not just an annual report.

---

Understanding what penetration testers consistently discover across organisations is essential for prioritising testing scope and remediation investment. The findings picture in 2024-2025 reflects a landscape where well-known vulnerability classes persist at alarming rates, while cloud and AI-related exposures are introducing new risk vectors that many testing programs are not yet configured to address.

Top Penetration Testing Findings: 2024-2025

Sources: OWASP Top 10:2025; Deepstrike Penetration Testing Statistics 2025; Appsecure Cloud Security Statistics 2025; Fidelis Security threat research.

The Severity Escalation Problem

The sharp increase in critical and high-severity findings is not simply a reflection of better testing - it indicates that attack surfaces are genuinely expanding faster than defenses. The concentration of severe findings in smaller organizations is particularly notable: mid-market and growth-stage companies carry disproportionate risk relative to their security investment, making them attractive targets for financially motivated threat actors.

Cloud Misconfiguration: The Persistent Root Cause

Cloud misconfiguration continues to dominate as a root cause of breaches, not just a testing finding. An estimated 99% of cloud security failures involve misconfiguration, with 82% attributed to human error in IAM policy design, storage bucket permissions, or network security group rules.

For security leaders, this means cloud security assessment should be a standard component of any penetration testing program - not an optional add-on. Testers should be explicitly evaluating IAM posture, storage exposure, workload security, and network configuration across AWS, Azure, and GCP environments, not just scanning for known CVEs.

Lion Security Perspective

We consistently see organizations that conduct thorough application and network testing but have never had their cloud environment professionally assessed. Cloud misconfigurations are often invisible to internal teams precisely because they are legitimate configurations - just wrong ones. A qualified cloud security assessment is one of the highest-ROI tests an organization can commission.

---

Beyond the established categories of network and application testing, several specialised disciplines are rapidly moving from niche to mainstream. Understanding these areas helps security leaders identify gaps in their current programs and ask the right questions of prospective providers.

Every organization deploying AI applications - whether building on top of foundation models or integrating AI into existing workflows - has introduced a new attack surface that traditional penetration testing is not designed to evaluate. The OWASP Top 10 for LLM Applications defines ten risk categories specific to AI systems, including prompt injection, insecure output handling, training data poisoning, and model extraction. OpenAI's own research has highlighted threat actors actively exploring LLM-driven attack capabilities.

AI security testing requires testers with hands-on experience in model architecture, inference behavior, and the specific failure modes of AI pipelines. It cannot be performed by general-purpose testers applying traditional web testing techniques to an AI endpoint.

Assumed breach testing - in which testers begin from a post-compromise starting position rather than attempting initial access from scratch - is gaining significant traction because it answers a question traditional pen testing often cannot: what can an attacker do once they're already inside? Starting from a compromised workstation, a set of valid credentials, or an established network foothold, testers evaluate lateral movement paths, privilege escalation opportunities, and the effectiveness of detection and response controls. This model is particularly valuable for organizations with mature perimeter defenses who want to stress-test their internal posture and incident response capability.

Where traditional red teaming operates covertly against a defensive team that does not know an engagement is occurring, purple teaming makes the collaboration explicit. Red team operators execute attack scenarios while the client's SOC or blue team observers watch, attempt detection, and tune their alerting and response playbooks in real time. The primary deliverable is not a list of findings - it is measurable improvement in detection coverage, validated SIEM rules, and defender skill development. Organizations with mature security operations programs are finding purple team exercises among the highest-value activities they can commission.

Breach and Attack Simulation platforms automate the execution of attacker techniques and tactics (TTPs) against live defensive tooling to validate whether SIEM, EDR, and SOAR controls detect and respond correctly. BAS is not a replacement for manual penetration testing - it lacks the creativity, context, and business logic understanding of a skilled human tester - but as a continuous validation layer between manual engagements, it provides meaningful coverage. Forward-looking security programs are beginning to combine manual PTaaS, periodic red teaming, and continuous BAS into a tiered assurance model.

The annual penetration test is giving way to continuous or program-based testing models that align with modern software delivery cadences. PTaaS platforms now enable organizations to run scoped tests aligned with release cycles, maintain ongoing retesting of known vulnerabilities, and receive real-time findings as they are discovered rather than waiting for a final report. 40% growth in continuous testing adoption was recorded in 2024, with DevSecOps-integrated testing growing at a comparable rate. For organizations running agile development or frequent cloud infrastructure changes, point-in-time testing is structurally insufficient.

Lion Security Perspective: Supply Chain - The Underserved Testing Gap

Supply chain attacks rose 22% in 2025 and now represent 30% of all security incidents. Yet very few penetration testing programs include explicit third-party dependency analysis or simulated supply chain exploitation scenarios. Security leaders should ask prospective providers how they approach supply chain risk as part of scoping conversations.

---

The procurement model for penetration testing is shifting as rapidly as the threat landscape. Understanding the dominant buying patterns, and their trade-offs, helps security leaders align their purchasing approach to their actual operational needs.

Three Dominant Procurement Models

Testing Frequency: The Gap Between Intent and Reality

There is a significant gap between how frequently organizations want to test and how frequently they actually do. Historically, annual testing was considered standard; leading organizations have moved to quarterly assessments aligned with major releases or infrastructure changes; and early adopters of continuous models are now testing daily or with every software release.

The limiting factors are consistent: 48% of organizations report testing less frequently than desired due to tester availability constraints; 44% cite budget; and a significant portion simply lack the internal program management capacity to run frequent engagements effectively. PTaaS platforms address all three barriers by providing on-demand access, predictable subscription costs, and workflow automation that reduces the operational overhead of running frequent engagements.

---

The quality of a penetration test is determined entirely by the skills of the people conducting it. The global penetration testing talent shortage is therefore not just a market dynamic - it is a direct risk factor for buyers who do not scrutinise tester qualifications carefully.

Generative AI and large language models will reshape the talent challenge, though the direction is not straightforward. AI-augmented tooling can accelerate reconnaissance, automate repetitive testing tasks, and help smaller teams cover more ground, partially offsetting the availability gap. But the same capabilities are available to attackers. Threat actors are already using AI to discover vulnerabilities faster, generate exploit variants, and scale attacks that previously required significant human expertise. Whatever delivery model an organization chooses, providers who have not integrated AI meaningfully into their testing methodology risk falling behind the threat actors their clients are defending against.

Workforce Gap at Scale

The global cybersecurity workforce gap stands at 4.8 million unfilled positions, with penetration testing among the most acutely affected specialisations. In the United States alone, over 34,000 penetration testing positions were open as of January 2025. This shortage has two important implications for buyers: it drives up the cost of quality testing, and it creates conditions where less rigorous providers field under-qualified testers to meet demand.

Certifications That Matter

Sources: Infosec Institute Top Pentesting Certifications 2025; Research.com Penetration Tester Career Data 2026; Deepstrike Certification Statistics 2025.

OSCP remains the gold standard for demonstrating hands-on exploitation competency because it requires candidates to successfully compromise machines in a live environment under time pressure - not pass a multiple choice exam. When evaluating providers, ask specifically for the OSCP and advanced Offensive Security credential counts on the team that would be assigned to your engagement, not just company-wide certification totals.

Lion Security Perspective

Generative AI and large language models will reshape the talent challenge, though the direction is not straightforward. AI-augmented tooling can accelerate reconnaissance, automate repetitive testing tasks, and help smaller teams cover more ground, partially offsetting the availability gap. But the same capabilities are available to attackers. Threat actors are already using AI to discover vulnerabilities faster, generate exploit variants, and scale attacks that previously required significant human expertise. Whatever delivery model an organization chooses, providers who have not integrated AI meaningfully into their testing methodology risk falling behind the threat actors their clients are defending against.

What the Talent Shortage Means for Buyers

The talent shortage has created a two-tier market. Tier-one providers with deep, credentialed teams command premium rates and maintain waitlists; lower-cost providers increasingly rely on automation and junior staff to fill demand. Buyers focused solely on price risk receiving assessments that miss the business-logic and chained-vulnerability findings that require experienced, creative testers to surface.

The appropriate response is not to simply buy the most expensive option - it is to ask the right qualification questions during procurement: Who specifically will work on my engagement? What certifications do they hold? What is the ratio of senior to junior testers on a typical project? How does the provider validate tester quality over time? The Lion Security Marketplace evaluation framework formally assesses these dimensions for every listed provider.

---

Every provider listed on the Lion Security Marketplace has been evaluated against a structured criteria framework built specifically for the Lion Security Marketplace. The framework is designed to give security leaders confidence that marketplace providers meet a defined quality threshold - and to surface the differentiators that matter for specific use cases.

The framework organises evaluation across eight categories, each subdivided into specific criteria rated as Foundational (baseline for listing), Signature (positive differentiator), or Horizon (forward-looking capability).

Providers are tiered as Gold (highest qualification threshold), Verified (meets all Foundational criteria with meaningful Signature coverage), or Listed (meets baseline Foundational criteria). Tier placement is reviewed annually and is adjusted based on client feedback, re-evaluation findings, or material capability changes. The full criteria framework is available from Lion Security as a standalone reference document.

---

The following themes represent the most significant forces that will shape penetration testing programs and provider capabilities over the next 12-18 months.

AI Attack Surface Assessment Becomes Standard

As AI-powered applications proliferate across the enterprise, the OWASP LLM Top 10 will become as familiar to security teams as the web application Top 10. Providers who cannot demonstrate credible AI/LLM testing capability will be disqualified from consideration by organizations with material AI deployments - and that category will encompass most of the enterprise market by end of 2026.

DORA Creates a New Benchmark for Financial Services Testing

DORA's Threat-Led Penetration Testing requirements are specific, auditor-reviewed, and cannot be met with standard commercial pen tests. Financial institutions in scope are rapidly discovering that most of their incumbent providers do not meet the independence, methodology, and documentation requirements. This will drive significant provider switching and consolidation around DORA-credentialed firms.

Continuous Testing Displaces Annual Engagements as the Default

The combination of PTaaS platform maturity, CI/CD integration capability, and growing organizational comfort with ongoing assessments will push continuous or quarterly testing from early-adopter to mainstream in the enterprise segment. Annual-only programs will increasingly be viewed as a compliance-minimum rather than a genuine security assurance posture.

Purple Teaming Graduates from Advanced Practice to Common Program Element

As threat detection investment (SIEM, EDR, SOAR) has grown, the question of whether those investments are actually working has become unavoidable. Purple team exercises directly answer that question. Organizations with mature security operations will integrate purple teaming into their annual assurance calendars alongside traditional red and pen testing.

Provider Consolidation Accelerates

Private equity interest in PTaaS has been significant, and platform capability requirements are raising the cost of competition. Smaller boutique providers will increasingly be acquired or absorbed into larger platforms. Buyers should assess provider stability and roadmap transparency as part of procurement - a provider acquired mid-engagement creates operational risk.

The Human-AI Balance in Testing Will Be Scrutinised

Fully automated pen testing platforms will make aggressive claims about coverage and speed. The market will begin to demand clearer disclosure from all providers about the ratio of human vs. automated testing, which finding categories are validated by humans, and what the false-negative rate of automated-only testing is for complex business logic vulnerabilities.

---

This report is a secondary research synthesis. All market data, statistics, and findings cited are drawn from independent third-party analyst reports, academic research, and industry studies published between 2024 and early 2026. Figures presented in this report represent the external research landscape, not proprietary Lion Security data.

Future editions of this Annual Report will incorporate Lion Security Marketplace proprietary data including: provider assessment scores across the eight evaluation categories, aggregated engagement outcome data contributed by marketplace clients, buyer satisfaction metrics, and emerging trend signals from the Lion Security provider network. These additions will be clearly distinguished from externally sourced data.

Primary Sources Cited

About Lion Security

Lion Security is an offensive cybersecurity research and advisory firm that helps organisations select, compare, and manage penetration testing vendors and solutions. We're vendor-neutral, data-driven, and built by experts who've seen what works and what doesn't.

The Lion Security Marketplace is being built to bring transparency, rigour, and buyer confidence to the penetration testing procurement process. By establishing a structured, published evaluation standard, and applying it consistently to every listed provider, Lion Security aims to make it easier for security leaders to find the right testing partner for their specific environment, compliance context, and risk profile.

This Annual Report is the first in a planned series intended to inform the security community, share market intelligence with buyers, and raise the bar for what organisations should expect from penetration testing providers.

---

© 2026 Lion Security. This report is provided for informational purposes. All third-party statistics are attributed to their original sources. Lion Security makes no warranty as to the accuracy of external data cited herein. This document is intended for security leaders and procurement professionals evaluating penetration testing programs.