Marketing vs. reality:
An (institutional) investor’s take on the leaked Bitmain financials
The anticipated IPO of Chinese cryptocurrency mining equipment manufacturer and miner Bitmain is currently attracting widespread attention, not only in the alternative media favored by crypto folks, but also traditional media, and, as rumors have it, among some of the world’s most prestigious investing firms, including e.g. Softbank and Temasek (albeit both denied involvement recently).
This is not surprising, given the IPO would be a rare chance to invest in a significantly sized IPO ($3 billion offering size was mentioned) in a hot growth sector — at a time when other promising companies often either get taken out early on or do not seem to want to IPO at all (think SpaceX, Uber, etc.). A rumored valuation range of up to $40–50 billion does not appear unreasonable if one takes the company’s statements of $10 billion sales and $2 billion in profits in 2018 at face value. In short, the FOMO is understandable.
But … does the hype stand up to scrutiny? Recently leaked data about Bitmain’s financials now helps us to construct ever so imperfect estimates of our own. Let us combine the leaked data with publicly known data points and see where we get to. We start by forecasting expected full year 2018 revenues.
The most obvious recent change in the company’s environment has been the sharp decline in cryptocurrency prices. Coupled with the fact that (BTC) difficulty has been steadily increasing, this meant mining equipment prices had to drop significantly, in order to maintain an ROI that would induce anybody to buy mining equipment — and this is precisely what has been happening. The below chart shows the Amazon retail price development of Bitmain’s flagship Antminer S9 (in its 13.5TH version — note other S9 versions, as indeed other miners offered by Bitmain, have suffered similar price declines).
· For 1Q18, eye-balling the chart indicates an average retail price for this version of the S9 of around US$3000, which compares to the leaked 1Q18 unit price for the S9 of US$1719 — the difference likely being significant bulk discounts for wholesale.
· For 2Q18, eyeballing the chart gives us a retail price of approx. US$1100 (and approx. US$650 if making similar assumptions on the impact of wholesale sales).
· For 3Q onwards, we can effectively “mark-to-market” and the retail price would be US$480. Bitmain’s own website (in both English and Chinese) also shows a price (as of 17-Aug-18) of US$485 for its latest (14.5 TH) Antminer S9 version. Including the impact of wholesale sales, the current price might well be below US$300.
So much about prices. Sales quantities are harder to guess. Given the current breakeven electricity cost for Bitcoin mining is now significantly lower (around $0.10/kWh) as it used to be, this should clearly have an impact on demand, as e.g. a retail user in a typical European country with $0.20/kWh cost cannot profitably mine BTC at current prices anymore. (Imperfect) read-acrosses do also not bode well: e.g. Nvidia, in its 2Q18 results , noted that GPU revenue from crypto came in at US$18 million, vs. an expectation of US$100 million, and US$289 million in Q1. Having said that, Bitmain may, for now, continue to have orders from large mining projects that have access to low electricity prices and already have secured funding. This may, by the way, include demand by Bitmain’s own mining division — to the extent Bitmain has inventory anyway and can deploy it (read: have sufficient low-electricity-cost locations — and note the recent announcement about a mining facility in Texas in this context), their own mining division may cover part or all of any unit sales shortfall to third parties. And inventory Bitmain does have: according to the leaked data, some US$1.24 billion at the end of 1Q18 — to put this in context: this is about 135 days in inventory and, if (untrue, but just for illustration) this inventory was all finished goods inventory of Antminer S9s, some 2.5 million units.
In any event, trying to get to revenues:
· On 1Q18 run rates, one could theoretically get to the company’s FY18 revenue number of US$10 billion by assuming some 20% quarter-on-quarter growth for the rest of the year, which would seem a modest assumption compared to likely recent actual growth.
· However, even heroically assuming flat S9 unit numbers (716k in 1Q18), but adjusting for Antminer S9 prices as per the above, and making similar assumptions for non-S9 revenues (32% of revenue in 1Q18, and substantially other mining rigs, own mining, and mining pool revenue, none of which is likely to have fared well recently), FY18 revenue comes out to approx. $3 billion rather than $10 billion, which would constitute a somewhat dramatic “miss”, to put it mildly.
Margins, of course, are brutally impacted by the price erosion:
· In 1Q18, the leaked data shows a gross margin for Antminer S9s of 70% (implying unit COGS of US$517)
· Even assuming a (fair) continuing decline of unit COGS, the price erosion assumed above to US$650 in 2Q and US$300 from 3Q would imply a gross margin of 22% in 2Q, and a solidly negative gross margin from 3Q onwards.
This is, of course, unless Bitmain decides, or is compelled to, write down its significant inventory — which may well be on the cards given the price declines so far (and, as time goes on, there is obsolescence risk, too, causing further price erosion).
The revenue and margin picture on non-S9 miners is likely even bleaker, given that altcoins have mostly sold off even more than Bitcoin. Combining all these datapoints, and assuming some R&D spend (US$13 million in 1Q18) as well as other total operating expenses running at a similar level to that at Canaan Creative (approx. 20% of revenues) leads us to scenarios in which Bitmain would hardly make any profit in 2018 — and this is before a potential huge write-off on coin holdings, discussed below, which may or may not run through the income statement.
For completeness, it should be noted that we made the entire analysis above without knowledge of Bitmain’s relevant accounting policies, notably regarding revenue recognition and inventory (and as e.g. internet posts at the end of 2017 frequently mentioned order delays of around 2 months, recognizing revenue on order vs. on delivery does make a difference). However, the analyses should still directionally be correct.
Let us change tack and speak of Bitmain’s mining division and coin holdings.
The leaked data shows Bitmain had 91,250 mining rigs running at the end of 1Q18 (and an implied average of 81,640 during the quarter). Let us simplistically assume all rigs were mining Bitcoin Cash (BCH). Using historical difficulty during 1Q18 and assuming an average electricity cost of US$0.04/kWh (implied from leaked data), Bitmain’s own mining rigs would have produced over 80k BCH. Now, if you were a pure mining company and assuming you still think about your financial performance and exposures in fiat currency terms (as, it seems, is also Bitmain’s case — after all, they gave out those estimated FY18 numbers in US$, not in BTC or even BCH), the expected normal course of business would seem to sell your mined coins rather than take active bets (unless you were substantially operationally hedged, e.g. if you could pay your suppliers in coins, but to our knowledge this is not the case).
As we know from the leaked data, that is not what Bitmain has done. Their BCH balance kept on growing — e.g. from 841,866 BCH at the end of FY17 to 1,021,316 BCH at the end of 1Q18. As per the above, 80k of this approx. 180k increase in BCH may have come from its own mining. With regard to the remaining 100k increase in BCH, we note that Bitmain’s BTC holdings decreased from 36,877 (end of FY17) to 22,082 (end of 1Q18), i.e. by 14,795. Assuming an average BCH/BTC rate of 0.15 during 1Q18, we get to pretty much exactly the missing 100k in BCH by assuming that Bitmain sold off their BTC for BCH. In other words: if this is indeed the case, not only has Bitmain not hedged/sold their production output, but they have also engaged in active coin picking (biased towards BCH, which is no surprise given their known opinions). A humble suggestion to Bitmain: if you want to take bets like this, open an asset management or prop trading division and hire professionals.
Of course, that prop trading has not been very successful. In fiat terms, the value of Bitmain’s coin portfolio as at the end of 2017 would have decreased from US$2.1 billion then to approx. US$800 million now (of course, Bitmain will have mined further coins since the end of 1Q18 and may also have engaged in continued coin trading, e.g. perhaps swapping more BTC for BCH). This in itself is a whopping US$1.3bn write-down Bitmain has experienced. Just the BTC-to-BCH coin trade that we think took place during 1Q18 (swapping approx. 15k BTC for approx. 100k BCH) would have produced over US$40 million of paper loss (BCH/BTC is down from approx. 0.20 at the end of 2017 to currently approx. 0.08; in general, altcoins are down vs. BTC, whose dominance has increased from approx. 40% at the end of 2017 to currently over 50%). This is all assuming market prices and not even giving any liquidity discount on Bitmain’s significant BCH holding.
In summary, this is a company whose fortunes unfortunately have significantly changed since the end of 2017. Building up significant inventory at end of 2017 and beginning of 2018, reducing R&D spend at a time when competition is catching up, and a huge bet on a relatively illiquid altcoin (BCH) are just a few examples of questionable management decisions. As a result, the company, unless circumstances change, will now likely produce a topline in the vicinity of US$3 billion (vs. a communicated US$10 billion) for FY18 and be operating at negative gross margins.
Of course, we could analyze further elements of Bitmain’s story, e.g. its other mining products, AI plans, or potential contingent liabilities regarding order commitments, but shall stop here and leave a few closing thoughts:
· Bitmain is clearly a key company of the crypto ecosystem and enjoys an enviable market position with its popular miners, however the company has failed to produce any new chip over the last 2 years and its smaller competitors Canaan and Ebang (who are also filing for IPO) have bypassed it on innovation coming up with new and more efficient 7 nm chips.
· The company is currently dependent on a key exogenous variable (cryptocurrency prices) and it should be noted that company management has been clearly happy to take very aggressive bets around this variable, both in terms of the inventory of its core product as well as its coin holdings. Recent large bets have so far gone awry.
· Depending on whether and when cryptocurrency prices rebound, things may still work out well for Bitmain. However, the company is about to ask (private and, it seems, soon also public) investors for significant incremental cash. Given the company’s recent track record, investors should carefully reflect whether this is the right time to give the company more money and, if so, at what valuation. Clearly at our estimated 2018 year Revenues of $3bln and Profits of zero the rumored latest round valuation of $14 bln is extremely high.
· Potential investors should most certainly ask Bitmain about the use of proceeds of their private and public fundraising. In a worst case scenario, Bitmain may be stuck with a huge inventory facing obsolescence, potential additional order commitments requiring cash, and a large and fairly illiquid holding in BCH (which, as e.g. this article claims, it may have started to try to liquidate). So, investors will probably want to understand whether they are (1) giving the company cash for profitable investment opportunities or (2) effectively giving the company cash as a lifeline!
· Furthermore, in any event, if the company’s financial performance and hence valuation ambition (as noted above) are dependent on a significant rebound in cryptocurrency prices (in order for an Antminer S9 to have the same profitability to an end user that it had in 1Q18, the Bitcoin price would need to go to approx. US$24k), investors should also analyze whether they may not realize a better risk-adjusted return by simply investing outright in cryptocurrency.