Every day, Wall Street analysts upgrade some stocks, downgrade others, and "initiate coverage" on a few more. But do these analysts even know what they're talking about? Today, we're taking one high-profile Wall Street pick and putting it under the microscope...

It's been five days since First Solar (NASDAQ:FSLR) stock jumped in response to a series of analyst reratings -- reratings that themselves came in response to a colossally bad earnings report from First Solar itself.

Bravely ignoring First Solar's reporting of a $6.92-per-share quarterly loss, analysts at Roth Capital mused that, well, at least "expectations may have bottomed." R.W. Baird went so far as to opine that the company seemed to be "gaining momentum."

"Posh!" replies Axiom Capital. First Solar is a mess, and you should just sell it. Here are three reasons why.

Construction worker holding solar panel.

Would you like to buy a few shares of First Solar stock? Axiom wouldn't recommend it. Image source: Getty Images.

1. First things first

Expectations for a turnaround at First Solar hinge upon the company's development of a Series 6 module which First Solar says it will be able to produce in a couple years for a cost of just $0.22 per watt (and sell for a bit more). In so doing, First Solar plans to skip right over building a Series 5 module, and jump a generation -- straight from Series 4 to Series 6. The generation-head move, says First Solar, will prepare it to compete against producers of cheap Chinese solar modules with its even cheaper thin-film product.

2. Read the fine print

But here's the problem, says Axiom, in a write-up this morning on StreetInsider.com: Jumping from Series 4 to Series 6 requires that First Solar invest "~$900mn into an unproven technology." (My Foolish colleague Travis Hoium points out that it also requires First Solar customers to take a leap of faith in the new product, and ante up to buy it). And what will First Solar get from that investment?

Axiom's chief solar analyst Gordon Johnson explains that it's a function of simple math: Assume that two years from now, First Solar is manufacturing about 2 gigawatts' worth of solar modules at $0.22 per watt, and selling them for $0.28 per watt. (Axiom thinks that conventional solar modules from China, which cost a bit more than $0.30 per watt today, will sink below $0.30 wholesale two years from now.)

Axiom calculates that First Solar can earn about $124 million in annual gross profit from that business, and another $265 million from sales of entire solar systems. Throw in a bit more profit from project sales, and the analyst believes First Solar can gross perhaps $424 million annually. Building a 10% reduction in operating expenses into its model, Axiom emerges at the end with a best guess of $91 million in operating profit once the switchover to Series 6 is complete.

3. Solar math

And here's the real problem: $91 million in profit, divided among First Solar's 100 million shares outstanding (Axiom may be assuming some buybacks between now and then, because according to S&P Global Market Intelligence, First Solar had 10million shares outstanding at last report), works out to $0.91 per share in operating profit.

In comparison, Wall Street estimates are assuming that First Solar will earn $5.27 per share in 2020. (Again, S&P data differs, with the consensus number there shown is $4.75 per share.) Still, tom-ay-to, to-mah-to, the upshot is still this, says Axiom: What First Solar will actually earn a couple years from now "is a far cry from Consensus' 2020 non-GAAP EPS [estimate]."

Final thing: What to do now

As Axiom sees it, module prices (that already aren't so steep) are going to absolutely collapse once China implements its midyear cut in the feed-in tariff. That's going to be good news for big consumers of solar cells, such as Tesla. (It doesn't buy from First Solar, by the way. But according to S&P Global, the electric vehicle maker does buy solar components from a whole lot of cheaper producers -- in addition to producing panels itself.) That's going to prevent First Solar from charging a whole lot more to sell its new Series 6 modules than it costs the company to build them. It also makes the near-$1 billion cost of upgrading its facilities to produce Series 6 a probable money-loser.

Now, it is worth emphasizing at this point that Axiom's prediction is a real outlier from what the rest of Wall Street is predicting -- and perhaps for good reason. After all, while First Solar's operating profits did decline in 2016, the company did still book more than $340 million in operating income. First Solar hasn't had an operating profit year below $100 million in more than a decade, and has generally grown earnings from year to year, rather than shrunk them. Any prediction that the company will earn less two years from now than it earns today must be taken with a few grains of solar salt.

That said, this is the prediction Axiom is making. The analyst sees First Solar shares "fading" as investors' clue in to the approaching train wreck, and rates the stock a sell with a $21 price target. If Axiom is right about that, the stock could be cut nearly in half over the course of the next 12 months.