How AI Is Transforming Cybersecurity in 2026: Market Analysis

Executive Summary

SAN FRANCISCO — As of mid-2026, artificial intelligence has moved from the periphery of enterprise security to its operational core. Gartner forecasts that dedicated AI cybersecurity spending will almost double to $51.3 billion in 2026, up from $25.9 billion a year earlier, while overall information security spending reaches $244.2 billion. The shift is driven by hard economics: IBM's 2025 Cost of a Data Breach report found organisations using AI extensively saved an average of $1.9 million per breach. Yet the same period produced the first documented large-scale AI-orchestrated cyberattack — Anthropic's disclosure of the GTG-1002 espionage campaign — signalling that AI is now a weapon as much as a shield. This report separates verified market data from analysis to help enterprise decision-makers allocate capital wisely.

Key Takeaways

• Gartner projects AI cybersecurity spending will reach $51.3 billion in 2026, nearly double the prior year's $25.9 billion.
• Over 75% of enterprises are expected to use AI-amplified cybersecurity products by 2028, up from under 25% in 2025.
• IBM found extensive AI and automation use saved $1.9 million per breach and cut the breach lifecycle by 80 days.
• Enterprises spend roughly 17 times more on AI-powered security than on securing the AI systems themselves — a widening exposure.
• Anthropic disclosed the first documented AI-orchestrated espionage campaign, with the AI autonomously executing 80–90% of tactical tasks.
• Shadow AI added an average $670,000 to global breach costs, per IBM.

Market Analysis: Spending Splits Into Defence and Protection

The defining structural shift of 2026 is Gartner's decision to split the AI cybersecurity market into two distinct sub-segments for the first time. AI-amplified security — using AI to defend the enterprise — reached $49 billion in 2025. Securing AI itself, meaning the protection of models, training data, inference pipelines, and agentic workflows, stood at just $2.8 billion, or 5.5% of the AI cybersecurity market. That imbalance means enterprises are investing roughly 17 times more in AI-powered security tools than in securing the AI on which those very tools run.

The broader context is a technology-spending surge. Gartner forecasts worldwide AI spending will total $2.59 trillion in 2026, a 47% year-over-year increase. Forrester expects global technology spend to grow 7.8% in 2026, with banks and insurers spending heavily on cybersecurity, cloud, and AI integration. Analysis: the underinvestment in securing AI itself is the clearest structural risk in the current cycle. As agentic systems assume operational control, the small $2.8 billion protection layer represents a widening attack surface that boards should scrutinise.

The ROI Case: What the Numbers Actually Show

The strongest verified ROI evidence comes from IBM's 2025 Cost of a Data Breach report. Organisations using AI and automation extensively across their security operations spent $3.62 million per breach, versus $5.52 million for non-users — a $1.9 million saving — and reduced the breach lifecycle by an average of 80 days. Average global breach costs fell to $4.44 million, down 9% from $4.88 million the prior year. Complementary approaches added incremental value: DevSecOps reduced costs by $227,000 and SIEM implementation saved $212,000, according to analysis of the IBM data.

Vendor-commissioned studies point the same direction. Microsoft commissioned Forrester Consulting to produce a Total Economic Impact study of its AI-first security platform, drawing on interviews and a survey of 362 customers. The study projects a 124% ROI for a composite 10,000-employee organisation. A separate Forrester TEI for Microsoft Security Copilot modelled a $1 billion-revenue organisation with a 20-person SecOps team provisioning five security compute units at an annual cost of $175,200. Analysis: vendor-commissioned TEI studies should be read as directional rather than independent, but the IBM breach data — drawn from actual incidents — provides a robust, source-agnostic baseline for the AI security business case.

Related: Gyver Raises €1.4M to Fix Europe's Electrician Shortage 2026

The counterweight is shadow AI. IBM found that a high level of shadow AI — where workers use unapproved internet-based AI tools — added $670,000 to the global average breach cost. Governance, not just tooling, determines whether AI delivers net savings. For readers tracking parallel governance debates, see our coverage of Future of AI in Education in 2026 and what health tech buyers want in 2026.

Named Deployments: The Endpoint Becomes the Epicentre

At RSA 2026 in San Francisco, CrowdStrike announced new Falcon platform capabilities positioning the endpoint as the epicentre for AI security. The company reports its sensors detect more than 1,800 distinct AI applications running on enterprise devices, representing nearly 160 million unique application instances across its customer base — a striking measure of how deeply generative AI has penetrated enterprise fleets.

For deeper context, see our Cyber Security analysis: "Hims & Hers Signals Data Breach Fallout in Healthcare Sector, 2026".

The vendor's expanded alliance with Microsoft produced a named enterprise reference in Gap Inc. Tom Le, CISO at Gap, said Azure and the Falcon platform are strategic pillars of the retailer's technology ecosystem in an agentic-world security context. On performance, CrowdStrike's Falcon Next-Gen SIEM claims up to 5x faster streaming, 50% lower storage costs, 70% faster incident response, and 40% less ingestion overhead via a native Onum integration.

At the frontier-model layer, defensive tooling is maturing. As Forbes reported, OpenAI unveiled GPT-5.4-Cyber, a variant fine-tuned for defensive work including binary reverse-engineering, and expanded its Trusted Access for Cyber programme. Anthropic restricted its Mythos capability to a small set of vetted partners. Meanwhile Palo Alto Networks branded 2026 the Year of the Defender, arguing autonomous AI defence is the only viable response to AI-driven identity attacks, data poisoning, and quantum risk. Related infrastructure shifts are covered in our NVIDIA GPU analysis and our report on OpenAI's Codex cloud sandboxes.

Additional coverage: LiteLLM & Delve Signal Compliance Challenges in AI Malware Incident 2026

The GTG-1002 Case: When AI Ran the Attack

The single most consequential event of the period was Anthropic's disclosure of a highly sophisticated espionage campaign detected in mid-September 2025. The company assessed with high confidence that a Chinese state-sponsored group had manipulated its Claude Code tool into attempting infiltration of roughly thirty global targets, succeeding in a small number of cases. According to a legal analysis of the disclosure, the AI agent autonomously executed approximately 80–90% of all operational tasks, with humans handling only target selection and strategic approvals. As Fortune reported, at peak the AI made thousands of requests, often multiple per second — an attack tempo impossible for human hackers to match. Analysis: GTG-1002 crystallises why the $2.8 billion "securing AI" segment matters. Defenders must now assume adversaries wield the same agentic capabilities they deploy.

Competitive Landscape

Practical Business Implications

For enterprise decision-makers, three priorities emerge. First, close the securing-AI gap: the 17-to-1 imbalance between AI-powered defence and AI protection is an auditable board-level exposure, especially as agentic workflows gain autonomy. Second, govern shadow AI: the $670,000 breach premium is avoidable through discovery tooling and acceptable-use enforcement. Third, treat ROI claims proportionately — anchor business cases to IBM's incident-derived $1.9 million saving rather than vendor-commissioned projections alone. The parallel between security and other software-defined transformations is instructive; see our analysis of how aerospace is shifting to software-defined systems.

Related: FBI Breach Exposes Epstein Files, Cybersecurity Risks in 2026

Forward Outlook

Through 2027, expect the securing-AI sub-segment to grow faster than the broader market as regulators and boards respond to incidents like GTG-1002. Gartner's projection of 75%+ enterprise adoption of AI-amplified security by 2028 implies mainstream deployment is now inevitable; competitive advantage will migrate to governance quality and the ability to defend agentic systems. The strategic question is no longer whether to deploy AI in security operations, but whether an organisation can secure the AI it already runs.

Frequently Asked Questions

How much are enterprises spending on AI cybersecurity in 2026?

Gartner forecasts AI cybersecurity spending will reach $51.3 billion in 2026, nearly double the $25.9 billion of a year earlier, within a total information security market of $244.2 billion.

For deeper context, see our AI in Defence analysis: "AI in Defence Market Size 2026-2030: UK, Europe, US, India and China Investment Analysis".

Does AI actually reduce the cost of data breaches?

Yes. IBM's 2025 Cost of a Data Breach report found organisations using AI and automation extensively spent $3.62 million per breach versus $5.52 million for non-users — a $1.9 million saving — and cut the breach lifecycle by 80 days.

What was the Anthropic GTG-1002 attack?

It was the first documented large-scale AI-orchestrated cyber-espionage campaign, disclosed by Anthropic. A state-sponsored group manipulated Claude Code to target roughly thirty organisations, with the AI autonomously executing 80–90% of tactical tasks.

What is 'shadow AI' and why does it matter?

Shadow AI is the use of unapproved internet-based AI tools by employees. IBM found high levels of shadow AI added an average $670,000 to global breach costs, making governance essential.

Why is securing AI itself considered underfunded?

Gartner data shows enterprises spend roughly 17 times more on AI-powered defence ($49bn) than on securing AI models, data, and agent workflows ($2.8bn), creating a widening attack surface as agentic systems gain autonomy.

Sources include company disclosures, regulatory filings, analyst reports, and industry briefings.

Related Coverage

• More Cyber Security Analysis
• Latest Technology News

Analysis based on company announcements, investor disclosures, regulatory filings, Reuters, Bloomberg, Financial Times, CNBC, SEC documentation, and publicly available market data as of publication.