Description

*Given the ~$50M market cap and fairly illiquid trading, this idea is best suited for small funds or PAs*

TABLE OF CONTENTS

Introduction

Technology – What do they do?

Management Change & New Business Model

Financials

Recent Developments

Valuation

Why does this opportunity exist?

Other Considerations

Risks

Catalysts

Introduction

I am long Intermolecular (IMI), a materials discovery company that primarily serves the semiconductor industry.

IMI is a high margin “semi”-recurring revenue technical services business currently valued at <1x Sales, with net cash of ~60% its market cap. The company was slightly FCF-positive over the last year. Given their attractive 65%-70% GMs and a highly fixed OpEx structure, cash flow rapidly accrues to the bottom line as they layer in additional revenues.

A combination of a recent short term drop in revenues with their two largest customers due to some pauses in new programs, and recent forced selling by several top holders have pressured the stock. This is despite two significant customer wins since November that provide needed customer diversification and collectively present a significant incremental revenue opportunity.

Between growth at existing customers, additions of new ones, and the operating leverage in the model, I believe the stock has 5x-10x upside over the several years.

Technology – What do they do?

IMI has focused on being a leader in materials science since its inception in 2004 (IPO in 2011). Currently, as the semiconductor industry struggles with the limits of Moore’s law, materials discovery has become increasingly important. IMI’s proprietary PVD & ALD tools and software form the basis for their platform that allows for significantly more rapid materials discovery. IMI’s platform allows their (primarily semiconductor) customers to test materials for certain properties that they are targeting (capacitance, leakage, thermal stability, etc), and to do so much more rapidly than through other methods. IMI’s processes can accelerate materials discovery by 10x-100x, which also significantly reduces costs to the customers. While IMI has served a variety of industries in the past, their predominant focus today in within memory. The below graphic from IMI’s investor deck gives a simple overview of their technology versus the convention.

A natural question one might ask here is if IMI is as unique in the market as I’m claiming here, who do they compete with? The simplest answer is they essentially compete against other uses of their customers’ R&D dollars. As IMI puts it:

“We are not aware of any companies that currently compete or have to date competed with us in the use of combinatorial methods for thin-film development; however, we do believe that we compete for the R&D resources of our customers with third-party IP licensors, equipment suppliers, industry consortia, alliance partnerships and university research teams.” [IMI 2018 10-K]

While this may potentially imply a contentious relationship between IMI and their customers’ internal R&D groups, I do not view it this way. Where it makes sense (for speed and cost) IMI serves a synergistic role alongside these groups. IMI will not be used when they don’t save money/time, but when they do, they are nicely complimentary to their customers’ internal teams.

Management Change & New Business Model

While IMI’s core technology has remained consistent (along with the value-add to their customers), the business has undergone some transformation over the better part of the decade that they’ve been public.

When current CEO Chris Kramer joined the company about 2.5 years ago, the company’s business model was one that relied heavily on largely unprofitable up-front collaborations and then a reliance on back end royalties (royalties which themselves were highly dependent on the fickle nature of success of their customers’ products developed using IMI).

Most recently, though, under Kramer’s leadership, the company fundamentally altered its model into a pure outsourced services model in which the services engagements IMI takes on are priced for profitability and there are no requirements that their customers pay royalties. Under the new model, IMI gets more certainty and the customers aren’t subject to potentially large future royalties (which are generally loathed in the semi world) on their successful products.

To summarize, the benefits of IMI’s new model are several-fold:

1) IMI is not stuck in a series of unprofitable service engagements (in the hopes of one day getting made whole through back end royalties)

2) The new model is more attractive to customers: again, semi customers, particularly in memory, are loathe to be locked into back-end royalty engagements

3) The IMI revenue model much more closely resembles the more familiar high margin recurring revenue model adored by the street

Under the current model, IMI’s customer engagements typically involve a “multi-year umbrella agreement” (2018 10-K), under which the customer can have multiple ongoing projects, each of which would typically last for 2-4 quarters.

In addition to this fundamental change to the business model, Kramer made two other key strategic moves:

1) The company took dramatic steps to reduce its OpEx to position the company to be cash-flow positive at current revenue levels. The company has reduced its quarterly OpEx run-rate from $10M+ to below $7M/quarter under Kramer.

2) While the company has historically served customers in the Solar, LED, and industrial glass industries, Kramer directed the company to focus exclusively on semiconductors (which can include chipmakers, materials suppliers, and equipment makers). The insights here were that semis where the company has historically by far added the most value and also that the revenue opportunity just within semis is large enough on its own.

Financials

A few items on the above:

1) Essentially all current revenue is now service program revenue (as described above). The only exception is $200k-$300k/quarter of license/royalty revenue. The last meaningful royalty stream was $1.25M/quarter from Micron which ended in Q417, as can be seen above.

2) Revenue has obviously fallen off in 2H18, which can been seen to be driven by IMI’s top 2 customers, Samsung and SK Hynix

3) Intel became a new customer in Q418 under an $8M+ 1 year deal. With this, Intel should become a ~20% customer in 2019 (they had never before been an IMI customer)

4) Adjusted EBITDA was what I chose to show for the bottom line because CapEx is de minimis at ~$1M/year (while D&A runs $4M+/year) and the company has large NOLs and thus pays no taxes. Adjusted EBITDA (less ~$1M for CapEx) is the best proxy for FCF

5) Under customer-specific revenue numbers, blank spaces indicate either zero revenues or no information

Recent Developments

IMI picked up an $8M+ contract with INTC lasting ~1yr for $8M+ which will run for one year, with ample opportunities for follow-on behind it (the first INTC revenue was >$0.6M in Q418).

And while not named, IMI picked up another huge customer (TSMC) for a trial contract in Q119. This was their first entrée with TSMC and their expectation is that this leads to more business subsequently.

Valuation

Cash backstop of $30M+ ($0.61/share):

• The company has recently been slightly FCF-positive – I don’t see this changing meaningfully (i.e. turning to cash burn)
• The company hasn’t made an acquisition in its ~8 year history as a public company. Recent conversations with management indicate no interest in pursuing this as a use of their cash.
• They announced a $10M buyback in November ’18 and are pretty straight-forward that they see this buyback authorization as a direct way to deal with the aged VC funds, etc that are forced sellers. I expect that by the next time they announce in May, they’ll be able to discuss putting their cash to use to take out large exiting holders. 

So we’re left with the current ~$22M EV. It’s well less than 1x Sales on a trailing basis (and recall ~65%-70% GMs on this revenue).

Base Case

First, let’s start with Samsung/SK Hynix. As can be seen above, both have seen a large drop-off in revenues to IMI, particularly Q4. The Q4 revenue levels represent 1 program each at Samsung and SK Hynix versus 3 per customer a year ago. Management believes the likely trajectory here in both cases is up. According to a recent discussion with the CFO, his sales team tells him 9 projects at Samsung are highly likely. I’m going to take the approach that they can get back to the recent past where they were doing ~$5M/quarter with Samsung and ~$2M/quarter with SK Hynix.

The biggest new customer win in the last year plus for the company has been Intel, which they’ve announced will be an $8M+ deal over a year (began late Q418), so assume ~$2M/quarter through the balance of 2019.

While the company has significantly downplayed, on the company’s February Q418 earnings release they disclosed that they had won an initial deal with a global Tier 1 semi OEM. This is TSMC. They had never worked with TSMC. The project involves helping TSMC with leading edge memory technologies that they hope to license. I’ll give them credit for $1M/quarter from TSMC.

My base case quarterly revenue buildup (near term – possibly by Q219/Q319):

So here we have nearly over $8M annualized FCF (~$0.17/share), just based on their current customer base at revenue levels specified in new contracts (Intel) and revenues rebounding to recent levels for existing customers (Samsung, SK Hynix).

Growth Case

Assuming the company can grow top line a 20%/year (somewhat a guess, but a level management thinks is achievable) through a combination of expanding business at current customers, plus layering in some new ones (the company believes both Toshiba and GlobalFoundries are strong prospects for new deals).

While the above is not intended to be a precise forecast, it does show the inherent leverage in the operating model and a range of potential price targets.

One natural question is whether the $80M+ in annual revenues in Year 3 is realistic. Here’s how I think about it by customer:

Samsung: Has run at $20M+ based on 3 concurrent projects. Company believes 5-6 is realistic, while more optimistic sales guys think 9 projects at a time is feasible. Given that it’s well-established that Samsung can maintain $20M+/year in revenue for IMI, I think $30M-$40M is not unrealistic several years out

Intel: Given that Intel’s R&D spend (see chart below) is essentially on par with Samsung’s spend, I do not think that a doubling of the 1 year $8M+ contract that IMI currently has to $16M/year is unrealistic

SK Hynix: SK Hynix has been at a $10M+ annual revenue run-rate within the last year, so plugging in $10M doesn’t seem unrealistic

TSMC/Toshiba/GF/Other: Given the above, ~$20M would be required from the remainder here in aggregate in order for IMI to reach $80M+ in annualized revenue. Non-named customers already generate ~$2M/quarter, so this group in aggregate reach $20M+ does not seem like a stretch

Putting these together, $80M+ in annual IMI revenue 3 years out does not seem unrealistic.

Growth Case with Buyback

Given the aforementioned forced sellers and the company’s $10M buyback put in place with a likely targeted usage, I think it’s worth examining the upside should the company take out 10M shares @ $1.00. 10M shares is my best estimate of stock for sale from Redpoint and the Lloyd Miller III estate. The below reproduces the Growth Case, assuming 10M shares bought back by the company at $1.00/share:

As can be seen, where factoring in a large buyback (which I view as likely), the 3 year out PR range move from $5-$9/share to $6-$11, so could be highly accretive to future value.

Why does this opportunity exist?

There’s no shortage of answers here when dealing with a $20M EV company, but the key reasons IMO:

• 8 years as a public company – no real growth or profits to speak of
• $50M market cap; just 1 sell-side analyst covers
• Recent drop-off in quarterly revenues, driven by declines in revenue from both Samsung and Hynix (top 2 customers)
• Lack of progress in in China, after announcing a high profile partnership over a year ago to go after China customers - recent discussions with management acknowledge de-emphasizing these efforts in the current geopolitical environment. To be clear, this is a lack of gaining new business, not a loss of any kind.
• Swath of non-discretionary selling from end of life VC funds and the estate of a large holder
• USVP distributed shares in late 2018 – pressured shares in late 2018, but likely these have now been largely absorbed by the market

• Lloyd Miller III Estate started selling in February 2019 – there’s likely still a couple million shares to come here (no longer a 10% holder so not getting regular updates)

• Redpoint Ventures (15% holder) just began selling at the beginning of the month

Other Considerations

• 30% activist shareholder (Raging Capital)
• Likely has strategic value to an equipment company
  • Over $200M of cumulative R&D spend in the last decade
  • Could allow a semicap equipment company to get closer to customers’ R&D teams (specifically around materials).
• A maker of PVD/ALD tools such as AMAT would seem like a reasonable fit.
• IMI owns the rights to Resistive RAM (ReRam) technology – Sandisk/Toshiba once had exclusivity to develop but they let their rights lapse, so now it’s owned outright by IMI again. ReRAM is thought to be one possibility for the next generation of NAND-like (solid state) technology. I consider this merely a far out of the money call option.
  • While I don’t believe anyone can say for for sure, MRAM (Magnetic RAM), and ReRam (Resistive RAM) seem to still be the primary areas of focus on those looking to the next generation of NAND/solid-state

Risks

• Negative cyclicality in semis
  • IMI’s revenue comes out of R&D budgets, so should prove less cyclical than equipment makers reliant on CapEx dollars, but this nonetheless is a risk
• Large sellers currently remain in an illiquid name
• Customer concentration: Samsung and SK Hynix were a combined 75% of revenue in CY18
  • Recent additions of INTC (Q418) and TSMC (Q119) have already helped here
  • Company is hopeful on additional near-term deals with Toshiba and GlobalFoundries, which would future reduce concentration

Catalyst

Catalysts

• Rebound in revenues from major customers Samsung and SK Hynix
  • As of Q418, each customer was generating revenue from only 1 project versus 3 each a year ago; I expect a rebound from each within a quarter or two
• With some revenue growth, significant FCF generation that will force investors to appreciate how underpriced the stock is
• Customer diversification
  • Added INTC in Q418 and TSMC in Q119 and are hopeful on deals near term with both Toshiba and GlobalFoundries
• Clearing out of large sellers (company has a sufficient buyback in place to make this happen)