Black Hat USA 2026 puts cloud identity and SaaS risk in focus

At a glance

What this is: Black Hat USA 2026 is expected to centre cloud, SaaS, identity, and AI risk, with the article arguing that modern attack surface management is being outpaced by visibility and governance gaps.

Why it matters: For IAM, NHI, cloud, and security teams, the message is that identity now governs access across SaaS, APIs, automation, and AI workflows, so control gaps quickly become data exposure paths.

Context

Black Hat USA 2026 is a security conference preview focused on the risk patterns likely to shape research and practitioner discussion in the coming year. The article treats cloud identity, SaaS expansion, and AI-assisted workflows as connected governance problems, not separate domains.

The underlying issue is control drift: organisations are adding SaaS tools, machine identities, and AI-enabled workflows faster than they can maintain visibility, access discipline, and policy consistency. That makes the identity layer, including non-human identities, a practical enforcement point rather than a back-office admin concern.

Key questions

Q: What breaks when SaaS sprawl is treated as a discovery problem instead of a governance problem?

A: Discovery alone cannot answer who owns an app, who still has access, or what happens to licenses and data when people move on. The failure is that the organisation ends up with inventory without lifecycle control, which creates hidden access, wasted spend, and audit gaps across the SaaS estate.

Q: Why do cloud environments increase non-human identity risk?

A: Cloud environments increase non-human identity risk because automation, APIs, and service accounts multiply faster than manual review can keep up. Those identities often carry persistent credentials and broad permissions, which makes them difficult to track and easy to over-extend. If they are not inventoried and rotated, they become durable access paths for attackers.

Q: What are the signs that SaaS security controls are not keeping pace with the current threat landscape?

A: Common warning signs include unclear visibility across SaaS applications, slow response to risky configurations, and dependence on deep product expertise for each app. If security teams cannot quickly assess exposure or remediate issues across many SaaS platforms, their controls are not scaling with the environment. That usually means hidden identity and integration risk remains unaddressed.

Q: Should organisations treat AI features in SaaS tools as a separate risk class?

A: Yes, because embedded AI changes how information is processed, summarised, and redistributed inside business workflows. Governance should distinguish ordinary application usage from AI-enabled data handling, especially where copilots or assistants can ingest regulated, confidential, or operationally sensitive content.

Technical breakdown

SaaS sprawl and cross-tenant visibility gaps

Modern SaaS environments are no longer isolated applications. They are ecosystems of connected apps, APIs, webhooks, and third-party automation that move data across trust boundaries faster than conventional inventory and monitoring processes can follow. The operational problem is not simply that there are more tools, but that data lineage and access paths become fragmented across multiple control planes. When that happens, security teams lose the ability to answer basic questions about where sensitive data sits, who can reach it, and which integrations can move it onward. In identity terms, the access model is distributed across many delegated actors, which makes ownership and review harder to sustain.

Practical implication: build SaaS inventory, data-flow mapping, and integration governance as one control set, not three separate programmes.

Why machine identities become the control plane

The article correctly frames identity as the central enforcement layer in cloud environments because services now depend on service accounts, API tokens, automation pipelines, and AI agents as much as on human users. These are non-human identities, and they often carry broader privileges than their human counterparts because they are created for function rather than for long-term stewardship. The technical risk is that these identities operate continuously, cross-system, and often without meaningful session boundaries. That makes token theft, privilege misuse, and trust relationship abuse especially effective, because the attacker is not breaking the identity model, but using it exactly as configured.

Practical implication: treat machine identity lifecycle, entitlement scope, and token governance as core cloud security controls.

AI-enabled workflows expand the data processing boundary

AI changes cloud data protection because data is no longer only stored or transmitted. It is also interpreted, summarised, and repurposed inside copilots, embedded models, and workflow assistants that are connected to enterprise systems. That creates a larger trust boundary, especially when AI features are turned on inside SaaS platforms without clear governance. The risk is not only model abuse; it is uncontrolled business use of systems that can ingest sensitive data, generate new outputs, and propagate that data into other services. In practice, AI becomes another intermediary in the data path, which means access, logging, and policy enforcement need to follow the workflow rather than the application name.

Practical implication: govern AI-enabled SaaS features as data-processing paths, with explicit review of what information they can ingest and emit.

Threat narrative

Attacker objective: The attacker wants to turn distributed cloud trust relationships into direct access to data and privileged workflows without needing a traditional perimeter breach.

1. Entry occurs through SaaS expansion, exposed APIs, weak authentication, or ungoverned AI-connected workflows that broaden the attack surface.
2. Credential access follows when attackers target tokens, over-permissioned machine identities, or delegated trust relationships between cloud services.
3. Impact comes through data exposure, lateral movement across integrated SaaS systems, or misuse of automated workflows to reach sensitive enterprise information.

Breaches seen in the wild

• Sisense breach 2024: A credential in Sisense's GitLab reportedly opened S3 buckets of customer tokens, passwords and certificates; CISA urged a full reset.
• United Nations breach 2021: Sakura Samurai used exposed Git credentials to reach 100,000+ UNEP staff records, then reported the flaw through the UN disclosure programme.

Read and download The State of NHI & AI Agent Breach Report 2026, covering 200+ breaches impacting Non-Human Identities including AI Agents.

NHI Mgmt Group analysis

Cloud identity sprawl is now a governance problem, not just an inventory problem: the article shows that SaaS, APIs, automation, and AI are all increasing the number of identities that can reach sensitive data. That changes the security question from who owns the application to who controls the delegated access paths. The practical conclusion is that identity governance must extend beyond human accounts into every software actor with data reach.

Machine identities are becoming the default attack path in distributed environments: service accounts and tokens are attractive because they often carry persistent or excessive access and rarely receive the same scrutiny as human users. That creates a structural imbalance between how access is granted and how access is governed. Practitioners should treat non-human identity governance as a primary control plane rather than a niche hardening exercise.

AI adoption is widening the policy gap between data intent and data behaviour: embedded copilots and AI-assisted SaaS tools can process information in ways traditional DLP and access reviews were not designed to observe. The result is a verification trust gap, where policy says one thing and workflow execution does another. Security teams need controls that follow the data path, not just the application boundary.

Misconfiguration remains the common denominator because cloud complexity outruns control hygiene: over-permissioned IAM roles, exposed storage, weak API authentication, and inconsistent encryption enforcement are still the practical failure modes. Black Hat’s relevance is that it surfaces the techniques defenders must assume adversaries will keep using. The right response is continuous governance of entitlements, configurations, and integrations, not periodic clean-up.

From our research library:

• 82% of breaches involved data stored in the cloud, according to IBM (2024).

What this signals

Cloud identity governance is moving toward control of delegated trust rather than control of endpoints alone: the practical issue is no longer just who signs in, but what software actors can inherit access, pass tokens, and move data across SaaS boundaries. That makes machine identity lifecycle management a board-relevant control question, not a niche technical detail.

SaaS and AI together create a policy execution gap: many organisations can write acceptable-use rules faster than they can enforce them across connected applications and embedded assistants. The gap matters because every new workflow can introduce a new access path, a new processing step, and a new exfiltration route.

Black Hat-style research matters because it exposes the next control failure before it becomes routine: if teams wait for incidents to define their programme, they will always be responding to yesterday's architecture. The better signal is whether identity, data movement, and configuration governance are managed as one operating model.

For practitioners

• Map SaaS applications to data owners and trust paths Build a current inventory of SaaS apps, third-party integrations, and automation routes that touch sensitive information, including shadow IT adoption.
• Audit non-human identities as first-class assets Review service accounts, tokens, and automation identities for scope, ownership, expiration, and revocation so machine access is governed like production access.
• Separate AI enablement from data approval Require review before enabling AI copilots or embedded LLM features in SaaS tools that can ingest regulated or confidential data.
• Tighten cloud configuration drift monitoring Continuously check for over-permissioned roles, exposed storage, weak authentication, and inconsistent encryption settings across cloud and SaaS environments.

Key takeaways

• The article's core message is that cloud risk is now shaped by the interaction of SaaS sprawl, machine identities, and AI-enabled workflows, not by infrastructure alone.
• Visibility gaps are the recurring failure mode because they allow sensitive data, tokens, and delegated trust to move faster than governance can track them.
• Security teams should respond by governing identity, integrations, and data movement together, with continuous review of non-human access and AI-connected workflows.

Key terms

• Non-Human Identity (NHI): A digital identity assigned to a non-human entity such as a software application, service account, API key, bot, machine, or AI agent that enables it to authenticate and interact with systems without direct human involvement. NHIs now outnumber human identities in most enterprises by 25 to 50 times.
• Cloud Identity Sprawl: Cloud identity sprawl is the accumulation of users, service accounts, roles, and automation credentials across multiple environments without consistent lifecycle control. It increases the chance that access remains broader than the data or workload now requires.
• Delegated trust: Delegated trust is the decision to let another system or organization issue, validate, or transmit access on your behalf. It is common in cloud and SaaS environments, but it becomes risky when scope, duration, and revocation are not tightly controlled. In NHI governance, delegated trust must be explicit and continuously reviewable.
• SaaS Sprawl: SaaS sprawl is the uncontrolled spread of software-as-a-service applications across teams and business units. It creates fragmented ownership, duplicated functionality, and weak visibility into who can access what. For IAM and NHI teams, the main risk is not only cost but persistent entitlements that outlive business need.

Deepen your knowledge

The NHI Foundation Level course covers NHI governance, machine identity security, and secrets management in a practitioner-focused format. It is designed for security teams that need to govern non-human access as part of a wider identity programme.