FORTUNE — Amazon might have netted nearly $900 million in the fourth quarter. Or it might have made just north of nothing. Wall Street doesn’t seem to have a clue.
What’s clear is someone is bound to get a huge surprise when Amazon (AMZN) delivers its fourth-quarter earnings Thursday afternoon. On a per-share basis, Wall Street analysts think the online retailer earned anywhere from $1.88 to just a penny.
That’s the second biggest range of estimates, behind Google, of any company in the S&P 500, and it’s the biggest, by far, when you consider what Amazon is likely to earn, according to an analysis of Bloomberg data. The average range of disagreement for earnings per share among analysts on stocks in the S&P 500 is just $0.17. Amazon’s gap is 10 times that.
And the range is not just because of outliers. There are six analysts who believe the company will earn less than $0.50 a share. And eight think the company will earn more than $1.50.
The wide range of estimates is just another example of how Amazon has upended the normal way Wall Street works. Back in October, investors and analysts cheered Amazon’s $40 million loss in the third quarter. Amazon has missed earnings estimates in five of the last six quarters. That’s usually the kiss of death for most stocks. Instead, Amazon’s shares have more than doubled — to a recent $385 — in the past two years.
One reason for the wide range is that Amazon isn’t very open with its analysts. Most corporate executives give analysts earnings targets. Amazon CEO Jeff Bezos doesn’t do that. At the end of the third quarter, Amazon told investors that the company’s operating income could be anywhere from a $500 million loss to a $500 million gain for the last three months of 2013. Analysts have been more optimistic, but their estimates also have a $1 billion range.
The uncertainty may also have to do with Amazon’s wide following on Wall Street. There are 46 analysts who cover Amazon on Wall Street. Apple (AAPL) has 44 analysts and, excluding one of them, the range of disagreement in estimates of Apple’s earnings per share is $0.50.
The gap in Google’s earnings estimates is nearly $3 per share, but as a percentage of the $12.00 a share the company is supposed to earn, the amount by which any one analyst will be wrong is not that much.
For other companies, major disagreement in earnings estimates, especially for large ones, is rare. The range for next quarter earnings at Wal-Mart (WMT) is $0.10 per share, according to Thomson Reuters. The earnings gap at Hewlett Packard (HPQ) is $0.06. There is more downside to being very wrong than there is upside to being very right.
Stock analysts tend to be a risk-averse crowd, especially among Wall Streeters. Every now and then, a few analysts, like Henry Blodget or Meredith Whitney, make a name for themselves with bold calls, and controversy. But generally, the simplified view of Wall Street analysts is that they are a kind of mouthpiece for corporate executives, largely passing along the earnings estimates, with some analysis, that corporate executives give them. It doesn’t sound like a great job.
But then you have the situation with Amazon. Being an analyst on Amazon actually sounds kind of fun. With little guidance, you imagine Amazon analysts actually having to sweat over the kind of models that are taught in business school, carefully calibrating their assumptions, coming out with very different conclusions than their peers, and then getting to see who got it right.
Except for this: No one really cares. Not even the analysts.
I e-mailed Carlos Kirjner of Sanford Bernstein about the fact that his estimate at $1.88 a share for Amazon is the highest on the Street. I expected a full-throated defense of his prediction, which he had no doubt spent countless hours on. What I got was a shrug. “EPS [earnings per share] is meaningless for the stock,” wrote Kirjner.
Indeed, some of the analysts didn’t even seem to know how their estimates stacked up against competitors. Kerry Rice of Needham thought his estimate was one of the highest. But at $0.44 a share it is one of the lowest. “I don’t think people who own Amazon are overly concerned about the bottom line right now,” says Rice.
Not worrying about what a company earns, of course, breaks another rule of investing; one of the most basic rules, in fact. But that’s the story Amazon has been selling, and Wall Street has been buying it for a while. The company is thinking long-term and reinvesting the money that would normally be profits to continue to build its business. All Bezos and company have to do is pull back on that investment — and they will, eventually — and Amazon will be hugely profitable.
The problem with that story is that when a company reinvests in its business, what you normally see is asset growth, all the stuff the company is acquiring with the money that it’s pouring back into its business. But you don’t see that at Amazon. Assets have not really been growing, or least not as fast as its stock price. Amazon shares are now valued at 22 times the company’s book value or net worth. This calls to mind another investing rule: companies can trade for a lot more than they are worth, but not for long.