Systems in production

Real work.
Real production.

These aren't demos or prototypes. Every system on this page runs in production, managed by X402. We designed it, built it, and we're the ones who get the call when something breaks at 3 a.m.

4,073+ industrial machines
instrumented
35,000+ bank customers on
rails we built
500+ autonomous drones
in the fleet
6,000+ endpoints on a
private zero-trust mesh

4,073 machines. 21 factories. 3 countries. Pilot live in 40 days.

A global manufacturer ran 4,000+ production machines on clipboards and guesswork across 21 factories in 3 countries. We wired every one to speak — and built the system that now runs the entire operation.

Before

Production was counted by hand. A manager learned a machine had stopped hours after it happened — and rarely knew why. The numbers that ran the business lived in spreadsheets, a full day behind the factory floor — across 3 countries.

What we built

We designed and built the hardware that taps each machine controller directly — opto-isolated signal conditioning laid out by electrical engineers on our own payroll — feeding PoE-powered single-board agents, each firewalled and remotely manageable to the point of hard power-cycling any individual unit from anywhere in the world. All 21 factories connect over an encrypted open-source mesh network. The platform ingests 80,000+ live signals fleet-wide, computes real-time OEE and fault detection per machine, and integrates two-way with the ERP — consolidating six fragmented business data sources into one custom operating environment X402 built and runs for the client.

The result

4,073+ machines now instrumented across 21 factories in 3 countries. Pilot live in 40 days — hardware R&D, in-house firmware, custom electronics, and first-factory commissioning. Full fleet rolled out over the following 10 weeks (~2 factories/week). A second customer deployment is already underway: 2,000 machines at a second client site.

4,073+machines instrumented
21factories connected
3countries
40 daysto pilot live
80,000+signals per day
Time to production: X402 vs. traditional integrator
Traditional
12–18 months
X402
40 days ✓

Actual deployment: 4,073 machines, 21 factories, 3 countries. Pilot: 40 days. Full fleet: ~4 months.

custom firmware opto-isolated electronics encrypted mesh PoE fleet real-time OEE ERP integration 21 factories · 3 countries
Read the full case study Talk to us about your factory
Data pipeline — sensor to operations 80,412 signals today · fleet-wide
X402 designed every layer — custom electronics · firmware · mesh · platform · operations

X402 electrical engineers designed the opto-isolated signal conditioning circuitry and PoE hardware — we own the full stack from the factory floor circuit board to the operations dashboard, with no dependency on a third-party IoT platform.

Three banks. One rail. Stablecoins in and out of bank accounts.

We believe global financial settlement migrates to blockchain within the next 2–5 years. We're not waiting — we already build the rails banks run on today.

The need

An American fintech needed to integrate with a major settlement blockchain so banks could offer their customers crypto on- and off-ramp of USDC and USDT. Three banks signed up for the service.

What we built

As part of the core engineering team on a major fintech build, we co-built a payment-rails platform — open source under the hood, built to order — that creates a dedicated on-chain address for every end user. Customers send USDC to and from it; funds are credited straight to their bank account, and from there move on-chain or over legacy rails — their choice.

The scale

35,000+ end users across the three banks — ~10,000, ~20,000 and ~5,000 each — with the platform handling ~1,400 stablecoin transfers a day, roughly 42,000 a month, half a million a year.

3banks live
35,000+end users
~1,400transfers / day
~500Ktransfers / year
stablecoin rails open-source core on-chain addressing legacy rail bridge fintech co-build
Read the full case study → Building fintech rails? Let’s talk

500+ drones. No pilots. The whole farm, watched from the air.

A large farm needed to see land no crew could walk. We built the autonomous flight agents that fly it for them — and the hardware they fly on.

The need

A large row-crop operation had more land than any crew could survey on foot. Crop health, problem spots and change over time went unseen until they showed up at harvest.

What we built

Autonomous flight agents on 500+ drones that plan and fly their own RTK-GPS-tracked routes over every field, capture high-resolution imagery on schedule, and complete each mission — takeoff to landing — with no pilot on the sticks.

The result

The farm now sees every acre on a cadence, catches trouble early from the air, and holds a dated, geo-tagged photographic record of its land over time — flown autonomously, start to finish.

500+drones in fleet
0pilots required
RTKGPS precision
autonomous agents RTK-GPS navigation boustrophedon sweep high-res imaging fleet management
Read the full case study

Enterprise campus, thousands of users: redundant everything, down to the door locks.

Before “smart building” was a marketing term, we built one. Access control, network and billing wired into a single system — the building enforced who got in, who got connectivity, and who didn’t.

The job

Servers, firewalls, switches, access points — redundant everything. A captive portal wired into the ERP: paying members got network access, late accounts didn't — billing enforced by the network itself. Access control by RFID and biometrics.

The doctrine

The best hardware, tied to none. We've run hands-on every class of firewall, network OS and desktop or server platform in the trade — so we pick the best for each case and combine technologies where one beats another. Redundancy isn't a feature, it's the default.

Buildings that run themselves

The same competency extends to enterprise automation: lighting, climate, sprinklers and camera surveillance integrated to real actions — open-source automation cores, hardened to enterprise grade. The building answered for itself.

6,000+ endpoints on a private mesh. The network defends itself.

Open source and closed source are not opposing camps — they're parts. We combine them where each is strongest.

What we built

A private network fabric — a zero-trust mesh built to enterprise standard. Multi-tenant: multiple customers, each carved into a catalog of ~1,000 isolated network segments, 6,000+ endpoints connected in total. We built the platform, and we operate it.

Three datacenters, active-active

The mesh spans three production datacenters on two continents plus a dedicated DR site on a third. Gateways run active-active between the two primary sites — if one goes dark, the other carries every tunnel immediately. Encryption at rest on every server: LUKS on Linux, full-disk on Windows.

It defends itself

AI agents inspect every flow across the mesh in real time — flagging lateral movement, detecting exfiltration, catching anomalies before they spread. When the egress agent detects a compromised endpoint mid-session, it isolates it and pages the sysadmin immediately, before the breach can fan out. While the client's team sleeps, the network defends itself.

6,000+endpoints
~1,000network segments
3datacenters
+1DR site, separate continent
mesh gateway open-source virtualization LUKS encryption active-active replication BYOIP egress AI security agents
Read the full case study

14 federal APIs. One compliance engine. Zero audit failures.

Brazilian fiscal compliance isn’t a form you fill out once — it’s a continuous integration problem. We built the engine that runs it automatically.

The compliance engine

A mid-market distributor was generating hundreds of NF-e invoices per day, running monthly eSocial filings, reconciling Open Finance statements, and bracing for quarterly SPED audits — each in a separate system, each with its own team. X402 built a real-time compliance engine that connects SEFAZ, eSocial, Open Finance, SPED, and 10+ other government APIs into a single automated layer.

Authorization in under a second

NF-e is authorized via real-time SEFAZ API — the receipt arrives in under one second. eSocial events are generated automatically when the HR system fires a trigger. The SPED ledger is always current, never the product of a quarterly scramble. Zero dedicated compliance headcount for routine filings since go-live.

LGPD-native from day one

Every data entity is classified at creation. LGPD retention schedules are enforced at the database schema level — not by a process or a reminder. Consent records are first-class objects. Data subject requests are served by API. The DPO gets a dashboard, not a spreadsheet, come audit time.

14+federal APIs
0audit failures
<1 sNF-e authorization
2 FTEheadcount eliminated
SEFAZ / NF-e v4.0 eSocial SPED Open Finance Pix LGPD digital certificates

Decades running other companies' critical operations.

Long before AI agents, we were the ones whose phone rang when a switch died at 2 a.m. Physical operations are systems too.

Over thirteen years operating mission-critical infrastructure for companies that can't afford downtime — with the same discipline we bring to every system we build today.

The VC hub

For over a decade, we managed the entire infrastructure of Brazil's leading startup and venture-capital innovation hub: three buildings in São Paulo, ~1,000 active users a day — redundant firewalls, switches, PoE access points, cameras and access control, end to end.

The education campus

From 2019 to 2022 we built and ran both São Paulo floors of a Silicon Valley online-education company: 100+ people a day, everything under our management — ERP, access control, cameras, alarm system — with fiber backbones joining the two offices into one network.

The financial group

For 13 years we ran the US operations of a South Korean financial group — private-equity, VC and trading offices from New York to San Francisco: three floors, 50+ professionals, 100+ managed devices — voice, network, firewalls, file sharing, email.

Ready to be next?

Your operation, instrumented.

Every system on this page started with a conversation. Tell us what's breaking or what you want to see — we'll give you a straight answer within a day.

No sales team. The engineer who scopes your project builds it.