For me the most memorable scene in the classic action movie The Matrix is where Neo (Keanu Reeves) enter the Matrix (computer-generated reality) for the first time and fights his mentor, Morpheus (Lawrence Fishburne), in a kung fu battle. Morpheus uses the experience to explain how the Matrix works to the rookie: “[The Matrix] has the same basic rules like gravity. What you must learn is that these rules are no different than rules of a computer system. Some of them can be bent, others can be broken.”

To my mind, this is an apt analog for valuing stocks. Most of the time, companies trade on a multiple of future earnings, or more precisely on a multiple of expected future cashflows (themselves a function of future earnings in most cases). Where higher earnings and faster growth is expected, a higher multiple is accorded; the same goes for greater visibility, or predictability of future earnings. The idea that earnings power, or cash-generating power, drives equity valuation is so fundamental to the accurate valuation of all enterprises that if it did NOT apply to the vast majority of traded equities, capitalism would cease to function (how would you rationally value businesses otherwise?). It is the very glue binding capital providers (investors, banks, venture capitalists, etc) to capital employers (companies).

But, as in life, the stockmarket is a product of human emotion as well as rational calculation, and so markets are periodically subject to bouts of mania and speculation (bubbles in bull markets and massive crashes in bear markets) that most all end in tears. For short (or sometimes long) periods this can apply to the whole market, or large segments of it. But even when most of the system exhibits mostly rational pricing, there remain pockets of outright irrationality where the normal rules clearly do not apply. Hence, as in The Matrix: some rules can be bent, and some can be broken.

I’ve always found it fascinating how two companies in the exact same industry, producing the same product, can be valued so violently differently as to make all kind of rational explanation impossible, forcing analysts to jump through all kinds of intellectual hoops to try to justify valuations that most sane, neutral observers could not fathom. General Motors (GM) vs Tesla (TSLA) is perhaps one such intriguing pair: GM produces ~10mm vehicles a year and is valued by the market at ~$60bn, or ~3x current year EBITDA; TSLA is hoping to produce just 100k vehicles by the end of this year, yet is valued at $30bn, and trades at 55x EBITDA. I’m not trying to decry TSLA’s valuation specifically – rather, I just hope to demonstrate that despite the common product, the market is employing two fundamentally different sets of methodology to value these two companies (and the analyst pool, common to both companies, has bought in to both at the same time…).

In this sense, localized manias – sometimes isolated to particular sectors or even single names (‘cult stocks’) – exist all around us, surrounded by otherwise reasonably-valued, ‘normal’ stocks. What makes some companies so lucky? Why do they get to trade at astronomical valuations? A massive share price premium is effectively a free gift from the market that can help expensive companies create their own reality (by funding, through rich stock, the acquisition of less fortunate companies at cheaper multiples) – what’s the secret sauce?

While the variants are numerous, most localized manias exhibit some combination of the following: a revolutionary new product or technology that promises ‘disruption’ of existing large markets; the opportunity for a long growth runway into such large markets (‘blue sky growth potential’); clever management that massages and markets the company into ‘story stock’ (often by focusing attention on non-standard performance metrics that suggest higher future potential and therefore valuation); and, of course, the ability to execute and deliver on growth promises.

Most cult growth stocks exhibit all of the above but fail to consistently execute on their promises (when measured over years), and so generally tend to devalue to more defendable valuations as the mania dissipates. Sometimes this happens gradually, over years; more often these stocks are repriced viciously in short periods (a number of SaaS and biotech stocks suffered this fate last March, for example). Occasionally the very best names defy the odds and continue to hit high growth expectations (or maintain the mystique, at least), quarter after quarter, year after year. An incomplete shortlist of these kinds of names would include Amazon, Tesla, Under Armor, Salesforce.com, and, of course, Netflix. But it is important to remember that these few cases are the exceptions: in most cases, reality catches up with the manias, whether it takes a quarter or many years, and rationality returns to the pricing mechanism of the market.

For those looking to benefit from the irrationality of the market (other than riding the wave and hoping it continues), you can try to figure out when the party could end. This is difficult, to be fair, but eminently achievable, and amply compensated by the potential rewards. Furthermore it is intellectually valuable as an exercise to help avoid investing mistakes on the long side (ie, not falling victim to manias yourself).

Keeping this in mind, let’s look at Netflix (NFLX). In many ways it is the quintessential ‘story stock’: it offers a heady combination of disruptive technology, high growth now and ‘blue sky’ potential growth going forward; it generates a boatload of revenue but no profits (and burns a ton of cash); it guides investors to look at company-defined metrics (‘contribution profit’) rather than GAAP financials (which are ugly); it has a brilliant management team pulling the strings rather expertly; and, perhaps most importantly, it is readily visible and tangible to retail investors (crucial to a real cult stock is the irrational investor who loves the product and so blindly invests).

But before diving into the current picture, it’s worthwhile taking a quick survey of where Netflix came from, for the company’s current situation is a function of the various transformations it has gone through. Started as a mail-order DVD rental business, Netflix’s first vision – ‘Netflix 1.0’ – was a pretty good business: it provided a convenient service better than the lethargic competition (ie, Blockbuster) and prospered in the early-mid 2000s. DVD rentals was truly a business with scale (regular readers will know I am obsessed with operating leverage), since adding 10 customers doesn’t require you to buy 10 more copies of your DVD library.

Unfortunately for Netflix, the market changed, DVD rentals were going the way of the dodo, and so Reed Hastings (Netflix CEO) and co made the ingenious move to invest heavily in becoming an ‘over-the-top’ provider of on-demand video content (‘Netflix 2.0’). This worked great, initially, and subscriber growth accelerated. But unfortunately the studios – who initially had sold over-the-top content rights for a song, since the subscriber base started very small and presented little threat – quickly realized that the more Netflix grew, the more over-the-top threatened cable subscriptions and hence the value of their content, and therefore began to jack up the cost of their content to Netflix dramatically. This never-ending, negative leverage cycle – the more subscribers you attract, the more expensive your content acquisition became – threatened to destroy Netflix 2.0.

So management – brilliant as they were – decided to reinvent the company, again, around the provision of original, unique Netflix-created content (House of Cards, Orange is the New Black, etc), and in effect becoming “the HBO of the internet – or, as you probably guessed, what I call “Netflix 3.0.” Strategically, this is where we are now.

The quality of the new programming was, thankfully, very high, and Netflix continues to attract subscribers quickly; in fact Netflix for the last couple of years has been pushing aggressively to replicate their model internationally (mostly because the US market is becoming saturated). Unfortunately, though, non-Netflix content acquisition costs didn’t come down as much as they hoped, mainly because a) a number of competitors (Hulu/HBO Go/CBS etc) have entered the space, often the media companies themselves, creating bidding wars for content; b) as more Americans ‘cut the cord’ and cancel their cable subscriptions, the value of content in a 24/7 broadband world has skyrocketed (not just for Netflix but to all of the media companies); and c) expanding internationally costs more in licensing fees than just distribution in the US alone. Additionally, the creation of top-quality programming requires, well, a lot of money – so the net result is content acquisition – now a combination of original Netflix content and licensed from other studios – continues to outpace revenue and subscriber growth quite dramatically.

The net result of all these gyrations can be summed up by a few simple stats. In 2007 (the tail-end of ‘Netflix 1.0’), Netflix had ~7mm paid subscribers, generated $1.2bn revenues, made ~35% gross profit, spent 28% on opex, and generated ~8% operating margins. Additionally the company generated a modest amount of free cash flow (~$50mm) and had no meaningful long-term content liabilities. This is decent if not stellar financial performance. By 2014 – squarely in ‘Netflix 3.0’ mode – the company has grown the subscriber base astronomically to ~57mm (roughly 40mm in the US), and revenues to $5.5bn. But gross margin has contracted to 32%, and opex has kept up with sales growth, such that operating margins are actually lower (~7%) – despite the mammoth growth in sales and subscribers. More worryingly, the business has been consistently cash consumptive (burning ~$130mm in cash in FY14), even though content acquisition costs (and attendant cash requirements) only accelerate into the future. The company has $9.5bn of content liabilities, of which~$4.2bn are not even on the balance sheet yet – these liabilities have grown at roughly twice the pace of revenues since 2007. This picture would be scary enough – except that it doesn’t consider that DVD rentals – a business that will eventually disappear – still generates ~15% of Netflix’ revenues and ~35% of their ‘contribution profit’ (a non-GAAP measure that excludes content acquisition costs)…that is still a huge chunk of the company’s business, for a segment likely destined to melt away.

I don’t want to get too bogged down in the details here; the market has already made its judgement (the stock ripped 18% yesterday to a new high, with the market cap now $34bn). At this point there are two basic tenets to the bull thesis, and the stock won’t go down until one or both of those tenets are seriously challenged. They are: that Netflix’ subscriber growth opportunity is still massive (the ‘blue sky’ growth argument); and that – perhaps like Amazon – there is tremendous operating leverage built into the model once Netflix is the dominant player (hence we should not be worried about low profitability at the moment).

Let’s dig a little deeper into the first point. How massive, really, is the subscriber growth opportunity? As mentioned, Netflix ended last year at ~57mm subscribers; as of 1Q, they are near 62mm, over 40mm of which are in the US. There are only 117mm households in the United States, so we are already at ~35% penetration; and, assuming at least some keep their cable, use other subscription services, or frankly borrow their friends’ Netflix passwords (I’m guilty!), I really don’t see how much more growth there is in the US. In fact the company acknowledges this implicitly (via the push to international) and explicitly (by reporting growth in US subs has tailed off in recent years).

As for the international opportunity – clearly the absolute number of potential subscribers is much larger, though the cost of getting to them is also, exponentially larger (given different geographies, licensing requirements, more competition, and less broadband penetration). In fact the company has not yet generated any profits internationally and it remains very much an open question whether the Netflix model can work abroad. Nevertheless – in terms of pure subscriber adds, there is certainly green grass to be enjoyed overseas.

But what about the second – and I believe more important – pillar of the bull thesis? At this point, I think we really have to reconsider how much pricing power/operating leverage Netflix will be able to extract. Consider: direct over-the-top competition is only increasing (cf, the launch of HBO Go, Hulu, CBS streaming, and likely some kind of Apple solution in the future); and the cost of content is still accelerating. Price hikes were achieved last year, but as competition increases and other unbundled solutions become more available, the inelastic demand enjoyed by Netflix so far will likely dissipate (this, at least, is how consumers behave in other crowded, competitive markets). And on the cost side, is it really reasonable to believe that content costs will come under control at some point? If you don’t believe revenue/subscriber growth goes on forever, this is what you need to believe to think Netflix has a viable business (let alone a stock worth buying up here).

But consider the nature of Netflix’ cost base. Outside of marketing (which could, at some point, exhibit a bit of positive leverage), the vast majority is content acquisition (incidentally – this is why Netflix reporting ‘contribution profit’ by excluding content costs is a complete joke on the market and all analysts who parrot it). We have discussed why external content costs are unlikely to go down (in fact how they have gone up as fast as, or faster than, Netflix’ subscriber growth); this is a big problem when subscriber growth slows. As for internally-generated content – this is a fairly inelastic cost item on a per-unit basis (which is bad if you are trying to achieve scale). Whether Netflix has 50mm or 100mm subscribers, the cost of making high-quality TV series is basically what it is – TV shows are not iPhones. And the more subscribers you have, the more different shows you need (to support the more diverse interests of your base). This is another reason why Netflix’ content costs are accelerating so dramatically (they keep making new, expensive shows). I guess if Netflix’ subscriber base fell massively, the company could get away with making fewer shows, though in that case clearly the company would have bigger problems. But otherwise – I just don’t see how the company ever generates real operating leverage from the model. There is simply too much competition.

There is one way Netflix can extricate itself with honor from all this: forget about external content, abandon the plans for world domination (via aggressively adding subscribers at any cost) and become, truly “the HBO of the internet.” With a smaller, loyal subscriber base the company could focus solely on a limited amount of high-quality content, and charge a premium price as a result – this would be sustainable, as content costs would come down dramatically, and the need to increase advertising spend continually to enter ever-more markets would also be lower. Of course the company will not pursue this angle as then Netflix would have to be valued as a boutique media company (think AMC Networks, Lions Gate, or Starz perhaps) – all of which are valued by the market in the $3-$5bn range (Netflix is at $34bn currently). Even in the most optimistic case where Netflix is valued akin to HBO – and really this is beyond extravagant, as HBO’s catalog has been built over 30yrs – we are talking about ~$20bn in absolute $ in an acquisition scenario (at least that is what Fox was apparently valuing HBO at, when they thought about buying Time Warner last year – this was ~10x HBO’s EBITDA). Of course, independent, standalone valuations would be lower, and even then – this is ~40% below where the market is currently pricing the company.

Thus, Netflix is trading, to me, anywhere between 2-8x its ‘fair’ value – hence my contention that it is the ultimate story stock, breaking most all the normal valuation rules applied in other parts of the market. At this point I should mention – I am NOT short Netflix stock. While I clearly am skeptical the company can ever grow into and thus justify its valuation, as long as the market focuses on subscriber growth at the expense of all else, it is impossible to accurately time the short. However – the point at which subscriber growth meaningfully slows; or the market turns its focus, belatedly, to earnings power, will be an opportune time to short the stock (even if your absolute starting point is much lower).