Amazon is the non-profitable behemoth of the ecommerce world. Loathed by many for putting indy book stores out of business; lauded by at least as many for transforming the time-consuming chore of shopping into a quick, easy, cheap and less risky endeavour, offering the restoration of hours of free time to anyone with an internet connection.

Amazon started out, of course, selling books but sometime in the early ‘00’s began selling everything. Today Amazon sells 356 million products. Walmart.com sells 4.2 million. A typical Walmart store sells 200,000. In the latest quarter, Amazon, which began operations in 1995, reported trailing twelve month revenues of $95bn, roughly half the revenues of Walmart, which opened its first store in 1962.

So, Amazon is big and it is useful. But it is not profitable. And because of this, some fear it is an elaborate Ponzi scheme. Could this be true? How long can Amazon continue in business while failing to turn a profit? And if Amazon does not report profits, how can it possibly justify a $235bn market capitalization?

Amazon seems to have studiously avoided reporting net profits since it opened its metaphoric doors 20 years ago. Some investors maintain that Amazon operates at break-even on purpose. We have some sympathy with this view.

First, it’s a business strategy: Amazon has its own chart explaining its strategy which starts with customer experience and ends with lower prices. The implication is clear. Whenever it can (i.e. whenever it has just earned a few dollars of profit) the company will lower prices to enhance customer experience and drive growth.

Second, it’s tax-effective: No reported net income does lower a corporation’s tax bill. Reinvestment of profits is a tax-free alternative.

Third, it’s a competitive weapon: Amazon clearly wants to be a very big fish in the retail sea and operating at break-even makes it hard for competitors, some of whom have more traditional shareholders who prefer profits and dividends over endless investing for long term growth.

However, we also note that between 2000 and 2010, Amazon’s net income line swung from a loss of almost $1.5bn to a profit of $1.1bn. The profit reported in 2010 was not enormous on a percentage of sales basis (3.3%) but still, the upward trend was quite well-established. Then in 2011, this trend was interrupted by an explosion in operating expenses which rose almost five-fold from 18.2% of sales in 2010 to 30.3% of sales (last twelve months), easily overshadowing the increase in gross margin which took place at the same time. Coincident with this explosion in operating expenses was another explosion in capital spending which also soared five-fold from $1bn in 2010 to $4.8bn in 2014.

To us, these numbers are not really consistent with the theory that Amazon has just been plodding along, deliberately breaking even for strategic, tax and competitive reasons. These numbers suggest Amazon was plodding along, heading towards higher profitability and then changed its mind.

One thing Amazon clearly changed its mind about was the pace at which it would grow its fulfillment platform. From January 2010 to December 2014, fulfillment square footage increased from 17 million square feet to 102 million, an almost 6-fold increase in five years. This dwarfs the historical growth rate – a comparatively pedestrian 25% p.a. Not surprisingly, fulfillment costs as a percentage of sales soared over the same period, rising from 9% to 15% of product sales. But now we can all get our stuff next day.

Another thing Amazon changed its mind about was cloud computing. Around 2009, Amazon launched Amazon Web Services (AWS) to provide “storage, compute, database, analytics and applications for any type of business.” Now, the cloud computing business requires prodigious amounts of capital to be spent on technology infrastructure (servers) and so it should not come as a shock that Amazon has probably spent close to $7bn on servers for AWS over the last four years. And, because these purchases are financed with capital leases, this spending is over and above the explosion in regular capital spending we talked about in the paragraphs above. In 2014, Amazon’s total investments of $9.8bn were seven times larger than its total investments in 2010. Few businesses could withstand such a large increase in spending without adverse consequences, so it is totally understandable that Amazon’s nascent profitability wilted under the pressure.

Amazon has always encouraged analysts and investors to measure the profitability of its business on a cash-flow basis. Specifically, Amazon states in its 10K that its “financial focus is on long term, sustainable growth in free cash-flow per share.” While, technically, free cash-flow per share has grown at 7.5% p.a. since 2009 (and 22% p.a. since 2011), once we add back in spending on capital leases, free cash-flow has not grown at all. In fact it has moved from a surplus of $2.8bn in 2009 to a deficit of $2.1bn in 2014.

A break-even P&L can be overlooked if there is enough operating cash-flow to cover investment spending. But negative free cash-flow moves the burden of financing growth on to the balance sheet which is definitely not an infinite source of capital.

So, upon close inspection, Amazon’s recent results are neither sustainable nor free cash-flow positive; the two things Amazon’s management promised they would be. And, surely, if Amazon cannot even show a positive number (let alone growth thereof) for the metric it has emphasized it should be measured against for all these years, the company just has to be a Ponzi scheme. Right?

Well, we can think of a couple of mitigating arguments immediately. No doubt there are others. First, capital spending, at the end of the day, is discretionary – even in a competitive environment. The folks at Amazon are not stupid. Do they really intend to trash their balance sheet for the sake of a spending frenzy on servers for AWS and putting a fulfillment center on every corner in America?

Second, Q2 2015 numbers did show a sharp drop in the rate of growth of both sorts of capital spending. The year on year increase in spending on capital leases fell from 114% in Q4 2014 to just 17% in Q2 2015 and capital expenditure actually fell in H1 2015 year on year. Hallelujah. The result of all this restraint is that on a trailing twelve-month basis, Amazon was almost break-even on a free cash-flow basis in Q2 2015. Was the explosion in investment between 2010 and 2014 a one off? Maybe. We may find ourselves sufficiently curious about this point over the next couple of weeks to put a call into the company.

Swayed by these mitigating arguments, our supposition is that Mr. Bezos has not (yet) succumbed to hubristic clouding of the faculties and will not sacrifice his balance sheet on the altar of growth. He knows that nothing interferes with the “sustainable” part of his mission statement like a crumbling balance sheet. Our bet is that spending will become more controlled in the future and will be (mostly) financed with internally generated funds.

Amazon may not be a Ponzi scheme, but is it worth $500 a share? Amazon’s shareholders are a forgiving bunch and have been more than happy to ignore the bottom line and value the company on a price/sales basis for many years. At 2.5x Amazon is actually not that egregiously expensive. Both Dollar Tree and Home Depot are selling at 1.8x sales. And the stock looks like a bargain against Alibaba, trading at a starry-eyed 13.5x price/sales. The Consensus is of the opinion that it is fine to value Amazon on price/sales because the company could turn a profit any time it wanted to just by raising prices a smidgen. To reach a reasonable valuation, however, prices might have to go up by more than just a smidgen. A 5% increase in total revenue, ceteris paribus, would produce roughly $4bn in net profits and leave the stock on a still generous PE multiple of 58x.

No, to make the stock a buy here, investors have to hope that Amazon really never comes to be valued on profits but continues to change hands on the basis of a reasonable multiple of revenue.

One key to this continuation is the belief in “sustainable growth”. (Mr. Bezos, please see above analysis). Another key is the size of the market opportunity in front of Amazon. Ecommerce today has roughly an 8% share of US retail revenue. Amazon, with roughly 14% of the ecommerce market, has 1% of US retail sales.

Amazon is big – but could easily be a lot bigger.