DoubleZero’s networking layer is not built for a single ecosystem. It is built for a problem every distributed system shares: the public internet underneath validators was never designed for sub-second consensus across geographic regions.
For Solana, where latency directly impacts vote credits and staking rewards, this matters immediately. But as validators increasingly operate across multiple chains, networking performance is becoming a competitive requirement everywhere, not just on Solana.
Solving the Shared Infrastructure Problem
Theoretically speaking, any distributed system or blockchain could adopt DoubleZero with minimal lift. Or at least that is the goal.
Every distributed system and blockchain faces the same problem. Consensus algorithms improve, validator clients become more efficient, but their data still travels over the same public internet designed decades ago for email and web browsing. The shared networking infrastructure underneath every distributed system and blockchain with the public internet is a bottleneck.
DoubleZero was built to solve the shared networking infrastructure problem. DoubleZero creates an incentive for Network Contributors to connect underutilized, fibre-optic links. It is a dedicated, high-performance networking mesh. It gives distributed systems a faster, more reliable way to move data between nodes. DoubleZero operates at the networking layer, below both the execution and consensus layers. Therefore it can serve any distributed system or blockchain without requiring significant protocol changes, consensus rewrites, or new virtual machine logic.
The Problem: Jitter, Congestion, and a Network Built for Something Else
The public internet routes data through shared infrastructure: commercial ISPs, undersea fibre cables, internet exchange points. Each hop adds unpredictable delays. A packet from a validator in Frankfurt, to one in Tokyo may take a different path every time. Some packets arrive a few milliseconds faster one second and significantly slower the next.
Jitter is the variation in packet arrival times. For everyday browsing, Jitter is invisible and unnoticeable. For a distributed system or blockchain, trying to reach consensus across hundreds of geographically distributed validators, is a structural issue. Even a perfectly optimized validator hardware and software stack, can suffer delays from the networking layer beneath it. Delayed consensus voting, unpredictable transaction/trading execution, missed rewards are all symptoms of the same networking infrastructure problem.
If distributed systems and blockchains want to displace current financial infrastructure, their product offerings must be better than current offerings.
For Solana, this is on-chain NASDAQ. To accomplish this, IBRL, which stands for Increase Bandwidth, Reduce Latency (IBRL) is the rallying cry. The Solana ecosystem coined the term to describe the ongoing effort to push transaction throughput higher and block times lower. As validator clients like Agave, Rakurai, Jito-Solana, and Firedancer improved, the remaining performance ceiling is physical networking, not computational. As Max Resnick, lead economist at Anza, has noted, execution is no longer the bottleneck for Solana. Networking is.
On Solana, we observe measurable improvements tied to network performance, particularly in latency-sensitive environments where even small gains translate into increased vote credits and rewards.
DoubleZero is a meaningful component of our infrastructure stack at the networking layer. Their approach to reducing latency across regions allows Figment to balance geographic distribution with performance, rather than a trade-off.
What DoubleZero Does
DoubleZero is a global mesh network built from dedicated, fibre-optic links contributed by telecom companies, data center operators, and infrastructure providers. These organizations already own or lease fibre that sits underused. Dark fibre, as the industry calls it, is cable that has been laid but carries no traffic. DoubleZero gives these organizations a mechanism to pool that spare capacity into a single high-performance transport layer.
Today, DoubleZero carries over 9.52 terabits per second in aggregate capacity, with 448 validators connected, representing 47.6% of Solana’s total stake and over $17 billion in connected value. The network currently spans North America, Europe, and Asia – over 30+ metros globally.
| Aggregate Network Capacity | 9.52 Tbps |
| Connected Validators | 448 |
| Solana Stake on DoubleZero | 47.6% |
| Total Connected Value | $17B USD |
Table: DoubleZero network snapshot (source: doublezero.xyz/dashboard, updated hourly)
Why Any Chain Can Use It
Blockchains are built in layers: applications at the top, then execution (how transactions are processed), then consensus (how validators agree on what happened), and at the bottom, the network layer, which is the physical and logical infrastructure that moves data between machines. DoubleZero lives at that bottom layer.
It does not touch execution logic. It does not modify consensus rules. It does not care whether a chain uses parallel execution (like Solana or Sui) or sequential execution (like Ethereum), Proof-of-Stake, Proof-of-History, or a DAG-based consensus model. It moves data between nodes faster and more reliably than the public internet. That is the full scope of what it does, and that scope is precisely what makes it chain-agnostic.
A bridge protocol or cross-chain messaging layer has to understand the specifics of each chain it connects. DoubleZero does not. The performance improvements it delivers are useful to any distributed system where nodes need to communicate quickly and consistently. Whether that system is Solana, Sui, Aptos, Celestia, Avalanche, or a high-performance Layer 2, the benefits are identical: data arrives faster, with less variance, and with less junk on the wire.
No protocol redesign required. Validators install lightweight daemon software that creates tunnel interfaces and manages routing. If the DoubleZero network is unavailable for any reason, traffic falls back to the public internet automatically. No hard fork. No consensus upgrade. DoubleZero is additive by design.
DoubleZero is chain-agnostic by design, built for any distributed system. Blockchains like Solana were the first natural product market fit, with Solana user’s strong demand for latency sensitive trading and demand for blockspace. DoubleZero’s chain-agnostic model reflects broader industry evolutions. Validator service providers no longer operate in single-network silos, multi-chain participation is the norm. DoubleZero is built to be multi-tenancy by way of carrying traffic of any blockchain and growing its ecosystem of tenants widely. Currently the network has two tenants: Solana and Shelby, but intends to expand into other blockchain ecosystems over time.
What Chains Actually Gain
Lower latency. Across dozens of measured routes, DoubleZero consistently delivers faster round-trip times than the public internet. On some intercontinental routes, such as Oslo to Frankfurt or Strasbourg to Munich, the improvement reaches 40 to 50 percent. Even on well-served routes, DoubleZero matches or beats standard internet paths by 15 to 25 percent. For Solana, targeting sub-400ms block production, shaving 10 to 50 milliseconds off intercontinental links is the difference between hitting the window and missing it.
More predictable block propagation. Consistency matters as much as raw speed. A validator that usually receives blocks in 20 milliseconds but occasionally waits 200 milliseconds is worse off than one that consistently receives them in 40 milliseconds. DoubleZero’s dedicated links carry committed service-level agreements (SLAs), contractual performance guarantees that are continuously monitored and enforced. Links that fall below agreed standards can be penalized or removed. Block data, consensus votes, and transaction data all arrive within a tight, predictable window rather than subject to the routing instability of the open internet.
Reduced performance variance across validators. On the public internet, geography and ISP quality create persistent structural advantages for validators in well-connected locations. DoubleZero compresses that gap by giving the full validator set access to the same high-performance transport layer. The results from DoubleZero’s vote metrics are direct: validators joining the network see average skipped slot rates fall from 0.29% to 0.05%, and average earned credits rise from 99.62% to 99.93%. These gains come from better networking, not better hardware or software.
| Vote Metrics (median) | Before DoubleZero | With DoubleZero | Change |
| Average Skipped Slots | 0.29% | 0.05% | -0.24 pp |
| Average Earned Credits | 99.62% | 99.93% | +0.31 pp |
| Average Vote Latency | 1.04 slot | 1.04 slot | -0.6% |
These are not hypothetical projections. Figment’s own Figment-1 validator on Solana reflects these improvements directly, in earned vote credits, and reduced skipped slots.
Table: Validator performance before and after joining DoubleZero (source: doublezero.xyz/metrics/vote, data from app.vx.tools/.tools)
Validator Vote Metrics Performance Dashboard
Source: DoubleZero vote metrics
DoubleZero Round Trip Network Latency vs Public Internet
Source: DoubleZero Telemetry Metrics
DoubleZero vs Public Internet Latency
Source: DoubleZero Telemetry Metrics
Compressing validator performance variance matters for network health. When better-connected operators consistently outperform peers through no merit of their own, rewards concentrate, decentralization weakens, and the network underperforms its theoretical throughput. DoubleZero addresses that at the infrastructure level, before any chain-specific logic comes into play.
DoubleZero is Chain-Agnostic
DoubleZero is chain-agnostic because the problem it solves is chain-agnostic. Every distributed system and blockchain sends data over the internet. Every distributed system or blockchain suffers when that data arrives late, inconsistently, or loaded with spam. Every distributed system or blockchain benefits, when the network underneath it gets faster, more predictable, and more reliable.
DoubleZero is a network-layer innovation. It increases bandwidth and reduces latency for whatever data flows over it. That is why Solana validators like Figment already see measurably better performance after connecting. The same infrastructure can extend to Sui, Aptos, Celestia, Avalanche, and whatever high-performance distributed system comes next.
The public internet was a marvel of engineering for its era. The era of globally distributed, sub-second consensus demands something purpose-built. DoubleZero is that something, and the reason it can serve every chain is that it was designed from the beginning to serve none of them specifically. It serves the networking layer they all share.
Sources:
Anza, “2025 Mission: On the Road to 1M TPS Brennan Watt
Max Resnick: Execution is not the bottleneck
DoubleZero network dashboard, doublezero.xyz/dashboard, updated hourly.
Malbec Labs Documentation, “Architecture,” docs.malbeclabs.com/architecture/.
DoubleZero telemetry dashboard, doublezero.xyz/metrics/telemetry,
DoubleZero vote metrics, doublezero.xyz/metrics/vote, data from app.vx.tools.
