Uncategorized

How Bitdeer Is Transforming Bitcoin Mining Machines

Published

15 часов ago

14 марта, 2025

admin

Application-specific integrated circuit (ASIC) chips form the backbone of the bitcoin (BTC) mining industry. ASIC machines are made for a single purpose: To solve Bitcoin’s SHA-256 algorithm as fast as possible in order to collect block rewards.

They’re extremely good at it. One of the most widely used ASIC machines, the Antminer S19, is capable of making 82 trillion computations per second — 820 times the number of stars in the Milky Way. The $30 billion ASIC manufacturing market is dominated by Bitmain. The Chinese company’s machines power roughly 80% of Bitcoin’s hashrate, according to TheMinerMag.

But Singapore-based bitcoin mining firm Bitdeer (BTDR) intends to shake things up with the release of a new ASIC chip architecture. These new chips could bring a huge jump in efficiency, the company claims, while improving transparency in the ASIC manufacturing process.

“The two dominant players [Bitmain and MicroBT] are both private companies, and very opaque,” Jeff LaBerge, head of capital markets and strategic initiatives at Bitdeer, told CoinDesk in an interview. “They don’t really engage with the media or give any type of guidance about what they’re doing from an R&D standpoint, and that makes it very difficult for end-buyers to plan.”

“We want our customers to know where we’re at in our manufacturing process, what our roadmap is in terms of new chip designs, where we’re at in our production cycles,” LaBerge said.

Shanon Squires, chief mining officer at bitcoin hosting firm Compass Mining, told CoinDesk that increased visibility into ASIC production would help miners plan new hardware shipments and make it easier to predict Bitcoin’s difficulty growth. “Bitdeer’s commitment to transparency is great for the mining industry,” she said.

“While Canaan discloses its annual sales volume for various mining models, Bitdeer takes it a step further by providing more frequent delivery volume updates,” Wolfie Zhao, head of research at TheMinerMag, told CoinDesk. “Although both are smaller players in the hardware market, their efforts show good faith in promoting transparency. Hopefully, this will encourage the larger market incumbents to take note.”

Seeking efficiency

ASIC chips have used mostly the same blueprint since 2014. Over the last decade, the biggest increases in ASIC power efficiency have come at the foundry level, as leading global chipmaker TSMC has refined its manufacturing process. While miners have also made alterations to chip design, such modifications have only brought incremental gains.

Even so, progress has been tremendous. The very first ever ASIC, Canaan’s Avalon (2013) had a power efficiency of 6,000 joules per terahash (J/TH). Bitmain’s Antminer S21XP Hydro, the current most efficient machine on the market, boasts 12 J/TH efficiency.

Bitdeer, which is listed on Nasdaq, wants to create a completely new architecture for its ASICs. “We feel like it’s going to be necessary to break into what we call the single-digit efficiency range,” LaBerge said, referring to mining rigs with less than 10 J/TH in efficiency.

Scaling up with the traditional blueprint means using progressively thinner chips. But thinner means chips are more likely to be defective and yields per batch tend to fall. “You’re also competing with Apple and Nvidia and some of the biggest companies in the world for the same materials,” LaBerge said.

Bitdeer’s Chief Strategy Officer, Haris Basit, is leading a team of engineers to create a new framework. Some members of that team worked on designs found in Bitmain’s first ASIC chips back in 2014 — the chips whose architecture became the standard across the industry. (Bitmain did not respond to a request for comment.)

Bitdeer’s research has already had successes. The company’s most recent product, the SEALMINER A3, achieved a power efficiency of 9.7 J/TH during performance trials, the firm reported on Monday. That means the A3 — which still uses the traditional ASIC blueprint — could end up taking the efficiency crown from the S21XP Hydro.

Yet the miner’s SEALMINER A4, which will employ the firm’s new chip architecture, is expected to consume 5 J/TH. It will likely be the most efficient ASIC machine on the market by a significant margin.

“People have known for a long time that you could recycle [the electric] charge on a chip, but no one’s really been able to figure out how to do that in a way that allows for high performance… We’ve cracked the code on how to do this in a very high performance application,” Basit told the Coin Stories podcast in December.

“Instead of just using [charge] once and discharging it, we use it several times, four, five, six times. So we get [a] 75-80% improvement in efficiency by doing that,” Basit added.

“Our SEALMINER A4 chips will use this technology, but it should also be applicable more generally in digital chips, especially digital chips that are highly active, like GPUs and signal processing chips.”

Manufacturing chips

Making ASICs isn’t easy. Bitdeer’s research team is divided into two units (one in Singapore, another in Silicon Valley) that both work on new chip designs. “For such a simple machine — all it does is solve the SHA-256 algorithm — it’s extremely complicated to design. We’ve got some of the best engineers in the world working on this,” LaBerge said. The company spends approximately $6-8 million on research per quarter.

So far, the firm has been delivering new products at a fast pace. Bitdeer pushed out both the SEALMINER A1 and A2 in 2024 and is expecting the A3 to enter mass production in the latter half of 2025. It says the A4 should reach tape-out (the last stage of its designing process) in the third quarter of the year, with a release likely in late-2025 or early 2026.

When a new chip design is finalized, Bitdeer sends the plans over to TSMC. Not only is the Taiwanese firm the largest chip manufacturer in the world, it’s also the most advanced on a technological level, which makes Bitdeer’s partnership with it crucial.

“You can’t just go to TSMC and say, ‘Hey, I want 100 exahash worth of chips in the next three months.’ There’s a process for going through that,” LaBerge said. “You go in and ask them for chip allocation, and they’ll give that based on priority.”

Once it has the plans in hand, TSMC produces a mask, which essentially functions as a template for chips — like the platen in a printing press. The mask is sent to Bitdeer alongside risk chips (a small batch of chips that the company can use for trials) to make sure the design works properly. If the firm needs any alterations to be made to the design, that’s when it happens. In that case, TSMC makes corrections based on Bitdeer’s feedback and sends over a new mask with new risk chips. All of this happens at significant cost. Bitdeer spent $14 million on the A2’s tape-out and $26 million on the A3’s.

When Bitdeer is satisfied with a design, TSMC uses the mask to mass-produce wafers. LaBerge compared wafers to sheets, each containing hundreds of chips. Technically, a mask can be used to create an almost unlimited number of wafers, but TSMC has finite resources and can only produce a certain number of chips, so firms end up competing for them.

One of the advantages of the A4’s design, according to LaBerge, is that it’s supposed to make the firm’s chip allocation process easier. “[Basit] challenged the team to come up with a new architecture that didn’t need to undergo TSMC’s latest processes, but could step back a couple of generations, which would allow us to use a node that is much less in demand,” he said. A semiconductor node is basically a specific version of the firm’s chip manufacturing technology; TSMC is constantly building new nodes in an effort to refine its processes.

It takes roughly three months for Bitdeer to receive its mask and risk chips after first submitting its design to TSMC. Then, it’s another three or four months for the company to receive its chips once it has given the foundry the green light for mass production. The chips are sent straight to Bitdeer’s manufacturing facilities in Asia. From there, it can take four to eight weeks for the mining rigs to be fully built and packaged.

Aiming for the top

Despite all the costs incurred during production, some of the capital required for manufacturing ASICs comes from Bitdeer’s customers.

Miners interested in purchasing Bitdeer’s ASICs typically put down a deposit of 25% to 50% of the total cost of the order. The production cycle tends to average at six to seven months, so it doesn’t take long for the company to recuperate its funds and make a profit.

Building ASICs also creates advantages for Bitdeer’s own mining operations. Up until recently the firm, which was founded in 2021, focused on the hosting business, meaning that it provided facilities for other bitcoin miners to place their rigs. Bitdeer is slowly transitioning out of that model and expanding its own mining operations alongside its ASIC manufacturing arm.

The acquisition of ASICs is typically the most expensive part of building up a bitcoin mining operation. These machines usually only last around three or four years before newer models make them obsolete, so bitcoin mining firms are constantly looking to acquire more.

Not only is Bitdeer able to considerably reduce these costs by producing its own machines, but it also has the option of selling its mining rigs to other firms depending on its needs.

Down the line, Bitdeer aims to give Bitmain and MicroBT a run for their money, and disrupt what LaBerge called the duopoly of the ASIC market. “We want to be the top player in the market, absolutely,” LaBerge said. “We believe we have the team and the technology to do that.”

UPDATE (March 14, 2025, 15:30 UTC): Added the cost of the SEALMINER A3 tape-out.

Uncategorized

Cardano: Deep Dive on the Trump Reserve Token Whose Blockchain Ignores TVL

Published

12 часов ago

14 марта, 2025

admin

Trading volumes for Cardano’s ADA token have exploded of late with daily figures averaging around $720 million in February while exceeding an average of $1.4 billion in March.

This rise was spurred by a social media post by U.S. President Donald Trump, who mentioned ADA as one of the tokens that would be included in the nation’s strategic crypto reserve.

Although Cardano is enjoying its moment of mainstream attention, the layer-1 blockchain has been quietly emerging as a crypto juggernaut since it went live in late 2017.

Adoption metrics

The ADA token has a market cap of $25.6 billion but what’s more notable is what’s under the hood; data from Google shows that the Cardano blockchain has more than 5 million unique wallets and 1.3 million delegators, with thousands of new wallets being created per day.

The blockchain also has $329 million in total value locked (TVL), although Cardano Foundation CEO Frederik Gregaard believes that metric is overemphasized by crypto communities.

Instead, he points to «non-value transactions» associated with people conducting real-world – albeit non-financial – activities on blockchain rails: Minting a decentralized ID, tracking metadata, recording documents, that sort of thing. Cardano’s a hotbed of such activity, he said.

«I’m fighting to ensure that 50% of the activity is a non-value transaction,» Gregaard told CoinDesk.

One example of this is Cardano’s partnership with Veritree, which saw the Cardano community donate over 1 million ADA tokens to plant 1 million mangrove trees in Kenya, with each donation verified and tracked on the blockchain.

Last week, the Cardano Foundation also announced a deal with SERPRO — Brazil’s largest state-owned IT company – to accelerate blockchain adoption in South America. SERPRO processes 33 billion transactions annually for 90% of Brazil’s federal administration. Additionally, 8,000 employees will also receive blockchain training.

Cardano’s perspective differs from the likes of Solana and the slew of layer-2 networks like Base that pride themselves on total value locked (TVL) and hype-driven movements like memecoins and non-fungible tokens (NFTs).

TVL on Solana grew from $2.2 billion to more than $10 billion in 2024, Cardano meanwhile zipped from a modest $445 million to $537 million in the same period.

DeFi on Cardano

Whilst Cardano Foundation’s CEO said his focus is on real-world use cases, the blockchain still boasts a bustling DeFi ecosystem under the surface.

Minswap is Cardano’s native decentralized exchange (DEX). Its cumulative trading volume hit $3.4 billion this month with December alone notching a near-record $271 million, DefiLlama data shows.

There are also a number of lending protocols including Liqwid, Lenfi and Optim Finance, with TVL across Cardano’s lending sector exceeding $116 million.

But the key part of Gregaard’s mission, he insists, is not to exceed that 50% level for financialized transactions. He sees it as staying in line with the Cardano Foundation’s non-profit ethos, even if it limits potential exponential growth of hype-fueled movements like memecoins.

Cardano Foundation vs Hoskinson vs Emurgo

Fulfilling that ethos has its own challenges, mostly because the blockchain is run by three main entities: the Cardano Foundation, Charles Hoskinson’s IOG and Emurgo. The latter two are commercial businesses, which can cause friction between them and the foundation.

«The intent of having a non-profit was that you can optimize decision-making based on 10 years, it’s different than if you optimize decision-making tomorrow,» Gregaard added.

Some of the friction was highlighted by an anonymous Cardano community member in December, who penned an email on a path forward and detailed how the entities running Cardano were at loggerheads.

«CF’s recent burst of activity is part of a larger strategic play—an attempt to undermine Charles, IOG, Intersect, and the broader governance roadmap,» the email read.

«It’s been a long and difficult road, but I do agree with some of the sentiments of the whistleblower,» Hoskinson wrote in response on X.

Gregaard, however, was more diplomatic about any potential rift.

«There’s no monetary exchange going on between us, but we do work very closely together,» he said.

«We sometimes go to [a conference] and we share a booth. So we come together and we sponsor booths together, that’s the closest you will get to any affiliates, which is very different compared to both the Ethereum foundation or Tezos foundation, where they basically control the Treasury and control the disbursements.»

«On the flip side, we [Cardano Foundation] are the liability umbrella for the community and the blockchain, which means that we are the one who interacts with the SEC and the CSDC and the FMA, and I negotiated MICA with the European Parliament.»

Uncategorized

Court Approves 3AC’s $1.53B Claim Against FTX, Setting Up Major Creditor Battle

Published

13 часов ago

14 марта, 2025

admin

The Delaware bankruptcy court handling the FTX estate approved a petition on Thursday from Three Arrows Capital (3AC) to significantly expand its claim against the estate from $120 million to $1.53 billion, marking a major development in the ongoing fallout from the collapse of Sam Bankman-Fried’s crypto empire.

3AC, once a dominant crypto hedge fund with over $3 billion in reported net assets, collapsed in 2022 while it still had deep financial ties to FTX, Sam Bankman-Fried’s soon-to-collapse crypto exchange. The hedge fund initially filed a proof of claim worth $120 million against FTX in July 2023 — adding its name to a long list of users and investors in FTX who lost money as a result of its sudden insolvency.

In November 2024, 3AC’s liquidators amended their claim after discovering new evidence suggesting that FTX had liquidated $1.53 billion in 3AC’s assets just two weeks before the hedge fund commenced its own liquidation proceedings two years prior. They argued that FTX’s liquidation of 3AC’s funds was carried out to satisfy a $1.3 billion liability to FTX, an obligation that 3AC claimed was not sufficiently substantiated.

FTX’s bankruptcy said the $1.3 billion liability represented collateral for a loan FTX made to 3AC, but the court ruled in favor of 3AC, finding insufficient evidence to support FTX’s loan claim.

The ruling allows 3AC to pursue a significantly larger portion of FTX’s remaining assets, potentially reshaping creditor payouts.

FTX, which began distributing funds to creditors in February 2025, said the expanded claim should have come sooner, arguing that it would burden other creditors and complicate its reorganization plan. The court, however, determined that 3AC’s delay was justified, given that the liquidators only uncovered the full extent of their claim in mid-2024 due to missing financial records from FTX and a lack of cooperation from 3AC’s founders, Zhu Su and Kyle Davies.

3AC, founded in 2012, had grown into one of the most influential financial firms in the cryptocurrency industry by 2022. Its collapse was among the first and largest dominoes to fall before the broader crypto market imploded in 2022, which ultimately set off the chain of events that revealed fraud in Sam Bankman-Fried’s crypto empire.

Bankman-Fried is currently pursuing an appeal of his criminal conviction and 25-year prison sentence. Following the collapse of 3AC, Su was detained in Singapore and sentenced to four months in prison for failing to cooperate with 3AC’s liquidators. Davies did not face any charges connected to the hedge fund’s collapse.

The 3AC founders reunited in 2023 to launch a short-lived crypto exchange called OPNX — designed to allow users to trade bankruptcy claims of failed crypto companies — which shut down in February.

With the court’s decision, 3AC’s liquidators now have a significantly larger position in the FTX. bankruptcy proceedings, raising questions about how the expanded claim will impact distributions to other creditors. The ruling also underscores the lack of transparency at both FTX and 3AC — further complicating efforts to untangle both firms’ assets and obligations.

Uncategorized

Weekly Recap: Regulatory Wins, Market Doldrums

Published

14 часов ago

14 марта, 2025

admin

It was a week of red in crypto and traditional markets, with bitcoin plummeting below $80K on March 10 and ETH falling to $1,821 the same day. So much for the “Trump Bump.” With the new administration on a tariff-tear this week, the markets were spooked about a recession and crypto was not immune.

Still, progress in digital assets was all around, and our reporters reported it all with alacrity. BlackRock’s bellwether BUIDL fund topped $1 billion and tokenized treasuries hit $4.2 billion, Kris Sandor reported. MoonPay, a payments aggregator, made an important stablecoin acquisition, Will Canny wrote. Ripple won a payments license in the UAE (Shaurya Malwa). OKX won a license to operate in Europe, Camomile Shumba reported. Coinbase announced plans to offer 24/7 futures trading in the U.S., Helene Braun reported.

There was also big regulatory news. The U.S. House voted to overturn the IRS’s controversial “broker rule” in a big win for DeFi operators. And a Senate committee voted to send the GENIUS stablecoin bill to the floor, ahead of probable approval there.

The Trump Family continued to be front-and-center in the crypto news. World Liberty Financial completed a $590 million token sale (for accredited investors for now), with an assist from adviser/investor and TRON founder Justin Sun. The Wall Street Journal reported that a Trump family representative also explored buying a stake in Binance.US, through World Liberty Financial.

From our Asia team, Sam Reynolds examined how the latest draft of the GENIUS act aims to split stablecoin regulation between state and federal authorities.

Parikshit Mishra reported on Coinbase returning to India after a two-year hiatus, setting off discussion about the future of crypto in India.

Shaurya Malwa continued his excellent reporting on XRP, pushing out multiple reports on Ripple. Malwa also reported on the implications of over-leveraging in the crypto market, as Hyperliquid lost $4 million due to a massive leveraged trade in ETH.

Market maven, Omkar Godbole, pushed out a timely piece on bitcoin’s bullish signal ahead of the U.S. CPI report, and was also early to spot how Eric Trump’s tweet on crypto were setting up short-term traders for disappointment.

Meanwhile, Tom Carreras had an excellent feature on how Bitdeer, a Singapore-based miner, hopes to shake up the mining machine market.

Hopefully next week brings better news in the markets. But, either way, our reporters will be there to cover what matters.