Mitigate risk
Attacks start in the inbox and spread to the endpoint. We watch both with one AI that judges behavior, not a list of names.
Investor relations
Security that fixes itself. A company built to own the whole stack.
CloudCastle is shipping three products in sequence — Email Security, AI-native antivirus, and an autonomous AI technician — and training its own security model on partner compute. Every release cuts risk, cuts downtime, and keeps expensive attacks from ever landing.
3
Products, one platform
2028
Autonomous technician GA
$1M+
Partner compute we are applying for
100%
US-built and staffed
Why now
Three costs every business pays. We remove all three.
Minimize downtime
Downtime is the bill nobody budgets for. Our AI technician fixes routine problems before anyone files a ticket.
Stop expensive attacks
One ransomware event can cost more than a decade of protection. We are priced so that never happens to a customer.
Legacy tools recognize only threats someone already named and still need a human to clean up; CloudCastle judges behavior, acts on its own, and records what it changed.
The platform
One engine, three products stacked on it
Every product sits on the same CloudCastle security model, so email, endpoint, and IT operations all judge behavior with one brain. Each layer we ship makes the layers above it cheaper to build and better at their job.
Product roadmap
One release at a time, each one funds the next
Quarterly release cadence. Every stage ships to paying customers, gets a full test cycle, and is proven before the next one starts.
Q4 2026
Shipping nowCloudCastle Email Security — general availability
Two ways to protect the inbox, sold together or apart: a secure email gateway that sits in front of the mail server (the Proofpoint model), and API-based analysis that inspects mail already inside Microsoft 365 and Google Workspace (the Perception Point model). Built for small and mid-size businesses and for enterprise tenants on the same engine.
Q4 2026
FundraisingSeed round closes
Funds Email Security go-to-market and the compute-partner applications below.
Q1 2027
ComputeCompute partners locked in
Applications filed with every hyperscaler and model-maker program listed on this page. Credits fund the first training runs of the CloudCastle security model.
Q2 2027
Next releaseCloudCastle Advanced EDR — AI-native next-generation antivirus (NGAV)
Endpoint protection with the CloudCastle model built in. Judges behavior on the device to stop zero-day exploits and hands-on-keyboard attacks that have no signature yet, and explains every action it takes.
Q3 2027
AIModel v1 trained on partner compute
Our own model, trained on email, server, and endpoint telemetry from live customers, replaces the general-purpose models we use today. Goal: beat the incumbents on detection rate and time to fix.
Q4 2027
FundraisingSeries A closes
Scales sales and the training pipeline; funds the autonomous technician.
Q1 2028
PreviewCloudCastle AI Technician — early access
An autonomous agent that fixes non-security problems on endpoints the moment they appear: slow machines, broken updates, printer and driver faults, disk and login issues. Small businesses get an IT department without hiring one.
Q3 2028
PlannedCloudCastle AI Technician — general availability
Full autonomy with a recorded audit trail for every change. Positions CloudCastle as the single platform for security and IT operations.
Q4 2028
FundraisingSeries B
Growth capital for enterprise and MSP channel expansion and model v2.
Funding plan
Three rounds, each tied to a shipped product
Targets, not commitments. Each round opens only after the milestone before it is live with customers.
Seed
$3M target
Q4 2026
- Email Security GA and first 500 protected mailboxes-per-tenant deals
- Compute-partner applications
- 12 months of runway to the Series A
Series A
$15M target
Q4 2027
- Advanced EDR AI-native antivirus in market
- Model v1 trained and in production
- Sales team for SMB, enterprise, and MSP channels
Series B
$40M target
Q4 2028
- AI Technician GA
- Enterprise and international expansion
- Model v2 on dedicated training capacity
Investor cadence
- Monthly investor update email
- Quarterly product release and roadmap review
- Annual strategy briefing
Round sizes, dates, and outcomes are targets and forward-looking statements. They can change as the business and the market change.
Partner compute
Training our own model without spending our own capital
Hyperscalers and model makers give qualifying AI startups large compute grants. We meet the requirements for each program below and are applying to all of them, then stacking the credits to train the CloudCastle security model.
Amazon Web Services — Activate and Generative AI Accelerator
Up to $1,000,000
Activate Portfolio credits up to $100K; AI tier up to $300K; the Generative AI Accelerator cohort up to $1M.
Business domain, live product, institutional backing for Portfolio; accelerator application each summer.
Google Cloud — Google for Startups Cloud Program (AI-first)
Up to $350,000
$250K in year one, $100K in year two, plus $12K enhanced support.
Pre-seed to Series A funding from an institutional investor, founded within five years, AI-first product.
Microsoft — Microsoft for Startups Founders Hub
Up to $150,000
Azure credits unlocked by milestone, covering GPU compute and Azure OpenAI; up to $200K through the Microsoft Investor Network.
B2B, proprietary software, business verification, sustained Azure usage across 10+ services.
Oracle Cloud Infrastructure — Oracle for Startups
70% discount
Deep OCI discount plus always-free tier for GPU and bare-metal training.
New OCI customer with a validated product.
IBM Cloud — Startup with IBM
Up to $120,000
$10K per month for 12 months at the premium level, usable on watsonx and GPU infrastructure.
Under five years old, under $1M revenue, venture or accelerator referral.
NVIDIA — Inception
Partner access
Preferred pricing and partner credits across AWS, Azure, and Google Cloud, plus DGX Cloud access and training.
AI-focused company; free to join, no equity.
OpenAI — OpenAI for Startups
Up to $2,000,000
API credits from $2.5K standard to seven-figure cohort grants for startups training with OpenAI models.
Referral through a partner VC or accelerator for the larger tiers.
Anthropic — Claude for Startups
Up to $100,000
$5K standard, $25K partner tier, premium tier at Anthropic's discretion.
Incorporated company; partner referral for the upper tiers.
xAI — Grok API
Bundled credits
Grok API credits offered through accelerator and cloud partner perk programs.
Access through partner programs; no standalone startup program yet.
CoreWeave — Startup Accelerator
Up to $250,000
GPU cluster credits and discounts sized per company for large training runs.
AI startup with a training workload; application reviewed per company.
Together AI — Startup Accelerator
Up to $50,000
Credits for fine-tuning, inference, and clusters across Build, Scale, and Grow tiers.
Tiered by funding stage.
Cohere — Cohere for Startups
Up to $100,000
Discounted pricing and larger credit packages through VC-partnered tiers.
Venture-backed for the credit tiers.
Program terms as published September 2026; each provider sets final amounts and eligibility. Credits are valued at list price and usually expire within 12 months, so we schedule training runs to match.
Stacked total
$4M+
Combined ceiling across the programs above
We will not receive every ceiling. Winning a third of it funds model v1 and v2 without touching investor capital.
Back the platform that makes cyberattacks and IT tickets disappear
Talk to the founders, see the numbers, and watch the product work on a live endpoint.