Thirteen point five joules per terahash. That is the operating point of Bitmain's Antminer S21 XP — the efficiency ceiling that mass-produced SHA-256 silicon has reached in 2025. Now read ASICID Inc.'s spec sheet. The IDMINER HomeRack claims 0.54 J/TH. The IDMINER 2 claims 0.54 J/TH. The IDMINER 1 claims 0.61 J/TH.
That is not an incremental gain. It is a claim of roughly 25 times the efficiency of the best chip any foundry on earth can currently manufacture. Data speaks louder than sentiment, and the data here is not a breakthrough. It is a fabrication dressed as engineering.
I spent three months in 2018 auditing the 0x protocol's v2 contracts, and I learned one thing that has survived every cycle since: verify the mechanism, not the marketing. A whitepaper can claim anything. Silicon cannot. When a spec sheet violates the physical limits of the process node it is supposedly built on, the correct response is not excitement. It is dismissal. I have watched that principle save capital through two bear markets — the 2022 deleverage that preserved 60% of my book, the rug-pulls that wiped out 90% of my peers. Every one of those failures started with a number that was too good to check.
So let me apply the same standard to ASICID that I applied to those reentrancy bugs — read the code, check the physics, and let the numbers indict themselves.
Context
ASICID Inc. describes itself as a Hong Kong-headquartered mining hardware company with US operations. Its flagship pitch, distributed via a paid press release on BeInCrypto, is a "new generation" of dual-algorithm ASICs: SHA-256 for Bitcoin, Scrypt for Litecoin and Dogecoin. The company lists three SKUs, promises a "zero-fee mining pool," and headlines projected monthly revenue of up to $25,590 per unit. The stated target market spans home miners, professional operations, and institutional buyers.
There is no team. No named founders. No foundry partner — no TSMC, no Samsung, no disclosed process node. No tape-out record. No third-party benchmark. No patent. No funding round. No exchange relationship. No delivered customer, no installed base. The entire evidentiary foundation is ASICID's own self-description, repeated across fifteen information points, all sourced to the same entity.
That is a single-source, self-interested disclosure. In any audit, that is the first thing you flag. Not because it is necessarily false — but because it is unfalsifiable, and unfalsifiable claims are how capital gets destroyed. When I audited 0x, the vulnerabilities I found were not hidden in obfuscated code. They were sitting in plain sight, in functions that everyone assumed worked because the documentation said they worked. The documentation is never the audit.
The press release itself matters. BeInCrypto and comparable outlets routinely publish project-supplied product announcements without independent editorial verification. This is not journalism. It is advertising with a masthead, and it should be priced accordingly. Every mainstream miner launch — Bitmain, MicroBT, Canaan — ships with chip lithography, foundry attribution, and measured hashrate curves. This one ships with a headline number and a zero-fee promise. The absence is the signal.
There is also a technical ambiguity the release never resolves. Bitcoin runs SHA-256. Litecoin and Dogecoin run Scrypt. These are two entirely different algorithms requiring two entirely different silicon designs. ASICID describes "dual-algorithm" mining without stating whether this is a single die integrating both engines, a multi-board assembly, or something else entirely. No architecture, no die shot, no thermal design. A hardware claim that cannot describe its own silicon is not a hardware claim. It is a narrative.
Core
Run the arithmetic. The physics is unforgiving.
The IDMINER HomeRack advertises 9,600 TH/s drawing four 1,300-watt supplies — 5,200 watts total. Divide. You get 0.54 J/TH. The IDMINER 2 posts 2,400 TH/s at 1,300 watts. Same number: 0.54. The IDMINER 1 lists 1,150 TH/s at 700 watts: 0.61 J/TH. The internal consistency is almost convincing — until you compare it to reality.
Industry-best SHA-256 efficiency sits near 13–15 J/TH. ASICID's claim is 22 to 25 times better. To move from 13.5 J/TH to 0.54 J/TH is not a next-generation improvement. It requires crossing roughly four to five semiconductor process generations — the kind of leap that takes a decade of foundry investment and billions in capex, not a press release from an unlisted company with no fabrication partner. Even the theoretical scaling limits of the industry's announced nodes do not reach this number. It is not ambitious. It is impossible.
The Scrypt claim is worse. The Antminer L7 — the reference Scrypt machine — runs at approximately 360 W/GH. ASICID advertises 1.6 to 2.0 W/GH. That is a claimed efficiency advantage of 180 to 220 times. There is no process node, no architecture, and no theoretical path that produces that number. When a spec exceeds the physical envelope of the technology by two orders of magnitude, you are not looking at innovation. You are looking at a number designed to be quoted, not tested.
Then examine the deployment architecture, because this is where the real mechanism lives.
ASICID ships units "pre-configured" and steers buyers toward its own zero-fee pool. Two things fall out of that. First, a firmware you do not control is a firmware that can be changed — redirected, throttled, or skimmed — after you have paid. The unit arrives sealed, signed by the vendor, and permanently subordinate to the vendor's keys. Second, a "zero-fee" pool is not charity. Pools carry real operating costs — bandwidth, infrastructure, engineering, payout settlement. Free mining pools are loss leaders. The fee is not eliminated; it is relocated somewhere less visible, typically into hashrate redirection or future pricing power once lock-in is achieved.
This is the oldest structure in the hardware playbook: sell the machine, then own the pipe. The unit is the hook. The pool is the lock. Once your hashrate enters their infrastructure, the vendor controls the settlement layer and the data — and you control nothing. In DeFi, we learned to read the contract before the yield. In hardware, the contract is the firmware, and nobody is reading it.
Now the revenue math. A projected $25,590 per month assumes optimal conditions simultaneously: high coin price, low network difficulty, full duty cycle, no downtime. Every one of those variables is outside the buyer's control and most are currently moving the wrong way. Post-halving, miner revenue per unit of hashrate has compressed sharply. Hashprice is depressed. Home miners are precisely the cohort for whom breakeven horizons have stretched, not shrunk.
The projection is a hook. It is the same psychological lever as a 10,000% APY farm — a number engineered to short-circuit the part of the brain that does arithmetic. I have written the formulas for impermanent loss because the gap between theoretical and actualizable yield is where retail gets harvested. The same gap exists here, multiplied by hardware counterparty risk. And note the release's own escape hatch: it labels the revenue "estimated." That single word pre-loads the disclaimer. If the machine returns a fraction of the promise, the vendor has already told you it was never guaranteed.
And note what is missing from the release entirely: the price. ASICID never states what these machines cost. That omission is not an oversight. A spec sheet without a price cannot be compared, and a product that cannot be compared cannot be rejected on value. It is a deliberate blank in the sales document.
Contrarian
Here is the counter-intuitive part, and it is where most readers misplace their skepticism.
The instinct is to ask whether ASICID is a scam. That is the wrong question. The right question is why the claim propagates at all — and the answer is structural, not moral.
Mining hardware is an oligopoly. Bitmain holds roughly 70–80% of the market. MicroBT holds 10–15%. Canaan is the listed remainder. ASICID appears in none of the industry's market statistics. A new entrant with no tape-out history claiming to beat the incumbent by 25x is not a competitive threat. It is a category error. Real silicon competition happens at the margin — 5% efficiency gains, delivered on schedule, backed by foundry capacity. It does not arrive as a 2,500% leap from a company nobody has heard of.
So why does anyone believe it? Because the post-halving environment manufactures demand for exactly this story. When legitimate mining margins compress, the audience for "high-return home mining" expands. Panic sells, logic buys — and the seller of this narrative is betting that compressed margins make buyers stop doing logic. The information-asymmetry target is not the institutional operator who demands benchmarks. It is the retail miner in an emerging market, comparing a promised monthly return against a local salary.
The smart money does not chase the machine. It watches the pipe. Institutional flows — ETF arbitrage desks, large miners — price hardware on verifiable efficiency curves and delivered hashrate. They require benchmarks before wire transfers. Retail is asked to wire first and benchmark never. That asymmetry is the product. Not the miner. The asymmetry.
Takeaway
Three things to verify before a single dollar moves, and none of them require trusting me.
Demand a third-party benchmark on delivered hardware — power draw at the wall, measured hashrate, sustained over hours, not a spec sheet. Check the WHOIS record on asicid.com and the corporate registration in Hong Kong and the relevant US state; a recently registered domain and an opaque shell structure tell you more than any press release. And never prepay. Cash-on-delivery or escrow, or you are not buying a miner — you are extending unsecured credit to an entity with no verifiable balance sheet, across two jurisdictions chosen for their distance from accountability.
The question worth sitting with is not whether 0.54 J/TH is possible. It is why a number that violates semiconductor physics still travels faster than any correction. Liquidity dries up when trust breaks — and trust in hardware claims breaks the moment someone finally plugs the machine into a wattmeter.
Someone will. The only open question is whether they do it before or after you have paid.
