Oleksandr Shatyrov
In what has seemingly become a rite of passage in the evolution of the industry over the last year, Bit Digital, Inc. (NASDAQ:BTBT) recently became yet another Bitcoin USD (BTC-USD) mining company to start investing in building data centers for “artificial intelligence.” Bit Digital is one of the public companies in the crypto ecosystem that I personally held for much of 2023. I have covered the company a handful of times for Seeking Alpha; most recently in late August when I called it a buy due to its pro-BTC and pro-Ethereum USD (ETH-USD) strategies.
Company Performance
Since my last BTBT article, we have received several months of additional production figures to assess. From a Bitcoin mining standpoint, Bit Digital has scaled exahash per second from 1.8 at the end of July to 2.5 EH/s at the end of December. Looking at average monthly production, Bit Digital’s Q4 figure of 142 was 6% ahead of its 134 average monthly BTC mined in Q3. Perhaps more important than the company’s production growth quarter over quarter, no public miner was more efficient than BTBT in the second half of last year by a notable mining metric:
Source: Company disclosures, Author’s calculations
Taking it down to the second decimal, Bit Digital edged out HIVE for the most BTC per EH/s in the six months ended 2023. That said, we did see BTBT showing weakness in this figure when isolating to December as there were eight companies that beat Bit Digital’s 67.3 BTC per EH/s in the month. This is of slight concern given the notable spike in transaction fees that Bitcoin miners enjoyed in December.
BTC Transaction Fees (IntoTheBlock, Author’s Chart)
At 21.5% of total rewards for the month, elevated transaction fees from protocols like Ordinals and BRC-20 token swaps have been a significant tailwind for Bitcoin miners and can help mitigate some impact from the upcoming halving if they remain at these levels.
As has been previously noted in my past articles, Bit Digital’s approach to a crypto flywheel involves selling BTC from mining, exchanging it for Ethereum, and then staking the ETH both natively and through liquid staking protocols for additional blockchain rewards.
Bit Digital Crypto Treasury (Bit Digital, Author’s Chart)
As a result of this treasury approach, Bit Digital’s balance sheet has become more dependent on Ethereum than Bitcoin over the last year with ETH going from 40.1% of Bit Digital’s crypto treasury at the end of December 2022 to 59.3% at the end of December 2023. While Bit Digital’s crypto holdings have substantially increased over the last year from $26.2 million to $66.9 million, ETH has significantly underperformed BTC over the last 12 months so it’s fair to wonder if this strategy has actually been working so far:
While BTC offers no native yield as a proof of work coin, Bit Digital’s blended APY from ETH staking over the last six months has averaged about 4.2%. While small at just a quarter of a million in Q3 staking rewards revenue, there are positive signs from recent trends in ETH staking:
ETH Real Yield (Staking Rewards)
When transaction fees spiked on Ethereum’s main layer toward the end of 2023, it reduced new ETH issuance to the point of being deflationary once again. So while Bit Digital’s blended APY from staking has trended down over the last few months, the “real yield” story on the broad network has been positive over the last 90 days having improved by 92 basis points. With the approval of a Bitcoin ETF now in the rearview mirror and Ethereum now on the radar for firms like BlackRock, Inc. (BLK), I’ve hypothesized that ETH has quite a bit of catching up to do and Bit Digital would be a beneficiary of that.
And Then There’s AI
Of course, the biggest immediate driver of incremental revenue for Bit Digital might not be crypto-related at all. In recent months, the company has announced the purchase of servers containing Nvidia HGX H100 GPUs. The intention for these servers and cards is to build out the company’s Bit Digital AI product. According to Bit Digital, these servers have already begun to arrive at the company’s Iceland datacenter where they will be installed and tested before being put into service.
The exact details aren’t totally clear, so I’m personally not getting terribly excited about these numbers until we start seeing what I feel would be more concrete guidance from company leadership. But we know so far that Bit Digital has a client contract for these AI services already in hand. As part of this arrangement, Bit Digital has received a three-month prepayment for an undisclosed amount and a contract that could possibly generate $50 million in annualized revenue over the next three years.
| Monthly Revenue at the BTC Price of | ||||
| Monthly Mined | $35,000 | $40,000 | $45,000 | $50,000 |
|---|---|---|---|---|
| 120 BTC | $4,200,000 | $4,800,000 | $5,400,000 | $6,000,000 |
| 130 BTC | $4,550,000 | $5,200,000 | $5,850,000 | $6,500,000 |
| 140 BTC | $4,900,000 | $5,600,000 | $6,300,000 | $7,000,000 |
Source: Author estimate table
If we assume the company can achieve its goal of 6.0 EH/s at some point this year and maintain a stable monthly BTC production of somewhere between 120-140 BTC following the halving, Bit Digital stands to generate as much as $21 million in mining revenue per quarter. That, coupled with an annualized estimate of $50 million in revenue from Bit Digital AI and another $1 million in ETH rewards, puts Bit Digital at $135 million in annual revenue by the end of 2024. At a $265 market cap, BTBT is trading at roughly 2x forward sales. This would be below the information technology sector median of 2.8. It also assumes a Bitcoin price of just $50k which I believe is actually on the very low end for 2024.
Risks
All this said, Bit Digital is still a highly risky investment. The company has three distinctly different revenue sources but each of them carries quite a bit of uncertainty. Bitcoin’s next block reward halving is about three months away. When that happens, revenue from the company’s primary segment to this point will be cut in half virtually overnight all else being equal. Ethereum staking is a very interesting business and one that I happen to like, but it’s still a small portion of the company’s overall revenue.
Then there’s artificial intelligence. For all the excitement about AI and Bitcoin miners pivoting to data center models, this is not an area where I have a tremendous amount of confidence for BTC miners broadly. There are already at least five publicly traded miners that are trying to serve this market and if the margins on this business are as good as what is being presented, the competition is likely to get more intense. Furthermore, ASICs are not used for generative AI so pivoting to data center solutions will be more capital-intensive than Bit Digital’s proof of stake expansion.
Summary
If you’re looking for a Bitcoin mining pure play, Bit Digital probably isn’t the right fit. That said, the company is trying to scale production to 6.0 EH/s this year. Assuming it can do so, the company should be able to maintain current levels of mining production without a drastic move-up in transaction fees. This EH/s growth will obviously come with an increased expense to energize that exahash but if Bit Digital AI delivers on some of these ambitious revenue figures, the increased mining expense could theoretically be offset but the higher margin AI segment.
Additionally, the time may be right for Bit Digital’s Ethereum staking strategy to start paying off both from staking rewards revenue and from a capital appreciation standpoint. With the Bitcoin ETF in the rearview mirror and the halving quickly approaching, attention seems to now be turning to expectations for an Ethereum ETF in the US market. I think ETH still has quite a bit of upside and as an ETH staker with over 17k ETH in treasury, Bit Digital would be a big beneficiary.
