About AlgoX2
From NYSE to Streaming OS
AlgoX2 was founded by the team that built NYSE Pillar — the trading architecture that today powers $100B+ in daily transactions across 7 regulated exchanges. While building exchange infrastructure, they discovered that no streaming platform could meet the requirements. AlgoX2 is the general-purpose streaming OS they wished existed.
Why AlgoX2
Today's stack vs. AlgoX2
Today's stack
AlgoX2
Rebalance storms stall writes
A single slow consumer — or one added broker — triggers multi-second rebalances that pause production across the cluster.
Partition-free user API
No Kafka-style topic partitions to rebalance. Producers publish to a single stream; X2 handles ordering internally.
Storage coupled to brokers
Scaling disk means adding CPU you do not need — and vice versa. Capacity planning becomes a guessing game.
Compute-storage separation
Scale CPU and I/O independently with policy-driven tiering across NVMe and object storage.
Five systems duct-taped together
Kafka for transport, Flink for transforms, Connect for sinks, Schema Registry for types, Redis for state. Five things to deploy, monitor, and page on.
One unified data plane
App-app, app-storage, and storage-app handled natively in one engine. No connectors, no schema registry, no external state store.
Duplicate pipelines for analytics
Streaming goes one place, analytics queries go another. Data is copied, freshness lags, and infra bills double.
Streaming + analytics in one engine
Kafka-compatible delivery with inline transforms and enrichment. Query the same data you stream.
JVM pauses and kernel copies
Garbage collection spikes tail latency; every hop copies buffers through the kernel. Microsecond tails are out of reach.
Zero-copy, kernel-bypass runtime
NUMA-aware networking with no kernel transitions. Predictable microsecond tails at extreme throughput.
N×M fan-out tax
Delivering one message to N consumers costs N× egress and N× broker CPU. Scaling subscribers scales your bill linearly.
Native multicast fan-out
One write serves thousands of consumers, with no duplicate network or CPU work.
Team
Meet AlgoX2 Founders

Alexei Lebedev, CEO
Built the world's fastest matching engine (Algo M2, 16µs). Architected NYSE Pillar. Created the OpenACR formal programming paradigm. Founded AlgoEngineering, a cross-continent HFT firm he ran for 10 years.

Vladimir Parizhsky, CTO
IEEE Fellow. 50+ patents. Drove modem technology to the Shannon limit at USRobotics (acquired by 3Com). Chief Systems Architect at Flarion (acquired by Qualcomm). Co-architected NYSE Pillar.

George Levin, CBO
Founded Getintent, a real-time bidding platform (acquired by GPM Media). Founded Hints, an AI sales automation agent. Leads go-to-market, partnerships, and fundraising.
Investors
Seed: $3.5M led by Bessemer Venture Partners
“Foundational infrastructure for the next wave of real-time innovation” — purpose-built for sectors like AI and finance where speed is critical.
Get Started
The foundation for your next real-time system
Whether you're building trading infrastructure, unifying real-time and analytical pipelines, streaming to your lakehouse, or replacing Kafka — this is where you start.