Vendor claims below are dated at first sourcing and re-checked periodically; see the Source & Date column in the comparison table. We hold D3 Morpheus to the same disclosure standard we apply to every other vendor on this page.
The Short Answer
The best agentic SOC platform for an MSSP in 2026 is the one whose pricing structure and tenant architecture protect your margin as client alert volumes fluctuate. For a service provider, the agentic platform is your cost of goods sold. For MSSPs that want the full alert lifecycle (autonomous investigation, response, and case management) on one engine with hard tenant isolation, white-label UI, and subscription pricing that keeps cost-to-serve flat, D3 Morpheus leads the category. Conifers CognitiveSOC is the strongest mesh-architecture alternative, built MSSP-first, with vendor-stated tenant onboarding in hours. Torq suits large MSSPs with dedicated automation engineering teams and a workflow-first operating model. Dropzone AI offers the fastest entry for smaller books of business, with the caveat that per-investigation pricing couples your COGS to client noise. Ecosystem-native agents (SentinelOne Purple AI, CrowdStrike Charlotte AI, Microsoft Security Copilot) fit MSSPs standardized on a single vendor’s stack, a condition few multi-client books actually meet.
The nine platforms compared in depth below:
- D3 Morpheus: unified agentic engine; native multi-tenancy, hard isolation, white-label, subscription economics
- Conifers CognitiveSOC: MSSP-first multi-agent mesh; vendor-stated tenant onboarding in 2–4 hours
- Torq: hyperautomation platform with HyperAgents; proven at large MSSPs and service providers
- Dropzone AI: focused AI analyst with a dedicated MSSP program; per-investigation pricing tiers
- Stellar Cyber: agentic AI inside an Open XDR platform; strong mid-market MSSP fit
- SentinelOne Purple AI (Athena): for MSSPs building on SentinelOne’s stack
- CrowdStrike Charlotte AI: for Falcon-based service delivery
- Microsoft Security Copilot + Sentinel: for Azure Lighthouse MSSPs in Microsoft-heavy client bases
- Radiant Security: high-coverage triage layer atop clients’ existing stacks
This guide covers the MSSP lens specifically. For the full category view, see The 12 Best Agentic SOC Platforms in 2026, including the four-architecture taxonomy and the AL1–AL4 autonomy model referenced throughout this page.
Why MSSPs Are a Different Evaluation Entirely
An enterprise buys an agentic SOC platform to reduce its own alert burden. An MSSP buys one to change the economics of its business. That difference rewrites every evaluation criterion.
The MSSP business model has a structural problem the whole industry knows and rarely says plainly: revenue scales with clients, but cost has historically scaled with analysts. Every new client brings a new alert stream, a new tool stack, and a new slice of 24/7 coverage obligation. D3 Security’s 2026 analysis of over 1,000 US security job postings found that the median security operations posting names eight tools. For an MSSP, that’s eight tools per client, multiplied across the book. The same research found that MSSPs and MDR providers have quietly become the industry’s employer of first resort: nearly every genuinely entry-level SOC job in the dataset was at a service provider, not an enterprise. MSSPs are staffing the industry’s talent pipeline while running the industry’s thinnest margins.
Agentic platforms are the first technology category that credibly breaks the analyst-headcount coupling. But which platform you choose determines whether you actually capture that margin, or hand it back through the pricing model.
The MSSP Margin Math: Four Pricing Traps
Before comparing platforms, run every candidate through this framework. It is vendor-neutral; use it in your own RFPs.
Trap 1: Per-alert and per-investigation pricing. Your COGS is now coupled to your noisiest client. A client’s misconfigured detection rule becomes your invoice. Worse, it creates a quiet incentive to filter alerts before ingestion: a security decision disguised as a cost decision, and one your SLA lawyers should hear about before your CFO does.
Trap 2: Per-agent compute and per-workflow execution pricing. The bill scales with incident volume, which means your worst weeks, the weeks you earn your retainer, produce your biggest costs. Surge pricing on the service you exist to provide.
Trap 3: Per-GB ingestion pricing. Cost tracks data volume, not value delivered. Onboarding a log-heavy client punishes you before you’ve triaged a single alert.
Trap 4: Per-seat pricing that counts your analysts. If the platform’s value is doing analyst work, pricing by analyst seat taxes exactly the leverage you bought it for.
The alternative is subscription economics: a fixed platform fee that absorbs AI compute, so cost-to-serve stays flat while the client book grows. The floor is higher; the ceiling is predictable. For a business whose margin is the difference between revenue per client and cost per client, predictability is the product you are buying.
The Multi-Tenancy Depth Test
“Multi-tenant” appears on every vendor’s MSSP page. It means at least three different things:
Level 1: UI partitioning. One environment with client filters. Fine for reporting; indefensible when a client’s auditor asks how their data is segregated from a competitor’s.
Level 2: Logical isolation. Separate tenant workspaces on shared infrastructure, with role-based cross-tenant administration. The minimum for serious MSSP work.
Level 3: Hard isolation with white-label and per-tenant configuration. Full data isolation per tenant, MSSP-branded interfaces for client-facing views, per-tenant detection tuning and autonomy policies, and billing separation. This is the level at which the platform stops constraining how you package and price your own service.
Ask every vendor which level they actually ship. Ask to see tenant onboarding performed live, with a stopwatch. Time-to-tenant is your revenue activation speed.
How We Evaluated These Platforms
Each platform was assessed on seven MSSP-specific criteria, alongside the architecture and autonomy scoring from the full category comparison. Where a vendor’s claim could not be independently verified, it is labeled vendor-stated.
- Pricing structure under the four-trap test: how does cost behave when a client has a bad week?
- Multi-tenancy depth: Level 1, 2, or 3, per the test above
- Time-to-tenant: how fast a new client goes from contract to autonomous triage
- White-labeling: can the platform disappear behind your brand?
- Per-tenant autonomy governance: can Client A run AL3 while risk-averse Client B stays at AL2? (Autonomy levels per the AL1–AL4 model.)
- Cross-client stack coverage: because your clients did not coordinate their tool purchases with each other
- Audit artifact quality: when a client’s regulator asks why the AI contained a host, what do you hand them?
The 9 Best Agentic SOC Platforms for MSSPs in 2026
1. D3 Morpheus: Best Overall Agentic SOC Platform for MSSPs
Architecture: Unified Agentic Engine · Autonomy ceiling: AL4 (bounded, policy-gated) · Multi-tenancy: Level 3, native, hard isolation, white-label
D3 Morpheus is built around the exact economics the four-trap framework describes. The platform runs on subscription pricing with AI compute absorbed into the fee, so an MSSP’s cost-to-serve stays flat whether a client’s alert volume doubles or a ransomware weekend triples the investigation load. One reasoning engine, the Cyber Triage Reasoning Graph, investigates every alert end-to-end at L2+ depth, generates response workflows at runtime, and executes them through built-in orchestration, producing one audit trail per incident, per tenant.
Why MSSPs specifically choose it:
- Margin protection by design: Subscription economics with compute absorbed. Up to 95% of alerts autonomously investigated at L2+ depth in under 2 minutes (D3-verified customer-reported metric, Jul 2026). Analyst hours become leverage, not COGS.
- Level 3 multi-tenancy: Hard data isolation per tenant, white-label UI for client-facing views, per-tenant detection tuning, and per-tenant autonomy policy. One client can run bounded-autonomous response while another stays human-approved, on the same engine.
- Cross-stack coverage across 800+ self-healing integrations: Your clients’ stacks are heterogeneous by definition; the median job posting’s eight tools are eight different tools at every client. Self-healing integrations detect API drift and regenerate connector code autonomously, which matters at MSSP scale, where one vendor’s API change would otherwise break workflows across the entire book at once.
- The audit answer: One incident yields one replayable decision trail (evidence tree, logic chain, confidence scores) per tenant. When a client operates under NIS2, DORA, or EU AI Act obligations, you hand their auditor one artifact, not a log-stitching project. → EU AI Act & SOC automation
Limitations: Autonomous depth is a function of integration coverage. Clients with unusual or thin stacks see proportionally thinner investigations until connectors are wired. Onboarding a tenant with a complex bespoke stack takes real integration work upfront. MSSPs with very small books (a handful of clients, low alert volume) may find the platform floor higher than a per-investigation product’s entry price.
Best for: MSSPs and MDR providers who want the full alert lifecycle on one engine, flat cost-to-serve at scale, and tenant isolation they can defend in a client’s audit.
→ Morpheus for MSSPs / multi-tenant operations · Pricing
2. Conifers CognitiveSOC: Best MSSP-First Multi-Agent Mesh
Architecture: Multi-Agent Mesh · Autonomy ceiling: AL3 · Multi-tenancy: Native; tenant onboarding in 2–4 hours (vendor-stated)
Conifers is the rare platform designed for service providers from its first commit. CognitiveSOC runs a mesh of task-specific agents sharing memory across the defense lifecycle. A hunt informs detection engineering; an investigation improves intelligence. It tunes detections, hunts, and investigations per tenant. CEO Tom Findling states that onboarding a tenant takes two to four hours (MSSP Alert, May 2026). We have not tested that independently.
Strengths: The strongest MSSP-native design in the mesh category; shared-memory fabric across agents is a real architectural idea; works atop the tools clients already run, so no client-side rip-and-replace conversation; a $25M round from SYN Ventures (January 2025) and a later undisclosed investment from Washington Harbour Partners support the roadmap.
Limitations: Younger platform with a smaller reference base than the category leaders. The mesh architecture carries the standard audit-composition burden: each agent writes its own log, and someone (you, at client-audit time) composes them into one narrative per incident. Quote-based pricing; test it against the four traps.
Best for: MSSPs that prefer mesh composability and per-tenant personalization at scale, and have the operational maturity to own audit composition.
3. Torq: Best for Large MSSPs with Automation Engineering Teams
Architecture: Multi-Agent Mesh on a hyperautomation platform · Autonomy ceiling: AL3 in production · Multi-tenancy: Yes
Torq pairs its no-code workflow platform with HyperAgents coordinated by Socrates, its OmniAgent orchestrator, and has real service-provider proof: RSM has publicly credited Torq HyperSOC with driving efficiency for RSM Defense and its customers (Torq press release, September 2024). The company’s momentum is unambiguous: a $140M Series D at a $1.2B valuation announced January 2026, and hundreds of multinational enterprise customers. Torq states that Socrates auto-remediates up to 95% of Tier-1 incidents without human intervention. That figure is Torq’s own, not an analyst finding.
Strengths: Best-in-class workflow authoring; genuine multi-agent investigation; enterprise-grade brand roster; a strong fit for MSSPs whose differentiation is custom automation engineering per client.
Limitations: The operating model assumes you build and maintain the workflow inventory, per client. That is a product-engineering function inside your service business, and its cost belongs in your margin model alongside the license. Pricing combines a base fee with per-workflow execution and per-agent compute (see Trap 2): cost couples to incident volume, so surge weeks bill accordingly. Expect six-figure commitments.
Best for: Large MSSPs with dedicated automation engineering capacity that sell bespoke automation as part of the service.
→ Evaluating the two architectures side by side? Morpheus vs. Torq
4. Dropzone AI: Best Entry Point for Smaller Books of Business
Architecture: Focused AI Analyst · Autonomy ceiling: AL2–AL3 · Multi-tenancy: Yes, with a dedicated MSSP program
Dropzone’s AI SOC analyst investigates alerts 24/7 at L2 depth with fast cloud onboarding and pricing tiered by investigation count. The Standard tier covers up to 4,000 full investigations per year per AI analyst. Dropzone published list pricing through 2025 and moved to quoted tiers in 2026 (vendor pricing page, checked August 2026). For an MSSP with a small book and a triage bottleneck, it is the fastest credible entry into agentic operations.
Strengths: Genuinely fast time-to-value; investigation write-ups analysts trust; published investigation tiers in a category that mostly hides its pricing structure; a real MSSP program, not an afterthought.
Limitations: Per-investigation pricing is Trap 1 in its purest form: at MSSP scale, your noisiest client sets your unit economics, and the incentive to pre-filter ingestion grows with the book. Response execution, orchestration, and case management remain thinner than platform-class products, so the rest of your service delivery stack stays where it is.
Best for: Smaller MSSPs (roughly 20–100 alerts/day per client, modest books) that need 24/7 autonomous triage now and will revisit platform economics as the book grows.
5. Stellar Cyber: Best Open-XDR Route to Agentic Operations
Architecture: Agentic AI embedded in an Open XDR platform · Autonomy ceiling: AL2–AL3 · Multi-tenancy: Yes, MSSP-oriented
Stellar Cyber embeds agentic capabilities inside an Open XDR platform with long-standing MSSP focus and mid-market fit. For providers that want detection and the agentic layer from one vendor, while staying open to clients’ existing tools, it is the pragmatic consolidation play.
Strengths: One throat to choke across detection and triage; established MSSP licensing motion; sensible fit for mid-market client profiles.
Limitations: The agentic layer is newer than the XDR core. Test autonomy depth and audit artifacts against the AL model in a POV, not from datasheets. Consolidating detection and agentic layers with one vendor also concentrates switching costs.
Best for: Mid-market MSSPs that want platform consolidation and are early enough in their agentic journey to grow with the vendor.
6. SentinelOne Purple AI (Athena): Best for SentinelOne-Standardized Service Delivery
Architecture: Ecosystem-Native Agent, expanding via Athena · Autonomy ceiling: AL3 · Multi-tenancy: Partial
For MSSPs that have standardized service delivery on SentinelOne, Purple AI is the natural agentic extension. Investigation quality on S1 telemetry is excellent, and the Athena release, announced April 2025, begins opening third-party SIEM and data-lake support, softening the ecosystem boundary that traditionally limits native agents.
Strengths: Deep endpoint-native investigation; clean upgrade path for S1-based MDR offerings; Athena’s cross-environment ambitions lead the ecosystem-agent class.
Limitations: Multi-tenancy support is partial, not MSSP-native. Third-party depth via Athena is new and unproven relative to vendor-agnostic platforms, and SentinelOne stated general availability at announcement only for Purple AI Auto Triage. Test it on a real cross-stack incident before building service SLAs on it. Add-on module economics on tiered platform licensing.
Best for: MSSPs whose service catalog is built on SentinelOne and whose clients’ alert mixes are endpoint-dominant. For the cross-stack remainder, providers typically pair a vendor-agnostic layer. → Morpheus for SentinelOne
7. CrowdStrike Charlotte AI: Best for Falcon-Based Offerings
Architecture: Ecosystem-Native Agent · Autonomy ceiling: AL3 within Falcon · Multi-tenancy: Partial for Charlotte specifically
Charlotte brings agentic triage and response inside Falcon. CrowdStrike states that its Detection Triage release (February 2025) matches the expert decisions of its own Falcon Complete MDR team more than 98% of the time, a vendor-internal benchmark, not an independent audit. Charlotte Agentic SOAR followed in November 2025 with no published accuracy figure. CrowdStrike publishes no list price for the agentic tier and meters it with credits. For Falcon-first service providers it is the lowest-friction agentic add.
Strengths: Exceptional investigation quality on Falcon telemetry; credit metering lets you model consumption per client before committing; zero new console.
Limitations: Charlotte is optimized for alerts Falcon produced; cross-stack investigations fall back to your analysts, which, for an MSSP, means the margin leak reopens exactly where client stacks diverge from CrowdStrike. MSSP multi-tenancy for Charlotte remains partial.
Best for: Providers whose offering is explicitly Falcon-managed services. For multi-stack books, pair with a vendor-agnostic layer. → Morpheus for CrowdStrike
8. Microsoft Security Copilot + Sentinel: Best for Azure Lighthouse MSSPs
Architecture: Ecosystem-Native Agent · Autonomy ceiling: AL2 in production (AL3 agents in preview) · Multi-tenancy: Yes, via Azure Lighthouse
For MSSPs managing Microsoft-heavy client bases through Azure Lighthouse, Copilot’s inclusion in E5 puts agentic assistance in reach across a large slice of the client base, and Sentinel’s multi-workspace management is mature. Microsoft’s Product Terms make the inclusion effective January 1, 2026, with phased rollout from November 18, 2025. The allocation is capped: 400 Security Compute Units per month per 1,000 paid licenses, to a ceiling of 10,000 SCUs per month, with no rollover and overage billed per SCU.
Strengths: No incremental license cost for E5 clients up to the included SCU allocation; unmatched Microsoft telemetry depth; the partner-agent store creates extension opportunities for MSSPs that build.
Limitations: Capability is split across GA and preview. The Sentinel MCP server reached general availability in February 2026 and AI playbook generation in May 2026, while several agent-facing features stay in preview, including graph tools in the MCP server, the Agent 365 connector, and UEBA behavior anomalies. Production service SLAs should not depend on the preview half. Today it is assistive (AL2): your analysts stay in the loop for every action, which limits the margin transformation. The moment client stacks include CrowdStrike, Okta, AWS, or non-Microsoft SaaS, coverage thins.
Best for: Microsoft-practice MSSPs who want bundled assistance now and will layer production-grade autonomy separately. → Morpheus for Microsoft
9. Radiant Security: Best High-Coverage Triage Layer on Client Stacks
Architecture: Focused AI Analyst (triage + response) · Autonomy ceiling: AL3 · Multi-tenancy: Limited
Radiant’s promise is attractive for providers drowning in a specific client’s noise. Radiant states that its triage and research agents investigate 100% of alerts, including unknown types, and that roughly 90% of alerts are resolved before they reach the analyst queue, with explainable reasoning attached. Both figures are Radiant’s own claims (vendor product pages, checked August 2026).
Strengths: Unlimited-alert coverage removes the backlog conversation; adaptive triage handles novel alert types without playbook work; readable reasoning output your analysts can pass to clients.
Limitations: Multi-tenancy is limited. Radiant is an enterprise product an MSSP can operate per client, not an MSSP platform. Case management, orchestration, and white-labeling live elsewhere.
Best for: Providers needing a triage pressure valve on specific high-volume clients, not a book-wide operating platform.
Side-by-Side: Agentic SOC Platforms for MSSPs, 2026
Nine platforms compared on the criteria that determine MSSP margin: pricing behavior, tenancy depth, white-labeling, and cross-stack coverage.
| Platform | Pricing model (trap test) | Multi-tenancy depth | White-label | Per-tenant autonomy policy | Cross-stack coverage | Source & Date |
|---|---|---|---|---|---|---|
| D3 Morpheus | Subscription, compute absorbed (passes all four) | Level 3: native, hard isolation | Yes | Yes (AL1–AL4 per tenant) | 800+ self-healing integrations | D3-verified customer-reported, Jul 2026 |
| Conifers CognitiveSOC | Enterprise/quote (test in POV) | Native; 2–4 hr tenant onboarding | Partial | Per-tenant tuning | Runs atop client stacks | CEO statement, MSSP Alert, May 2026 |
| Torq | Base + per-workflow + per-agent compute (Trap 2) | Yes | Partial | Workflow-scoped | Large connector library, analyst-authored | Vendor-stated + press, 2024–2026 |
| Dropzone AI | Per-investigation tiers; quote-based since 2026 (Trap 1) | Yes (MSSP program) | Limited | Limited | 90+ | Vendor pricing page, checked Aug 2026 |
| Stellar Cyber | Platform licensing | Yes, MSSP-oriented | Partial | Platform-scoped | Open XDR + integrations | Vendor-stated, 2026 |
| SentinelOne Purple AI | Tiered platform + add-on | Partial | No | Platform-scoped | S1-first; Athena expanding | Vendor-stated; Athena announced Apr 2025 |
| CrowdStrike Charlotte | Credit-based; no public list price | Partial (Charlotte) | No | Falcon-scoped | ~150, Falcon-first | Vendor-stated; Detection Triage Feb 2025, Agentic SOAR Nov 2025 |
| MS Security Copilot + Sentinel | Included in E5, capped SCU allocation (Jan 2026) | Yes (Azure Lighthouse) | No | Preview-dependent | 300+ Sentinel connectors | Microsoft licensing + preview status, Q1–Q2 2026 |
| Radiant Security | Quote-based, unlimited-alert model | Limited | No | Limited | 100+ | Vendor product pages, checked Aug 2026 |
How to Choose: Four MSSP Scenarios
“Our margin problem is analyst cost per client.” You need the full lifecycle autonomous. Triage alone leaves escalation, response, and reporting labor in place. Shortlist platforms with built-in execution and Level 3 tenancy (D3 Morpheus; Conifers as the mesh alternative), and model the four pricing traps against your three noisiest clients’ real alert volumes.
“We’re a Falcon/S1/Microsoft shop.” Verify the premise across the book, not the anchor client. Service providers rarely stay single-ecosystem for long: every net-new logo is a new stack. Native agents are rational where 80%+ of a client’s alerts originate in that ecosystem; for everything else, the boundary is where your analysts’ hours go. Run the boundary test from the category guide.
“We need something live this quarter.” Dropzone’s MSSP program and Conifers’ hours-scale tenant onboarding are the speed plays. Accept the respective trades: per-investigation economics, or mesh audit composition. If the quarter-speed requirement comes from one noisy client, Radiant as a pressure valve buys evaluation time.
“Our clients are regulated.” The audit artifact is your deliverable now. One incident should yield one replayable decision trail per tenant: evidence, logic, confidence scores. Unified engines produce this natively; meshes require composition; preview-stage agents may not produce defensible artifacts at all. Map against your clients’ NIS2, DORA, and EU AI Act obligations before contract, not at first audit. → NIS2 · DORA
Frequently Asked Questions
What is the best agentic SOC platform for MSSPs in 2026?
D3 Morpheus is the strongest agentic SOC platform for MSSPs in 2026: native Level 3 multi-tenancy with hard data isolation and white-label UI, per-tenant autonomy policies across four autonomy levels, 800+ self-healing integrations for heterogeneous client stacks, and subscription pricing that keeps cost-to-serve flat as client alert volumes fluctuate. Conifers CognitiveSOC is the strongest MSSP-first mesh alternative; Torq suits large providers with automation engineering teams; Dropzone AI is the fastest entry for smaller books.
How should an MSSP price-compare agentic SOC platforms?
Model your three noisiest clients’ real alert volumes, including surge weeks, against each vendor’s pricing structure. The canonical question to put to every vendor, ours included: what does the bill look like in your noisiest month? Per-alert and per-investigation models couple your COGS to client noise; per-workflow and per-agent-compute models couple cost to incident volume; per-GB models couple cost to data; subscription models trade a higher floor for flat cost-to-serve. For a service business, the pricing structure matters more than the list price.
Why does multi-tenancy depth matter more than the multi-tenancy checkbox?
Because “multi-tenant” spans everything from UI filters to hard data isolation with white-labeling and per-tenant policy. When a client’s auditor asks how their data is segregated, or one client requires human approval on every action while another wants bounded-autonomous response, only Level 3 tenancy answers without workarounds.
Can an MSSP white-label an agentic SOC platform?
Some platforms, yes. D3 Morpheus ships white-label UI natively for client-facing views. Most ecosystem-native agents (Purple AI, Charlotte, Copilot) do not white-label, which places the vendor’s brand between you and your client. If your service identity matters commercially, make white-labeling a hard requirement, not a preference.
Do agentic SOC platforms replace MSSP analysts?
The hiring evidence says no. It says the work changes shape. D3 Security’s 2026 analysis of 1,000+ security job postings found service providers remain the industry’s primary entry-level employer while enterprise postings shift toward engineering and oversight. In practice, agentic platforms absorb Tier-1 triage volume while your senior analysts move to validation, hunting, and client-facing work. That is also the margin story: the same headcount serves a larger book at higher-value tiers.
What autonomy level should an MSSP run for clients?
Per client, per policy, which is why per-tenant autonomy governance is an evaluation criterion. A pragmatic pattern: onboard new tenants at AL2 (agent investigates, analyst approves), graduate proven alert categories to AL3, and reserve AL4 bounded-autonomous execution for high-confidence, low-blast-radius actions under command-risk policy. Platforms that run all levels on one engine make that graduation a policy change, not a product change.
Final Thoughts
For an enterprise, choosing an agentic SOC platform is a capability decision. For an MSSP, it is a business-model decision: the platform’s pricing structure becomes your cost structure, its tenancy architecture becomes your compliance posture, and its coverage boundary becomes the line where your margin leaks back into analyst hours. Run every finalist against your own book’s economics (noisiest clients, surge weeks, heterogeneous stacks) and demand the per-tenant audit artifact. Vendor claims are inputs. Your cost-to-serve at renewal is the decision.
Run the Margin Math on Your Own Book
D3 Morpheus gives MSSPs the full alert lifecycle on one engine: up to 95% of alerts investigated at L2+ depth in under two minutes, Level 3 multi-tenancy with white-label UI, 800+ self-healing integrations, and subscription pricing that holds cost-to-serve flat as your book grows.
Request a demo → · Morpheus for MSSPs →
D3 Security is not affiliated with the third-party vendors named above. All trademarks are the property of their respective owners. This comparison reflects publicly available information and our team’s evaluation as of August 14, 2026. Vendor-stated figures are the vendor’s claims, not independent audits.
