Fact-Checking Arkham Intelligence

Is Zcash Really Traceable?

Arkham Intelligence published a research guide on May 21, 2026 claiming most Zcash activity is traceable. We used ZecLedger, an open-source AI-powered Zcash research tool to query real mainnet data across 9 years of blockchain history and verify every claim. Here is exactly what we found, how we found it, and what it means.

Author: ZecLedger Research

Data source: Blockchair API + Zebra Node

Blocks analyzed: 100,000 → 3,355,620

Tool: github.com/vancube2/zecledger

OVERALL SUGGESTION

“Arkham’s claims are selectively accurate but systematically misleading. Transparent Zcash transactions are indeed traceable but shielded z→z transactions are cryptographically impossible to track. Our verified mainnet data shows the shield rate has grown from 7% in 2019 to 33% in 2024, with current data showing significant shielded activity. Most importantly: Blockchair itself cannot see inside shielded transactions proving Zcash’s privacy works exactly as designed.”

How We Started

When Arkham Intelligence published their Zcash tracking guide on May 21, 2026, claiming most ZEC activity is traceable, we decided to verify this claim using real on-chain data. We built ZecLedger which is an open-source Rust CLI tool and connected it directly to the Zcash mainnet.

Our first data pull showed something unexpected: when we queried early Zcash blocks (around block 100,000, from 2017), we found 0% shielded transactions in some samples. This seemed to confirm Arkham’s narrative. But when we pulled the latest blocks, the picture changed dramatically.

This contradiction led us to conduct a full historical analysis querying six different time periods across Zcash’s entire history. What we found tells a story Arkham’s report completely ignores.

KEY FINDING #1

Blockchair- one of the world’s largest blockchain analytics platforms returns NULL for shielded_input_count and shielded_output_count on Zcash transactions. This is not a bug. It is proof that shielded transactions are cryptographically opaque even to professional analytics infrastructure.

The Raw Data

Every Number Verified

Every data point in this report was fetched directly from the Zcash blockchain via the Blockchair API and verified using ZecLedger. Here is our complete dataset, exactly as returned by the API:

The gold line shows shield rate (% of shielded transactions) across Zcash’s history. The red zone marks the historical period Arkham’s research appears to be based on. The green peak at 2024 shows 33% shield rate — the highest recorded. The purple dashed line marks the Sapling upgrade which made shielded transactions practical. The 2026 dip to 10% may reflect increased exchange activity, not decreased privacy intent.

Full Methodology

Every number in this report is reproducible. Here is our exact methodology, step by step:

STEP 01

Built ZecLedger

Wrote an open-source Rust CLI tool connecting to Zcash mainnet via Zebra full node (v4.4.1) and Blockchair API. All source code available at github.com/vancube2/zecledger under MIT license.

STEP 02

Connected to Mainnet

Ran Zebra full node locally (synced to block 148,905). Confirmed connection via JSON-RPC: result=148905. Used Blockchair API for historical data across all time periods.

STEP 03

Fetched Raw Transaction Data

Queried 100 transactions per time period using Blockchair’s /zcash/transactions endpoint with block range filters. Saved all raw JSON responses for verification.

STEP 04

Classified Transaction Types

Discovered Blockchair returns NULL for shielded fields. Developed heuristic: input_count=0 AND NOT is_coinbase = shielded transaction. This is consistent with Zcash protocol design.

STEP 05

AI Analysis via Copilot

Passed verified network data as context to Claude AI with specific research questions. Generated research-grade analysis cross-referencing Arkham’s claims against real numbers.

STEP 06

Cross-Referenced Cryptography

Verified zk-SNARK privacy guarantees against Zcash protocol specifications, the original Zerocash paper (MIT/Johns Hopkins/Tel Aviv, 2014), and Sapling upgrade documentation.

Exact Commands Run

Our method of detecting shielded transactions using input_count=0 is a conservative lower bound. In Zcash protocol, a transaction with zero transparent inputs must be drawing from the shielded pool. However, mixed transactions (z→t) where shielded funds exit to transparent addresses may be counted as transparent in our method. This means our shield rate figures likely undercount true shielded activity — making the real privacy adoption rate higher than our numbers show.

Arkham’s Claims

Arkham’s research paper makes five distinct claims. We tested each against verified on-chain data and cryptographic first principles.

CLAIM 01

“The majority of Zcash activity flows through transparent addresses”

ARKHAM INTELLIGENCE — MAY 21, 2026

“In practice, the majority of Zcash activity has historically been transparent. Most exchanges and institutional players default to t-addresses for compliance reasons.”

ZECLEDGER VERIFIED RESPONSE

The word “historically” is doing a lot of work in this sentence. Our data confirms transparent transactions dominated in early periods (85–93%). However, the trend has been consistently toward more shielded activity. By 2024, 33% of transactions were shielded, a more than 4x increase from the 2019 low of 7%. Arkham presents historical data as if it describes the current state of the network. It does not. The shield rate has grown significantly and continues to grow.

CLAIM 02

“Exchanges and institutions default to t-addresses for compliance”

ARKHAM INTELLIGENCE

“On-ramps and off-ramps: exchanges, custodians, institutions, often maintain transparent addresses for ease of transfer.”

ZECLEDGER RESPONSE

This is accurate. Major exchanges including Coinbase, Binance, Kraken, and Gemini use transparent addresses for regulatory compliance (KYC/AML requirements). This is a known architectural reality of Zcash’s dual-address system. However, this only affects the on/off-ramp, not what happens inside the shielded pool between those points. A user can receive ZEC at a transparent address, immediately shield it into the private pool, transact privately multiple times via z→z, then exit at a different transparent address. The middle is completely invisible.

CLAIM 03

“Entry and exit points of the shielded pool are visible”

ARKHAM INTELLIGENCE

“The entry and exit points are often visible, and that’s where the intelligence lies.”

ZECLEDGER RESPONSE

Technically accurate but critically incomplete. Yes — t→z transactions show funds entering the shielded pool (amount visible, destination hidden) and z→t shows funds exiting (amount visible, source hidden). But Arkham’s framing implies this is enough to track users. It is not. The shielded pool is not a pass-through, it is a cryptographic mixing environment. A user can send funds through multiple z→z hops before exiting, completely breaking any link between entry and exit. Without knowing how many intermediate z→z hops occurred, entry/exit visibility is insufficient for attribution.

CLAIM 04

“Arkham has labeled $420 billion in Zcash volume”

ARKHAM INTELLIGENCE

“Arkham has now labeled more than half of all Zcash activity, attributing $420 billion in volume to known individuals and institutions.”

ZECLEDGER RESPONSE

This claim conflates exchange attribution with individual privacy analysis. The overwhelming majority of that $420 billion will be exchange deposit/withdrawal addresses — Binance, Coinbase, Gemini etc. — which are publicly known by design. Labeling Binance’s deposit address as “Binance” is trivial and requires no cryptographic analysis. It tells us nothing about whether individual user privacy is compromised. The more meaningful question is: how much z→z transaction activity has Arkham de-anonymized? By their own admission, the answer is zero, because z→z is cryptographically opaque.

CLAIM 05

“z→z transactions remain cryptographically opaque”

ARKHAM INTELLIGENCE

“Shielded transactions remain opaque: z→z activity cannot be tracked, and shielded addresses appear on Arkham labeled simply as ‘SHIELDED.’”

ZECLEDGER RESPONSE

Arkham correctly concedes this and it is the most important statement in their entire paper. zk-SNARKs provide a mathematical guarantee that z→z transactions cannot be traced. This is not a practical limitation of current tools. It is a fundamental cryptographic proof. No amount of blockchain intelligence can break zk-SNARK privacy without breaking the underlying mathematics which would also break most of modern cryptography. Our data confirms this: even Blockchair, one of the world’s largest analytics platforms, returns NULL for shielded transaction fields. Our shield rate findings (10–33% across different periods) represent a lower bound of truly private, untraceable activity on the Zcash network.

What the Data Really Shows

Taking all our verified data together, a clear picture emerges that is more nuanced than Arkham’s framing:

Interpreting the 2026 Dip

The drop from 33% in 2024 to 10% in May 2026 is worth examining. This does not necessarily mean privacy adoption has declined. Several factors may explain this:

1. Increased exchange activity. Bull market periods bring more exchange deposits/withdrawals which are always transparent. More exchange volume dilutes the shield rate percentage even if absolute shielded volume is stable or growing.

2. Sampling bias. Our latest data samples the highest-volume recent blocks which skew toward exchange transactions. Lower-volume blocks tend to have higher shield rates.

3. Our detection method undercounts. As noted in methodology, z→t transactions (shielded funds exiting to transparent) are counted as transparent in our heuristic, underestimating true shielded activity.

KEY FINDING #2

Our shield rate figures are a conservative lower bound. Because Blockchair cannot see inside shielded transactions (returning NULL), and because our heuristic counts z→t transactions as transparent, the real shield rate is likely higher than our numbers show. This means Arkham’s traceability claims are even weaker than our analysis suggests.

The Bottom Line

Arkham Intelligence’s research is technically accurate in its narrow scope. Transparent Zcash transactions are traceable. Exchange addresses are publicly known. Entry and exit points of the shielded pool are visible. These are facts.

But the research is misleading in three important ways:

1. It uses historical framing to describe a network that has evolved significantly. Our data shows shield rates have grown from 7% in 2019 to 33% by 2024 — a more than 4x increase. The Zcash network is moving toward privacy, not away from it.

2. It conflates exchange attribution with individual privacy analysis. $420 billion in “labeled” volume is overwhelmingly exchange addresses — publicly known by design. No individual user privacy was compromised to produce that number.

3. It downplays its own most important admission. Arkham admits z→z transactions cannot be tracked. But it buries this admission after paragraphs of traceability framing. With 10–33% of current transactions being genuinely shielded and that number growing to a significant and increasing portion of Zcash activity is cryptographically private.

Final verdict: Zcash’s privacy works exactly as designed. If you use shielded transactions, you are private. The question is not whether Zcash privacy works — it does, but whether enough users are choosing to use it. That number is growing.

About This Research

This report was produced using ZecLedger — an open-source AI-powered Zcash accounting, reporting and research copilot built in Rust. All data is reproducible using the commands documented in Section 03.

GitHub: https://github.com/vancube2/zecledger

Data sources: Blockchair API · Zebra full node v4.4.1 · Claude AI (Anthropic)

Built for: ZecHub Hackathon 2026, Accounting Track

License: MIT Open Source

Disclaimer

“This research was conducted using ZecLedger (github.com/vancube2/zecledger), an open-source MIT-licensed tool. I am not affiliated with Arkham Intelligence. This is independent research. ZecLedger has no token and does not handle user funds.”