Parameter (formerly Hex Security)
AI agents that continuously hack a company's own systems to find security holes before attackers do.
On the record
Checkpoint: Whether Parameter announces a Series A and at what size versus Neo's $100M and Terra's $38M
The call
Parameter has early pilots and a three-person team from PlayAI, AWS, and Codegen backgrounds, but faces better-funded rivals and platform buyers in a category still deciding who owns it.
Parameter builds software that uses AI agents to find security holes in companies' systems by pretending to be hackers, similar to how companies hire pen testers today but faster and cheaper. The hard part is that bigger security companies like Pentera and Horizon3 already do automated testing without the AI agents, and they have more customers and money, so Parameter's AI approach is not automatically better or harder to copy. The biggest risk is that OpenAI, Anthropic, or other AI labs could build this capability into their own products, making Parameter unnecessary, unless Parameter builds up a library of real attacks and security certifications that prove it works better than what AI labs alone could do.
Why
01
AI pressure: Frontier labs like Anthropic and OpenAI building agentic offensive-security reasoning directly into their models. If frontier models learn to run exploit chains natively, Parameter's agent harness becomes a thin wrapper anyone can copy. Could AI do the core job? Yes, likely.
02
Bigger players: Zscaler (after acquiring Splx AI) and automated platforms like Pentera, Horizon3.ai NodeZero. Zscaler already bought an AI red-teaming startup and folded it into its platform, showing big security vendors will buy this category instead of losing to it. Could a giant copy it? Yes, easily.
03
What protects it. Pricing far below the roughly $30,000 cost of a manual pentest and fast setup for small teams that skip annual audits entirely.
What would change this call
Parameter announces a Series A within 6 months and its size versus Neo/Terra/Beacon. Shows whether Parameter can match peer capital or falls behind in the funding race.
A second incumbent (after Zscaler/Splx) acquires an AI-agent-pentest startup. Would confirm platform players are consolidating the category faster than Parameter can differentiate.
Parameter names a paying enterprise customer with public renewal data, outside the YC network. Would show real distribution beyond friendly-founder pilots.
A second incumbent (after Zscaler/Splx) acquires an AI-agent-pentest startup. Would confirm platform players are consolidating the category faster than Parameter can differentiate.
Parameter names a paying enterprise customer with public renewal data, outside the YC network. Would show real distribution beyond friendly-founder pilots.
Key numbers
- Neo launched with $100M in funding
- Terra Security raised $8M seed, ~$38M total
- Beacon Security raised $13M seed
Named risks
- No confirmed funding round, ARR, or enterprise customer logos beyond vendor-claimed figures.
- Crowded, fast-consolidating market with far better-funded rivals like Neo ($100M) and Terra (~$38M).
- No proprietary exploit dataset or compliance seal; core product is a thin harness over frontier models.
- Brand confusion from rebrand plus namesake collisions with unrelated Hex, Hex Security UK, and parameter.io.
- Frontier-model alignment guardrails may refuse offensive workflows, limiting category-wide capability.
Rivals, as read in July 2026
| Rival | Direction | Evidence |
|---|---|---|
| Neo | losing to | Launched with $100M funding July 20 2026 versus Parameter's unconfirmed capital base |
| Zscaler | losing to | Acquired Splx AI, showing platform-player consolidation appetite in AI red-teaming |
| Pentera / Horizon3.ai | losing to | Established automated pentest platforms with enterprise distribution, not blocked by billable-hours model |
| Terra Security | contesting | Reported ~$38M total funding, launched TORCH product March 2026, closest agentic-pentest peer |
| Beacon Security | contesting | Raised $13M seed July 16 2026 in same AI-agent-security niche |
| Fabraix | contesting | YC-backed peer claims 78% attack success rate on AgentHarm benchmark and 10,000+ jailbreak library |
Questions people ask
- Is Parameter (formerly Hex Security) profitable?
- No profitability data is public. A secondary blog claims Hex Security hit $1M ARR in 8 weeks (vendor-claimed, reported June 3, 2026), but this is unverified and no independent financial reporting confirms revenue or profitability.
- How much funding has Parameter raised?
- No confirmed funding round has been publicly announced as of July 2026. An aggregator figure of $172.0M is excluded here as unreliable; the company's only verified backing is participation in YC's Winter 2026 batch.
- Who are Parameter's biggest competitors?
- Direct AI-pentest peers include Terra Security (reported ~$38M total funding, TORCH launch March 2026), XBOW, Fabraix, and Beacon Security ($13M seed, July 16 2026). Platform incumbents Pentera, Horizon3.ai/NodeZero, and Zscaler (which acquired Splx AI) also compete for the same budget.
- Could AI replace Parameter's own offering?
- Yes, that is the core risk: Parameter's harness sits directly on frontier models like Anthropic's and OpenAI's with no confirmed proprietary exploit dataset or certification, so improved base-model agentic reasoning could subsume its differentiation.
- Who owns Parameter (formerly Hex Security)?
- It is privately held and founder-led, part of Y Combinator's Winter 2026 batch. Founders are Huzaifa Ahmad (ex-PlayAI, AWS, Capital One), Ahmad Khan, and Prama Yudhistira (ex-Codegen, AMD). No outside lead investor has been confirmed beyond YC.
- Did Hex Security rebrand to Parameter?
- Yes. The company launched via YC in February 2026 as Hex Security and by July 2026 its live site, parameter.ai, showed the new name Parameter, though the exact rebrand date has not been confirmed by any dated third-party source.
Timeline
- 2026-07-20Competitor Neo launches with $100M funding in AI-native security
- 2026-07-17Pentera Alternatives roundup names automated pentest platforms as category threat
- 2026-07-16Competitor Beacon Security raises $13M seed funding
- 2026-07Company site shows rebrand from Hex Security to Parameter (parameter.ai)
- 2026-06-03Secondary blog claims Hex Security hit $1M ARR in 8 weeks (vendor-claimed, unconfirmed)
- 2026-05-21Research flags frontier-model alignment guardrails limiting AI red-teaming agents category-wide
- 2026-03-23NIST CAISI publishes large-scale AI agent red-teaming competition analysis
- 2026-02-16tryfondo blog covers Hex Security launch, cites Anthropic's Sept 2025 AI-attack disclosure
- 2026-02-08Hex Security publicly launches via Y Combinator, claims vulnerabilities found in dozens of YC firms
- 2025-04-22Peer Terra Security secures $8M seed for agentic AI pen testing
Revision history
27 July 2026
First assessed. Losing its protection. The text above is fixed as of this date and is not edited afterwards.
In the same band
Sources
- Hex Security: Agentic Offensive Security at Scale
- Launch YC: Hex Security: AI that Hacks before Attackers do (2026-02-08)
- Hex Security Funding: $172.0M
- Claude's Corner: Hex Security, AI Agents That Hack Before Attackers Do (2026-06-03)
- $8 Million (Seed) Secured For Agentic AI Pen Testing Solution (2025-04-22)
- Beacon Security Raises $13 Million in Seed Funding... (2026-07-16)
- Neo Launches with $100M to Secure AI Software Across the Enterprise (2026-07-20)
- Best AI agent red teaming tools in 2026
- Launch YC: Fabraix
- AI red teaming agents change how LLMs get tested (2026-05-21)
- Insights into AI Agent Security from a Large-Scale Red-Teaming Competition (2026-03-23)
- Hex Security Launches: AI That Hacks Before Attackers Do (2026-02-16)