I do not know any non-disclosed info but presenting the other side. Hindenburg went after Block last year:
Block Stock Bounces Back After Refuting Hindenburg Report Findings | PYMNTS.com.
Andrew Left of Citron fame went to prison (my bad, I thought Mr. Left had already been found guilty and sentenced. @Smorgasbord1 corrected me in a post below), Hindenberg uses the same business model that Citron used. They write lengthy damning reports of some business or another and cash in on short positions when the stock takes a dive.
I don’t understand where the line is drawn between a document full of misrepresentations and innuendo verses a document worthy of putting the author in jail due to injurious slander. Or if it is even possible to slander a company rather than an individual.
I don’t have a problem with people seeing what they perceive as a weakness in an otherwise successful company and publishing a document that sets for their analysis. Where I see a problem is when the weakness and failings identified by a short seller are simply untrue for the most part or indicative of problems that don’t actually exist which results in damaging both the company and the stockholders.
On the negative front Barclay’s has joined the short choir. Basically mirroring the short report but with perhaps more details (sources for this info was not given). The primary details is as follows:
I’m“MCI has been giving away DLC components for free to match or price below the air-cooled rack pricing provided by Dell (NYSE:DELL),” said analysts at Barclays.
If memory serves, SMCI talked about this at the earnings call. They were basically seeding the market. Without Blackwell, liquid cooling is still optional. So SMCI is seeding the market to gain share, firm up the supply chain, and be ready to capitalize on Blackwell.
The report also specifies that Blackwell GB200 - concern over SMCI’s share in regard. Yet, GB200 won’t go high volume until 2025. Won’t go low volume until next quarter. Barclays doesn’t specify anyone substituting for SMCI. And in fact has no info as to GB200 vendors.
GB200 must be liquid cooled.
So some details (negative) but also some piling on going on as well.
Whoa there:
- Andrew Left has been charged with fraud by the SEC, but is not yet convicted. Heck, I don’t believe he’s even been arraigned yet.
- It’s wrong to group all short sellers together. There are viable and legal avenues for short selling, and for making statements about one’s short position.
- I know of no charges against Hindenburg at this time. Indeed, their well promoted short report and position on Nikola was eventually shown to contain true and fair allegations, resulting in Nikola’s CEO going to prison, not the other way around.
Do you believe the Hindenburg report is entirely wrong? Or even mostly wrong? Checking the re-hirings of people who resigned over the prior faudulent accounting should be easy enough, as is establishing who owns which pieces of which companies that are Super Micro suppliers. The charges for whether SMCI products ended up in Russia are more complex, but given the company itself has acknowledged it’s working through a review of “internal controls,” there might be something to them.
At any rate, the charges against Left include misrepresenting his actual financial position in the stock, trades he had made, fabricating invoices and hiding payments to a hedge fund, etc. Not that his analysis of a company was flawed (which isn’t a crime even if true).
As to Hindenburg. Seems they also are not entirely “unbullish” towards SMCI. This I had not heard about until now:
Despite the negative outlook, Hindenburg did admit a bull case for Super Micro, which would entail “meteoric growth” in response to demand for AI chips. The short-selling firm cited Bloomberg, where analysts forecast 87% revenue growth in 2025—or at least, before the report came out.
Maybe everyone is correct here. The next great catalyst is the mainstreaming of Nvidia’s Blackwell chips that like its current chips did, is about to light up the market again. These systems need to be liquid cooled.
It was and remains in the original Hindenburg report:
Bull Case: Growth Runway Driven By AI Server Market Demand, Close Relationship With Nvidia, And Proprietary Liquid Cooling Technology
The bull case for Super Micro is that it offers exposure to meteoric growth in demand for AI chips.
which concludes:
Conclusion: A Serial Recidivist
All told, we believe Super Micro is a classic case of recidivism. Its actions suggest that it hasn’t changed from its checkered past regarding its revenue recognition and accounting practices.
While the company gained an initial lead in supporting the AI industry, it seems to be cutting corners on its accounting, sanctions compliance, and product quality, all while more credible competitors eat away at its margins and market share.
I’ve now read two short reports on SMCI. One is from June when an “independent” director who was on the audit committee left and NASDAQ sent a letter stating to fix this quickly. SMCI did so by appointing 3 “independent” directors into the audit committee.
There are two things to quickly discuss. The first is that Dell just gave margin guidance on AI servers that mirrors SMCI’s guidance in regard. Dell’s estimate GM’s were slightly lower than SMCI’s gross margins. And both companies have forecast that their AI server margins will improve over the next few quarters. The short report indicated that DELL was selling AI servers at nearly 0% gross margin. That is not true. You can read multiple analyst reports on this and listen to Dell’s earnings call.
Second is possible channel stuffing with related third parties. Specifically Ablecom and Compusa. Both are owned by brothers of the CEO and the CEO owns a minority interest in both. And yes, channel stuffing (which was a problem) can be done fraudulently in this manner. They can also just ship product to a warehouse or the like and call it a sale. Lots of ways to do this.
The Aug 30 notice indicated that “no material” changes were expected from info so far obtained to prior financials given.
It is quite possible SMCI is gaming the system or just lying about it when they put out that formal SEC filing, but if they are on the level then channel stuffing with related parties or otherwise did not materially happen. As that would require material restatements of financials.
All in all, reasons to be worried and reasons to not be so worried and to look to the future as liquid cooled Blackwell rolls out and data centers make direct liquid cooled servers a mainstream element.
We should know in less than two weeks. If we are being lied to someone will go to jail.
All this is of concern, but I still don’t plan to buy back in. Their moat is not large enough to keep Dell from eating some of that TAM. I worked for a server startup some years ago. We pioneered a new server configuration that was very friendly to server rooms. A couple years later, all the big names moved on our business. Our great business became a commodity business. Absent the SEC problems this seems like a similar scenario.
I underestimated SMCI’s competitive edge before, and I’m not going to guess how long it will last now. I have started buying in slowly (bought 25% of my position today) and will be paying close attention to learn more.
It seems that both TSLA and GOOG are using them for their AI data centers, which is a pretty strong vote of confidence on the product side. I’m choosing to believe that the financial mess will get sorted out, while the business continues to grow strongly and have great future prospects with Blackwell.
I’ve fallen for scams before (Worldcom, anyone??), but I don’t think this is one.
ActonUp
I agree with you. I think all the brouhaha over SMCI will blow over, but what does concern me is sustainability. I never expected the AI build out to last forever, but now I’m wondering just how long will it last?
Has Nvidia announced a follow up for Blackwell? How much better (however that gets measured) does the next act have to be for the hyperscaler and sovereign AI investments to continue? Seems like there’s got to be an upper limit where enough is indeed enough. Has that limit been quantified?
I don’t know the answers to these questions. What I do know is that SMCI will continue to prosper until they don’t. At that point (actually, somewhat prior to that point) SMCI stock will not continue to rise. In fact, I expect it will fall for legitimate reasons.
NVDA has. NVDA now intends to iterate every 12 months instead of every 18 months.
The typical NVDA new chip costs more nominally, but gives materially more computation power at lower energy cost per computation than the previous chip.
Thus, each new chip creates another upgrade frenzy as it enables new use cases, and does so at reduced total cost of ownership. Even though the chip costs more.
It seems likely Nvidia has a few more great upgrade cycles in them over the next few years. And inference servers are just beginning as training has produced more and more models to run.
These new chips will all be liquid cooled. The question becomes, can SMCI make money off this even as their volumes increase? Both Dell and SMCI have guided to returning gross margins later in the year. SMCI has historically had the best gross margins in the industry.
Alas, if I were an SMCI customer I’d be using this event as a reason to demand discounted pricing and ask, why am I paying a premium when you are discounting my competition?
Even after this 10K stuff is behind them the market will still be concerned by this and SMCI will need to show returning normal margins.
So agree with both of you.
I’m talking about just in servers. Dell and HP are conglomerates. They sell everything from enterprise software to PCs and beyond. During some of the years you are citing, as an example, EMC was a high gross margin subsidiary for
Dell.
For enterprise servers only, as that is all SMCI sells, their gross margins were highest in the industry.
That’s what worries me. They can drop their margins to 0 on enterprise servers and still survive.
Agree. In fact analysts were concerned Dell was doing just this to gain market share on AI servers. But in their most recent report Dell stated they expected margins to return to 11-14% (or thereabouts - may have been 9-15 or bit higher) into middle of the year just like SMCI guided.
SMCI is not a hill to die on. Still, if patient (and I know @SaulR80683 prudently sold. Why incur the lost opportunity costs) at this valuation. SMCI has strong appreciation potential IF its margins do return to normal into mid-2025.
The thing has kept me in is the narrative around DLC and Blackwell will be 100% DLC with SMCI leading the industry with DLC. That would give SMCI at least a period of undue advantage over the competition.
That may still play out. At present I had to sell out. But good and bad Tinker keep fighting over this. So I’ve been quite fickle on it.
Looks like the other shoe has dropped:
Down 16.5% today as I type this. At best, I think this is going to take months to clear.
Boom! I was willing to ride out the delayed 10K but not this. Once again I caught a falling knife. Besides the accounting accusation (premature shipment for recognition of revenue) there is now an allegation that shipments were being made to China on export restricted chips. The recent SMCI statement that the annual 10K report will not affect the overall numbers maybe makes sense. Perhaps there was some end of quarter to quarter revenue massaging that has no cumulative effect on the annual report. Given the physically adjoining five fabrication familial businesses, this seems all too easy a game to play a long way from SEC oversight. But at the end of the day, either the net income is there or not there. Perhaps we saw this slippage with the declining margins rather than due all to the capex investments? The signs were there.
So how does the US restrict chips and reseller shipments from Taiwan to China? I have so many questions that I am just not qualified to answer. You have to wonder why a leader, CEO Charles Liang, would jeopardize his company when business is booming and the stock flying so high.
While at Juniper Networks for 7 years, we worked with and bought SMCI appliances for many years; great products, great service, and good people to work with in Silicon Valley. So sad. How will this DOJ investigation affect sales and margins going forward in what should be a spectacular 2025? My head hurts from the whipsaw of information. Yes, a DOJ investigation will take many months to resolve, creating a overhanging dark clouds. I’m out. Sorry for my venting and this useless contribution to this board.
-zane
That accusation was actually made in the Hindenburg report. TBF, though, there’s probably a gray area covering to what extent companies need to ensure that the customer is not in/from an export-restricted country, or that he company being sold to will not in turn sell/trade to an entity in an export-restricted country.
And the WSJ report says:
The prosecutor has asked for information that appeared to be connected to a former employee who accused the company of accounting violations, the report added.
Interesting that this is DOJ and not SEC, unless they’re co-operating on the investigation, or it’s on multiple fronts.
Full disclosure having posted quite a bit about SMCI over the last few months - I have just sold out at $400. I sadly feel that there may be further to fall once the DOJ begin their investigation - and I wanted to protect the gains I have in SMCI (100% having first bought in May 23 at $197).
Once this clears - which may take some time - then I may buy back in again if the story is intact. But right now, there are too many question marks hanging over this company and so I have reluctantly sold.
Jonathan