Canaan Inc. (CAN) Earnings Call Transcript & Summary

Ladies and gentlemen, thank you for standing by, and welcome to Canaan, Inc.'s Fourth Quarter and Full Year of 2024 Earnings Conference Call. [Operator Instructions] Please note that this event is being recorded. Now, I'd like to hand the conference over to your speaker host today, Ms. Gwyn Lauber, Investor Relations Director of the company. Please go ahead, Gwyn.

Thank you, operator. Hello, everyone, and welcome to our earnings conference call. Joining us today are our Chairman and CEO, Nangeng Zhang; and our CFO, Jin James Cheng. Leo Wang, Vice President of Capital Markets and Corporate Development; and Xi Zhang, Senior IR Manager, will also be available during the question-and-answer session. Our CEO will start the call by providing an overview of the company and performance highlights for the quarter. Our CFO will then provide details on the company's operating and financial results for the period before we open up the call for your questions. Before I begin, I would like to refer you to our safe harbor statement in our earnings press release. Today's call will include forward-looking statements. These statements include, but are not limited to, our outlook for the company and statements that estimate or project future operating results or the performance of the company. These statements speak only as of today, and the company assumes no obligation to revise any forward-looking statements that may be made in today's press release, call, or webcast, except as required by law. These statements do not guarantee future performance and are subject to risks, uncertainties, and assumptions. Please refer to the press release and the risk factors and documents we file with the Securities and Exchange Commission, including our most recent annual report on Form 20-F, for information on risks, uncertainties, and assumptions that may cause actual results to differ materially from those set forth in such statements. In addition, during today's call and webcast, we will discuss both GAAP financial measures and certain non-GAAP financial measures, which we believe are useful as supplemental measures of the company's performance. These non-GAAP measures should be considered in addition to and not as a substitute for or in isolation from GAAP results. You can find additional disclosures regarding these non-GAAP financial measures, including reconciliations with comparable GAAP results, in our earnings press release, which is posted on the company's website. With that, I will now turn the call over to our Chairman and CEO, Nangeng Zhang. Please go ahead.

Hello, everyone. This is Zhang Ji, CEO of Canaan. Thank you for joining our call. James, our CFO, and I are pleased to speak with you from our headquarters in Singapore and share our updates for the fourth quarter and the recent developments. In the fourth quarter of 2024, the trends of Bitcoin price and the total network hash rate were relatively favorable for miners. The Bitcoin price steadily rose from USD 61,000 at the beginning of the quarter to USD 93,000 at the end. The total network hash rate also increased from 622EH/s to 842EH/s. That said, in Q4, Bitcoin's price grew faster than the hash rate. This is a typical electricity cost of USD 0.06 per kilowatt hour for miners using our A50 machines. So their gross margin rose from about 40% at the start of the quarter to as high as 60% by the end. This created a more profitable environment for mining and boosted miners' confidence. We fully leveraged this positive market trend and pushed sales and the delivery of mining machines to grow the energized hash rate for our mining operations. We achieved approximately USD 89 million in total revenues this quarter, well above our guidance of USD 80 million. For the full year, we achieved almost USD 270 million in revenue, up 27.4% year-over-year. For mining machine sales, we focus on new product presales. The high-performance A15 series officially entered mass product delivery in Q4. We sold 9.1 million terahash per second of computing power, up 66% year-over-year and up 24.7% quarter-over-quarter, reaching a record high in our company's history. This contributed USD 73 million to mining machine revenue, up 64% year-over-year and making a 2-year quarterly high. In this quarter, we fulfilled large orders from North American public mining companies like Cipher, CleanSpark, and HIVE and received strong positive feedback. For the full year of 2024, North America became our largest regional market, accounting for approximately 40% of mining machine sales, showing the success of our expansion in the region during the past years. The rising Bitcoin price also helped us clear career generation inventory. In Q4, the period generation products remained popular in regions with energy advantages like Southeast Asia and Africa. To date, sales of inventory of traditional models like A16 and A14 have been mostly cleared, and A15 has become our main product. This completes the upgrade of our inventory portfolio. Our first consumer product, the NanoStream mining heater, continues to be welcomed by individual users around the world. By year-end, consumers from 80 countries around the globe have ordered our products through our online shop. Since their launch, we received a total of 24,000 Nano orders with more than 50% coming from customers in Europe and North America. This also helped increase our brand recognition in these regions. In 2025, we launched several new Avalon Home products, including the Nano 3S, Avalon Mini 3, and Avalon Q. These new products are available on our website for preorder, with small batches being delivered in Q1. As of now, total orders for these new consumer-grade products have reached 12,000 units with a total amount of over USD 5 million. Some of these orders also come from distributor clients who have now established cooperation with us. Considering that the Northern Hemisphere will soon enter summer, indoor heating demand will gradually decline. This level of presale performance has greatly exceeded our expectations. It also shows that many individual customers are eager to try and participate in Bitcoin mining. Our mining business also delivered a strong performance this quarter. We mined 186 Bitcoins, up 84% year-over-year with higher online hash rates and stronger Bitcoin prices, our mining revenue reached USD 15.3 million, up 313% year-over-year. We continued to take advantage of low electricity costs in Africa and expanded local mining capacity. By the end of 2024, our total installed mining hash rate reached 5.4 EH/s per second, with 4.8 EH/s per second energized. Our average all-in power cost remained at USD 0.04 per kilowatt hour. With Bitcoin price trending up in Q4, our mining gross margin increased to 42%, much higher than the 22% in Q3. Our efficient mining operations also helped grow our Bitcoin holdings. By the end of 2024, our own Bitcoins on the balance sheet reached 1,293, a new high. The rising price of Bitcoin at quarter end also resulted in a fair value gain to this asset. This gain shows the preliminary benefits of our mining operations as a long-term strategy and our decision to hold Bitcoin. As our mining business expands, we provide additional information to the capital markets, as is standard with the other public miners. Starting in January 2025, we began releasing monthly mining updates. By the end of February 2025, we announced that we had 7 active joint mining projects worldwide, with 6.28 EH/s installed and 5.73 EH/s energized. Our own Bitcoin holdings had reached 1,355 by the end of February. As we previously announced, since Q1 2025, we have expanded 2 North American mining projects through upgraded partnerships. We are moving toward our mid-2025 goal of reaching 10 EH/s per second deployed in North America and 15 EH/s per second globally. We expect our mining business to achieve a new high in revenue in the first quarter of 2025. Today, we are also pleased to announce that we recently signed new cooperation agreements with 2 partners to expand mining capacity at sites in Pennsylvania and Texas. Once these 2 projects gradually go online in Q2 2025, we expect to increase our operating hash rate in North America to approximately 4.7 EH/s per second. We have already built a strong and experienced operations and business development team in North America, and they are actively in discussions with several potential partners for joint mining. Based on our current observations, the cryptocurrency site resources in the North American market still show strong potential, which is in line with our earlier strategic assessment. The expansion of joint mining and hosted mining in North America has been progressing very smoothly. As a manufacturer with minor research and development capabilities, we are increasingly welcomed by mining site owners for joint mining ventures. The new U.S. administration is expected to ease regulations, provide more energy resources to the industry, and support Bitcoin mining at both the federal and state levels. Recently, our key competitor, a critical affiliate company, has been added to the U.S. entities list. We believe we will benefit from the growing tension and rising cost in the mining hardware supply chain and that many mining site owners will increasingly view Canaan as a preferred partner. We also hope to reinvest more of the funds raised in U.S. capital markets back to local projects and gradually build ourselves into a company with a strong mining footprint in the United States. Both our mining machine sales and mining operations rely on continuous R&D efforts. In the fourth quarter, we successfully ramped up mass production of the A15 series. Working closely with our foundry partners, we carried out ongoing process improvements to boost product performance and yield rate. Looking ahead, we expect to deliver more computing power using the same number of wafers, including more high-end A15 models. In addition to our standard IR code miners, we continue to provide customized mining solutions based on clients' specific needs and their site conditions. For example, in our previous collaboration with the public [indiscernible], we delivered 3,800 units of our 1566i emerging Cool miners. These machines ran stably on site, performed beyond the contract satisfactions, and received very positive feedback from the client. In 2024, although mining chips were generally not subject to U.S. export control restrictions at the time, we continue to proactively manage our supply chain to prepare for potential regulatory changes in the future. In January 2025, the U.S. BIS issued new rules that certainly imposed additional constraints on the global semiconductor supply chain. Outsourced assembly and testing vendors were now effectively limited to those on an approved white list, drastically narrowing available options overnight. Thanks to our early planning, we were able to swiftly transition A15 chip packaging to OSAT partners within the white list, minimizing delays in the manufacturing process caused by the new regulations, subsequently, through continued collaboration with our foundry partners to support license applications. We were informed this month that all our in-production wafers have received the necessary license to be released even to the OSATs result approved white list status. As a result, our choices of OSAT are largely no longer restricted. This put us in a strong position to scale up production and reduce costs more quickly in response to rising market demand. This is the result of our long-standing commitment to compliance. We will continue to work to anticipate and mitigate the potential impacts of future policy changes. Additionally, we have completed the R&D on our next-generation A16 mining machines and finished the tape-out process. Once again, we are pushing the limits of manufacturing technology due to the advanced chip design of our products. The production is very complex and time-consuming, but we continue to work closely with our foundry partners to bring the ultra-performance product to customers as soon as possible. The standard IR code 816 is expected to reach nearly 300 terahash per second. As usual, we will release the new products and send simple machines to customers after we finish the full system testing and obtain real-world data. In our consumer product line, during the first quarter of 2025, we successfully launched several new additions to the Avalon Home series. This includes the upgraded Nano 3S, desktop mining heater, the Avalon Mini 3, quiet home heater with integrated mining capabilities, and the Avalon Q, our first low-noise home mining device that supports 110 volts and is compatible with major global voltage standards. Through this protocol, silently, strategically designed, and functionally home-oriented consumer products, we aim to make Bitcoin mining more accessible to everyday users and the future promote the concept of decentralization. We invite everyone to visit our official website to learn more about these products. On the operations side, we achieved significant positive changes in the fourth quarter, thanks to the large-scale delivery of our A15 products, strong growth in mining revenue, and continued cost optimization. Our gross loss narrowed sharply by 70.5% quarter-over-quarter to USD 6.4 million. At the same time, after excluding noncash adjustments such as fair value changes and prudent accounting treatment, we turned profitable at the EBITDA level. In Q4, we achieved an adjusted EBITDA gain of USD 19.3 million. This marks our first EBITDA profitability since the mining machines market downturn began 2 years ago. Looking into 2025, we will continue to push forward with the sale and the mass delivery of our A15 series products. Thanks to their strong performance, the test bench of A15 products received high recognition from customers, and orders have been coming in actively. Our delivery schedule is now booked into May and June of this year. Our key focus now is to work closely with our foundry partners to ramp up production capacity and steadily increase our delivery volume. We also aim to improve computing power and yield rate through process optimization in the future, which will enhance profitability, product competitiveness, and customer satisfaction. In addition, we are actively balancing the allocation of mining machines between customer orders and our mining operations. Based on real-time market demand dynamics to achieve better synergy between the 2 business segments. The first quarter is traditionally a slow season. We completed the first phase of our supply chain compliance adjustment mostly during the Lunar New Year holiday. So, the impact on the chip production cycle was kept within a limited range. However, starting in February, major changes in the global political and economic environment have caused significant Bitcoin price volatility. This has negatively affected market sentiment, future expectations, and especially financing activities across the markets. Given these combined factors, we are maintaining a very cautious outlook for the first quarter of 2025. We expected Q1 revenue to be approximately USD 75 million. We expected that in Q2, the production capacity would ramp up and our mining project deployment would move towards completion. Our performance will see a big improvement compared to Q1. We currently estimate the total revenue for Q2 of 2025 to be in the range of USD 120 million to USD 150 million. We are maintaining our previously issued full-year 2025 guidance with total annual revenue expected to be in the range of USD 900 million to USD 1.1 billion. This outlook is based on the company's current market and operational conditions. However, given the recent market volatility and the uncertainty about how long it may continue, actual results may differ. From Beijing to Singapore to Silicon Valley, from delivering the first ASIC miner to going public on NASDAQ and launching home series mining machines, Canaan has been forging ahead on the path of decentralization. Looking back at the past year, we remain firmly committed to R&D and to delivering high-quality, high-performance, and customized Bitcoin mining solutions. Our A15 series has been embraced by demanding customers in the North American market. Our Avalon Home series, which combines Bitcoin mining and home-use functionality, is designed to meet the needs of a diverse global customer base. We have continued to strengthen our global presence, winning orders from and building our strategic partners with leading publicly listed miners in North America, thus raising our brand awareness and market share in the region. As a NASDAQ-listed company, we are inspired by America, empowered by America, and dedicated to America. We have actively leveraged the strength of U.S. capital markets to support our growth, and we are rapidly advancing heavy asset projects such as mining operations in the U.S., which are now beginning to take shape. At the same time, we continue to push forward our global strategy. We are optimizing our strategy for R&D, supply chain, manufacturing, and logistics. We are exploring local production to adapt to the evolving compliance environment. We remain confident in the long-term potential of Bitcoin. We will stay focused on innovation, operational excellence, and reinforcing our role as an important player in the Bitcoin ecosystem. This concludes my prepared remarks. Thank you, everyone. I will now turn the call over to our CFO, James. Thank you.

Thank you, Zhang Ji, and good day, everyone. This is James, CFO of Canaan. I'm very glad to share our Q4 financial results with you today. As Zhang Ji stated at the start of the call, in quarter 4 last year, the trends of Bitcoin price and total network hash rate were relatively favorable for miners. We fully leveraged this positive market trend, leading to increased sales and the delivery of mining machines and to the growth in energized hash rate of our own mining operations. Let me give a quick summary of our financial performance for the fourth quarter. First, the computing power sold reached 9.1 million terahash per second, breaking our previous record. Total revenue reached approximately USD 89 million, beating our guidance of USD 80 million and resulting in year-over-year growth of approximately 81%, as well as the highest quarterly revenue in the past 2 years. Secondly, through the increased deployment in our mining operations, our mining revenue reached USD 15.3 million in this quarter and increased approximately 313% year-over-year, exceeding the pre-halving level in 2024. We mined 186 Bitcoins in this quarter, an increase of 84% year-over-year. As we've reported in our monthly production updates in quarter 1, we continued the trajectory of accumulating more Bitcoins. At the end of February, our total Bitcoin holding reached 1,355. We are steadily advancing towards our target of achieving 10 EH/s per second in North America and 15 EH/s per second globally by the middle of 2025. As announced today, we further expanded our North American mining capacity by signing agreements with 2 new partners in Pennsylvania and Texas, which will add 4.7 EH/s per second to our North American mining capacity. This will bring our global deployed mining capacity to have the potential to grow to 11 EH/s per second with 6 EH/s per second in North America. Next, we completed the upgrade cycle of the product portfolio of our mining machines, and the A15 series is now our main sellable product, leading to USD 73 million of machine sales revenue. More than 17,000 units of the A15 series were delivered in quarter 4, of which over 80% came from the large orders of public mining companies, including Cipher, CleanSpark, and HIVE. These results demonstrate that our North American focus, which began just over 2 years ago, is having a positive impact on our business. On the consumer front, our first home series consumer product, the Avalon Nano 3, continued to gain adoption by individual users globally. In 2024, the total orders for Avalon Nano 3 reached approximately 24,000 units and contributed $2.7 million to the full-year revenue. At the beginning of 2025, our Avalon Home series was further expanded. To date, the orders for the new products reached more than 12,000 units valued at USD 5 million. Finally, thanks to the significant narrowing of gross loss and the upward Bitcoin price, we achieved a quarterly adjusted EBITDA of $19 million for the first time since the start of the market downturn 2 years ago. Additionally, our cash flow from production and operations turned positive, contributing $17 million in this quarter, strengthening our cash balance, and allowing us to end the year with $96 million. Moving to our profit and loss for the quarter. Total revenue was approximately $89 million. As I mentioned earlier, mining revenue contributed $15.3 million in Q4, an outstanding year-over-year increase of 313%. We mined 186 Bitcoins in the quarter, a year-over-year increase of 84%. This increase was primarily driven by more computing power installed for our mining business, which reached 5.4 EH/s at the end of this quarter, resulting in a historical high. As announced today, we have recently expanded our North American mining footprint by 4.7 EH/s per second through 2 new projects in Pennsylvania and Texas. When fully operational, these additions will bring our worldwide mining capacity the potential to grow to around 11 EH/s per second, with 6 EH/s per second in North America. We will continue to disclose the operational progress in our monthly mining report. Now, turning to product revenue. Revenue from machine sales was approximately $73 million, an increase of approximately 64% year-over-year. We delivered a total computing power sold of 9.1 million terahash per second, representing a year-over-year increase of 66% and marking a record high quarterly sales volume. The average selling price, or we say ASP, remained stable at $8.1 per terahash per second compared to $8.2 per terahash per second in the same quarter of 2023. With the destocking of our A13 and A14 series nearly completed, we expect the average selling price to return to a more reasonable level in the first quarter of 2025. More than 17,000 A15 mining rigs were delivered in Q4, of which nearly 80% were delivered to public mining companies in North America, including Cipher, CleanSpark, and HIVE. Turning to the revenue from our Avalon Home series. By the end of 2024, we received orders for approximately 24,000 units of our Nano products, reaching customers in 80 countries and contributing revenue of approximately $2.7 million since their launch. Our newly launched Avalon Mini and Avalon Q have also attracted significant attention to this consumer product line. To date, we have received orders for more than 12,000 units of this new Avalon Home product, with a total amount of $5 million. For our mining machine costs, we accrued $13.6 million for an inventory write-down in the quarter. With the successful clearance of the prior generation product inventory, inventory write-down decreased by 76% year-over-year. The noncash write-down is made on the U.S. GAAP rules, impacting our gross profit, but with no impact on our cash balance. Excluding the above write-down, our mining machine sales would have broken even for this quarter. Leveraging the rising Bitcoin price and our efficient management of mining operating costs, our direct mining margin improved to 42%. Please note that the direct mining margin is calculated by subtracting our mining operation cost for energy and hosting from our mining revenues without the depreciation of deployed machines. Demonstrating our robust top line growth and our effective cost optimization measures, our gross loss for the fourth quarter was $6.4 million, a substantial reduction of 88% year-over-year. Turning to the expenses. Our operating expenses totaled $49.3 million, increasing $10.1 million year-over-year. This rise was primarily driven by our global business expansion and our risk mitigation activities, including staff costs, professional service fees, and R&D expenditures. We recorded $9.3 million for one-off expenses in the quarter, including share-based compensation cancellation, onetime R&D expenditures, and consulting fees. If this one-off expense were excluded, operating expenses would have been $39.8 million, remaining stable year-over-year. As mentioned, in quarter 1 of 2024, we chose to adopt the FASB new accounting rules on cryptocurrency assets on January 1, 2024. These new rules allow cryptocurrencies to be carried at their fair market value. In the fourth quarter, the price of Bitcoin increased to around $95,000 at the year-end of 2024 versus around $63,500 on September 30, 2024, contributing to an aggregate gain on crypto assets of $39 million during this quarter. All things above considered, our adjusted EBITDA was a gain of $19.3 million, mainly driven by the solid revenue growth and significantly narrowed gross loss as well as the upward Bitcoin price. This is the first time we have achieved quarterly EBITDA profitability since quarter 1 of 2023. In quarter 4 of 2024, we recorded a valuation allowance against our deferred tax assets as well as an unrecognized tax benefit liability. This follows the conservatism principle in accounting and resulted in a noncash income tax expense of $85 million. When we generate profits in the future, these deferred tax assets can still be used to offset our income tax expenses. In the quarter, we issued 30,000 Series A1 preferred shares with gross proceeds of $30 million. The third tranche of the Series A preferred shares issued in quarter 3 of 2024 was still recognized as a convertible liability at fair value as of the year-end. This financing incurred an excess of fair value over proceeds received and fair value changes. This noncash accounting treatment hit our Q4 bottom line for a total of $4.8 million. In order to represent our performance more accurately and more comparably, we excluded the impact of this accounting treatment from our non-GAAP measures. Turning to our balance sheet and cash flow. At the end of quarter 4, we held cash of $96 million on our balance sheet, an increase of around 34% compared to the end of Q3. During the quarter, our cash flow from production and operations turned positive, generating $17 million. A cash outflow of $71 million for production and operation was offset by a cash inflow of $82 million from sales and $6 million from export VAT refunds. In quarter 4, we received $30 million from our preferred share financing, and we paid $22 million to secure our wafer supply. With stable cash flow and the forthcoming completion, we have leveraged strategic financing tools to secure product capacity through prepayments. These activities will allow us to capitalize on the upcoming bull market while preparing for long-term mining demand despite the short-term volatility of Bitcoin prices. At the end of 2024, we entered into an ATM agreement from the end of 2024 to February 19, 2025. We utilized the ATM for fundraising with net proceeds of $42.5 million at an average price of $2.08 per ADS. We did not utilize the ATM after February 19, 2025. In early March, we closed the first tranche of new Series A1 preferred shares with net proceeds of $99.7 million. Total proceeds of approximately $145 million will be used to fund the expansion of our production scale and manufacturing or investing in our mining operation in North America. At the end of Q4, the aggregate balance of inventory, prepayments, and other current assets was $185.5 million, a decline of $40.6 million compared to the end of Q3. This decline was driven by the clearance of prior-generation product inventory and upgraded inventory composition. Now, turning to our Bitcoin assets. Bitcoin held as our own holding assets increased in this quarter, reaching a record high of 1,293 Bitcoins as of December 31. This is 81 more than 112 at the end of last quarter. On December 31, 2024, the fair market value of our owned Bitcoins totaled around $123 million, and our hold-on strategy gain was approximately $66 million higher than the original value of the Bitcoins that we gained from mining or other operations. As of February 28, our total Bitcoin holdings increased to 1,355, as already disclosed. By the end of 2024, we pledged 600 Bitcoins for secured loans with an aggregate carrying value of $24 million, which we believe is a reasonable interest level. The secured loans enable additional liquidity for our production expansion and operations. We also transferred 100 Bitcoins into a fixed-term product with a guaranteed minimum annual return. In the future, as part of our hold-on strategy, we will explore more ways to increase capital liquidity through our owned crypto assets. Please note that Bitcoin pledged or transferred into fixed-term products are recognized as cryptocurrency receivables in our balance sheet. And classification between current and noncurrent assets is consistent with the period of corresponding secured loans of fixed-term products. In the first quarter of 2025, as the CEO mentioned, we anticipate revenue of $75 million due to the seasonality of our business, compounded by the uncertain economic environment as well as the volatility of Bitcoin prices. Based on our current visibility on orders and wafer capacity, we anticipate a revenue of $120 million to $150 million for the second quarter of 2025. This concludes our prepared remarks. We are now open for questions.

[Operator Instructions] We will now take our first question. This is from the line of Nick Giles from B. Riley Securities.

Congrats on the progress here more recently. My first question is, can you give us an update on any site acquisition activity or how you're approaching procuring power infrastructure in the current environment? Curious if you've observed any changes or challenges here in 2025, as demand for power assets remains very strong.

Let me take those questions. I think we have been actively looking for sites in energy-rich regions across North America. Our cooperation model is very flexible. We can do like fixed-rate hosting, joint mining with profit sharing, or even acquire and build our own sites. Right now, the site resources we have secured are high, actually more than enough to support our current deployment plan, which is a target of 10 EH/s in the first half of the year. Since shifting our focus to the U.S. last year, we have seen more opportunities than we originally expected. Recently, due to some changes in U.S. policy, it seems that new power generation capacity may become available in the near future, which would benefit us even more. We have also noticed that some institutional players in North America are thinking about allocating part of their power resources to AI or HPC hosting. We truly hope those efforts succeed. AI and HPC business can bring more predictable cash flow for our customers, which in turn, could attract more institutional interest and raise the overall valuation of their operations, which would help them broaden their business, including having more power and capital available than they can go to Bitcoin mining to consume their electricity. So well, Bitcoin mining has much lower CapEx needs than AI or HPC, it consumes far more energy. On the other hand, AI and HPC offer very stable cash flows. That's why we believe these 2 sectors are actually complementary in the mid to long-term. I hope we can see both Bitcoin mining and AI applications go side by side going forward.

My next question is just on overall demand. Obviously, your 2Q guidance is a nice improvement from the 1Q guidance. And so, I was wondering if you could give us a sense of what the second half could look like. Is demand still improving? Your annual guidance does imply a stronger second half. So, wondering if you could comment on how you're seeing demand today, particularly on the pricing side, and how that translates to your annual guidance.

I will say something, and maybe James will add some numbers to your question. I think look at our actual sales performance in Q4, it was quite good. We sold 9.1 million terahash of computing power and have already achieved our best historical results. It includes the deliveries to North American listed companies like Cipher, CleanSpark, and HIVE. Yes, this reflects the success we have had in expanding our presence in the North American market. In fact, customer order volumes at the end of the year were driven by the rise in Bitcoin price starting in November. So yes, and as a result, we clear our inventory, then we start to build our new inventory. Most of the wafers still come out from Q2 to Q3. The machines will come out major in Q2 and Q3. So we can foresee that from Q2, the number will be much larger than Q1. For the second half, I think the key may be the bitcoin price or the economic environment, which is really hard for us to predict today. But what I can say is if the bitcoin price like we don't need another historical high, maybe only like 100,000 is very significant for us to achieve our annual target.

I will add some color, Nick. I think from our guidance on Q1, it has already shown in 2025, quarter 1, the revenue is more than double of last year same quarter. So the demand was strong. But as we both know, our orders and revenue are highly correlated with Bitcoin prices. So recently, the turbulence in Bitcoin prices has definitely had an impact on the speed of accumulating more contract orders and also the trajectory of the price increases. But it looks like our customer orders are still quite strong from the market. We also have our North American mining sites building up on target. That's why I think we are still confident to work towards the full-year target. I think we are working on that, and Angie just shared the capacity perspective. But, from the demand perspective in quarter 3 or quarter 4, we definitely need the macro environment to turn more positive and for the Bitcoin price to climb to a new high. And then, at that time, the demand could be strong, and the price trajectory could be favorable to us, and then we will try to maximize our revenue for the full year. Currently, I think we are still sticking to the annual target, and we are working towards this target. Thank you, Nick.

We'll now take our next question. This is from the line of Kevin Cassidy from Rosenblatt Securities.

Congratulations also on the great progress. Just as we're talking about the full-year revenue, can you say what your ASIC orders are with your foundry partner that may not be accounted for by customer orders just yet? How far out are your orders with the foundry versus what's in your backlog?

I think we just come from the bear market. In bear markets, we typically produce based on sales orders, which helps the company control the risks of the product inventory. Currently, we are just out of the bear market and at the early stage of the full market for mining machines. And maybe this month, we will be in the turbulence of the full market. So, we are gradually increasing our production, but we have not placed orders at maximum capacity. When all the wafers, we must consider many factors, including but not limited to our cost, how tight the supply chain is, and what the minus expected profitability about hash price and the competitive landscape in the market for the next, say, 6 to 12 months. So, for commercial confidentiality reasons, we are not disclosing the exact number for how many wafers we have already ordered at this time. But it appears that the machine production will continue through Q3. If the market demand doesn't increase significantly, it should be sufficient to meet both ourselves and our own mining operations. If demand increases, we still have time. So we have room to place additional orders to a few tens of percentage of our capacity to fulfill the additional orders. And if the demand weakens, we can increase allocation to the number to our own mining operations. So I think for the current capacity in the production line, it shouldn't be a major issue for the inventory.

You announced that the A16 ASIC is 300 terahash per second. That's a great improvement. Is there anything more than besides going to the better or lower process node, any architectural changes that you're making to the ASIC that's giving you this improvement?

Yes. I think the A16 is indeed using the very cutting-edge technology like the GA technology and even smaller transistor sizes compared to A15. But it's not simply a straightforward part of our previous generation. I think if you look back into history, every generation of our products has often achieved much more significant performance improvements, and many different generations of our many machines are in one node process or even one node process upgrade or even using the same process. These performances are far greater than the percentage improvements indicated by wafer manufacturers during each process node iteration. This is because our designers are doing their innovations and many technical advancements ongoing. Also, to this stage, besides process and design, we are already at the DTCO, design technology coization stage to increase the performance even more. Perhaps you can notice that each generation of our product shows about 20% to 30% of the improvements, but what I must say is there isn't a magic number there to give us such a big improvement. It's more likely that we have like 50, 80 individual technical improvements every contribute like less than 0.5% to lead to collectively together to lead to the overall result. So, as we move to advanced nodes, each new generation of technologies also needs time to improve. And I think the performance of ASICs and the GPUs on the same new process may also vary significantly. As a design company, we need to spend a lot of time in the early stages now to reduce these uncertainties and minimize the risks to the tape-out. So maybe your question is why there are many newcomers even when they try to have key technical talents from the industry but still struggle to produce competitive products.

Just one other is the 300 terahash per second for traditional air cooled; do you have an estimate of what it would be for immersion or just liquid cooled?

I think that, for today, the emerging version of the machines is more common in the market, not constrained to the U.S. market, many different market. We see that we already observed this. So Yes. I think maybe the reason is that mining machines are now built with a much higher energy density and need better energy efficiency. This will bring more strict environment control needs. So, liquid cooling can help to manage heat more effectively. Another reason maybe the mining hardware now must stay in use longer than old days. So, this makes it worthwhile to invest more in the site set up to improve the scalability and uptime.

We'll now take our next question. This is from the line of Kevin Dede from HCW.

I think, firstly, I'd just like to congratulate you on managing your supply line and deliveries to the U.S. in light of the regulatory changes. I just thought that was very savvy of you, gentlemen and the team. First, I was hoping you could offer a little more detail on how customer service has changed over the years. I've been following you for a long time. And I'd just like to understand what you think you're doing better now in the U.S. and winning customers.

I think that in recent years, as Bitcoin mining has become more global, the company has to provide maintenance and service in a way that's closer to our customers. In addition to our standard 1-year warranty for mining recorders, we now have 26 service stations for spare parts warehouses worldwide. This helped shorten the parts delivery cycle time and allowed us to provide more time maintenance and parts replacement services for our customers. We also adjust the location of our service plans based on customer distribution for future improvement of service efficiency. Additionally, based on customer feedback, our products have shown strong operational stability, particularly in very high temperature and humid environments, where they continue to run effectively and very reliably. Also, the return of the repair is actually very low now, especially for the A15 models. We also offer on-site customized mining solutions based on customers' specific needs, such as emerging cooling and co-development for some systems. It ensures that our products and services can be closely aligned and adaptable to our customers' diverse needs, and we are receiving positive feedback. Thank you for recognizing our success, but we're still striving for improvement. We believe we can do even better, and we aim to outperform 2 of our Chinese competitors in the following areas. First, we want to keep the failure rate of our products below the industry average, which comes from better design and high-quality control during the production process. This approach particularly benefits high labor cost regions like the U.S. Second, we aim to ensure our North American team and management in Singapore are one cohesive unit. Myself, along with our CFO and R&D and supply chain leaders, are all committed to ensuring high customer satisfaction. Finally, from the very beginning, we have been a company driven by the value of decentralization. Therefore, we remain open to all customer requests, evaluating them only based on whether they make business sense and benefit the industry development. We don't want to limit customers' options proactively.

And can you talk a little bit about your North American customer pipeline and what the sales funnel looks like? The press release alluded to a new customer for the A15XP. Can you talk about that a little bit and maybe give us a little more comfort regarding where or how you see the second half coming together?

I think that last year, we actively expanded into the North American market and secured several major clients like Cipher, Hive, and CleanSpark. North American customers' orders now represent about 40% of our product sales revenue for the full year of 2024. In Q4, with the large-scale delivery of A15 orders to North America, sales from the region reached a historical high. And also we are in the second half of this year, our sales system now is increasing their global market coverage. Last year, we had nearly 200 professional offline customers. Following the launch of the Avalon Home series, we attracted a large number of retail customers, resulting in exponential growth in customer numbers. And in 2024, the total number of customers who have placed orders reached about 2,000. Now, the company continues to go globalized. We have established special teams in regions such as North America and South Asia, building a multilayer and comprehensive sales team that reaches customers worldwide. Currently, the number of new customers is substantial, and they generally have a strong competitive basis and high purchasing intention. However, I think large orders from new customers typically take some time to settle down. If the global political and economic situation, along with the Bitcoin price influenced by these sectors, continue to rise steadily, I think this process will certainly accelerate. So because in Q2 and Q3, the machines will come out majorly in Q2 and Q3. I think the capacity issue will happen, especially in Q1. I think, personally, I'm quite optimistic in Q2 and Q3.

Last question for me, Zhang Ji. You mentioned that the A16 is now taped out. I'm wondering when you start placing orders for wafers for those machines.

We started to place orders back to -- I think our first big batch was in October to November last year, and we had a second batch followed in December. So the wafers will continue to come out from Q1 this year, and we are in the ramp-up stage. The major part will come out in Q2 and Q3. Sorry, I think your question is about A16.

A16, right? Yes, no, no. Okay. Yes. I was just wondering about the takeout orders.

Okay. A16, we will send the machines to our customer I think it's in Q3. So we will follow the risk run and our first mass production batch at that time.

We'll now take our next question. This is from Joe Flynn from Compass Point Research & Trading.

I was hoping if you could just touch on the ramp of A15 in a little more detail. I was curious how much of the initial capacity was ultimately covered by North Americans' upfront deposits. And is there any balance on those going forward? And then, ultimately, what's the strategy of managing the liquidity risk given the 6-month lead times? And what approaches are you ultimately going to continue pursuing from that end?

I think for our 18 orders we can deliver most of them in May. Yes, they are preordered.

I'm sorry, I was talking about the upfront deposits from the North American customers as it relates to securing fab capacity.

I think it mostly comes from North America, yes. And most of them are prepaid for about 50%, payment.

So then what's ultimately kind of the approach to funding the additional fab capacity 6 months out as those start to roll off? Given the risk of the lead times?

I think the production for the machines today is about 5 months. But usually, our customers cannot wait for 5 months from the first payment to receive the machines. So, usually, we will keep our pipeline for about 3 to 4 months. That means our customers will pay the first down payment like 50% and then pay the other 50% before we ship the machines, or they receive the machines in 3 or 4 months, yes.

And most of my questions were asked, but maybe go in a different direction. Just curious if there is any portion of the R&D still dedicated to the risk 5 ASICs because there are a lot of reports out there that expect to see a huge increase in demand in China. I was wondering if anything came about the partnership with Alibaba a few years ago and if you guys ultimately have the opportunity to compete in that market at all.

Yes. I think it's a very sensitive question. I think, given the tightening regulation environment we've observed recently, especially from the middle of January, what I can say is we are staying very, very cautious about developing anything related to AI now. So yes, so that side, we are actively prepared for the future. Over the next 2 years, our plan is to invest more in building out our infrastructure in the U.S., which I already mentioned tonight. This is our way, I mean, personally, I think this is our way to hedge against the risk of missing out on the AI phase in the next maybe 2 or 3 years. If one day, AI and HPC become the mainstream computing use case, replacing mining, we want to be ready with the power and size and all infrastructures and technologies there. We are already in place, we will be very strong in a very strong position to pivot quickly into the new direction. This is what I can say today.

We'll now take our next question, this is from the line of Bill Papanastasiou from KBW.

Zhang Ji, may you share some color on your progress of reaching self-mining hashrate growth targets and how this may impact external sales of hardware? Can the internal production of mining equipment sustain self-mining growth targets while also ramping your external sales to grow market share?

Yes, at the end of February this year, our company operations, we have 7 active mining partnership programs globally with a total deployed hash rate of more than 6 exahash and the power is close to 6 exahash per second. In Q1, we further expanded 2 existing North American mining projects through upgrades to increase their capacity. So, we are keeping our target of 10 EH/s in North America in the first half of this year. In the first quarter, I think our North American self-mining business will be very close to 5 EH/s per second. Yes, because I think currently, the company's operation is very smooth, and we will build ourselves money, follow our cash flow, and keep our cash flow remain healthy. I think the U.S. government is implementing a series of policies supporting the cryptocurrency industry. The trend towards a Bitcoin market remains intact because despite short-term price fluctuations since the advanced process capacity required by our industry must be deployed in advance. We have chosen to use financing tools to fundraise to lock production capacity and fulfill our machine sales and money. For the how to expand our market share, we are NASDAQ listed Singapore company alongside the 2 private Chinese companies. Our market share is still small compared to our major main competitors. So for us, we have opportunities and we must focus on how to gain market share, then we can have much higher revenue. Historically, I think our major competitors have the industry-leading product and/or improved advantages in North America. They expand their influence with global Chinese… [Technical Difficulty]

[Operator Instructions]

Hi this is Leo Wang, Vice President of Capital Markets and Corporate Development. I think Zhang Ji is offline because of technical difficulties. We are more than happy to take your questions after this. I suggest we conclude this session.

Thank you. In that case, I will turn the call back over to the company for any closing remarks. If there are no further closing remarks, then we can conclude the call today. Thank you, everyone, for attending, and you may now disconnect.