Recent Posts:

Solyndra Looking for $350M in High-Wire Financial Play

Michael Kanellos: August 12, 2008, 12:10 PM
Solyndra, the secretive developer of copper indium gallium selenide (CIGS) solar cells, is trying to raise $350 million in an effort to get into mass production, but the deal is causing some in Silicon Valley to shiver. The Fremont-based company, which raised $79 million in venture funds in 2007, wants the money to build a 420-megawatt solar plant, according to sources and documents reviewed by Greentech Media. Under the deal, Goldman Sachs, the lead bank on the deal, is trying to sell $120 million of convertible securities to existing investors and $230 million to newcomers. The convertible securities could be converted to Solyndra shares in the event of an IPO. Assuming an IPO can be pulled off, the owners of the securities would get Solyndra stock at a discount of the IPO price. If Solyndra eventually sells shares at $10, for instance, investors holding the convertible securities would be able to get shares for $8 dollars. The exact discount isn’t specified in the document, but sources say Goldman is telling investors that they will get a discount of 20 percent. But the discounts increase if Solyndra fails to hit certain deadlines. If the company does not file a registration statement to go public or undergo a fundamental change (i.e. get acquired) within twelve months of the issuance of the convertible securities, the security holders get another 10 percent discount. If the SEC has not declared a filed registration statement effective in 18 months, security holders also get a 10 percent discount. Thus, at the $10 hypothetical price, the holders of these securities would get shares at $7 or even $6. Solyndra, after all, could file a registration statement in 13 months (a ten percent penalty) and then get approval in 20 months (another ten percent.). And it’s not over yet. The holders of convertible securities earn 6 percent interest on their shares in the pre-IPO phase. The interest rate will be bumped by 1 percent if Solyndra misses its 12-month deadline and by 1 percent if it misses the 18-month deadline. The interest rate jumps to ten percent or the prevailing interest rate, whichever is greater, if an IPO or a fundamental change transaction has not occurred in 24 months. Solyndra did not return calls for comment. Other companies have used convertible funding to move into mass production, but the numbers and circumstances surrounding this one appear to push the stakes somewhat high. First, Solyndra hasn't sold any products yet. Second, it's got quite an unusual product: a cylindrical solar cell. By being cylindrical, the solar cell can gather direct sunlight, as well as convert light reflected from the roof into power. In theory, it sounds great. Taming the chemistry of CIGS, however, is not easy--most CIGS companies are not in production yet and some have faced delays. The vast majority of solar cell makers, CIGS and otherwise, are or are planning to build their solar cells on flat, planar surfaces, not cylinders. Some of Solyndra's early key employees--such as Benny Buller, Ratson Morad and Jonathan Michael--have left the company. Many VCs have begun to discuss a solar bubble and fear that many companies are overvalued. Combined, the leadings CIGS companies--a list that includes HelioVolt and Nanosolar--have already raised hundreds of millions of dollars. Several analysts (including some at Goldman) also expect the silicon shortage to ease next year, which could dent some of the appeal of CIGS by lowering the price of silicon solar cells. Many silicon solar cell manufacturers are expected to bring manufacturing capacity online over the next several years. So, in a nutshell, Goldman Sachs wants to raise over one third of a billion dollars on an experimental product that will enter an increasingly crowded market. Then again, Solyndra recently landed a contract to supply Solar Power with $325 million worth of solar cells between 2008 and 2012. Germany's Phoenix Solar also announced a deal to buy approximately $681 million worth of solar cells from Solyndra over the same period of time. That's over a billion in sales, assuming mass manufacturing can take off. Nonetheless, two VCs contacted by us deemed the deal risky and are declining to join. After the investment, Solyndra may be too expensive as an acquisition, noted one. The people who could be hurt the most in this, potentially, are employees because of the dilution that will occur when the new stock is issued. The additional discounts that the later investors may get if Solyndra fails to hit its deadlines will further depress the value of existing shares, said some.

Share a computer with a co-worker, get a rebate from utilities

Michael Kanellos: August 12, 2008, 4:10 AM
Virtualization--the ability to use a single server for multiple tasks--has taken off like crazy in the data center world as a way to both cut hardware costs and electricity costs. And in the near future, utilities may start pushing virtualization to cut desktop computing costs. That's the word from Mark Bramfitt, Principal Program Manager, Customer Energy Efficiency, at Pacific Gas and Electric. Mark is the guy that sets up and monitors programs at PG&E to cut computing costs. In these programs, PG&E effectively gives companies rebates to adopt energy efficient devices. (We're on a panel later today at the Flash Memory Summit.) PG&E, he said, may initiate a desktop virtualization program. Potentially, it could be quite popular. Garden variety data servers without virtualization software only utilize their processors about 15 percent of the time. With virtualization, a data center manager can crank that up into the 85 percent range and get more computing per watt. Desktops are even more profligate with energy. If you have a Windows computer, hit control-alt-delete and check your CPU utilization. You are probably only using 6 to 15 percent of the CPU's power. Thus, there is room to share. (I have a Mac--easily the worst computer I have ever owned in my life--and haven't figured out how to check CPU performance. But believe me, the time required in getting the smallest task done with thus thing means it guzzles the same or more than my old Acer.) Desktop virtualization would essentially be a form of thin client computing. You no longer would have a desktop. Instead, you and your co-workers would have thin clients that tap into a server in a computer room. Thin clients each consume about two watts and a server running ten of them would consume 160 watts. A desktop PC can consume 65 watts or more. Thus, a ten-unit thin client installation would consume 180 watts (ten two-watt thin clients and a server) while a ten unit PC fleet will consume 650 watts. PC power consumption will be This doesn't include monitors but that's equal for PCs and thin clients. Notebooks can rival and beat thin clients in power consumption, but they cost a lot more too. You could also set up a network where worker A has a PC on his or her desk and worker B has a thin client that shares the PC's CPU, memory and hard drive with A. Each worker would be separated by a virual partition so that they couldn't read each other's email. Thin clients have been the next big thing for, oh, a good fifteen years or so, but energy consumption may finally give them the oomph that they need for mass adoption. NComputing, a thin client start-up in Redwood City, already has 1 million seats deployed. They've only been selling computers for about two years. Pretty impressive. Bramfitt, though, said that not all programs set up by PG&E gain favor with IT managers. They have responded to programs to promote virtualization and new types of energy efficient air conditioners for server rooms, but not other programs. IT managers, he said, aren't quite used to working with utilties. Also, they don't pay the power bills so they still aren't completely tuned to cutting their utility costs.

Solar IP and Stealth Solar Roundup: Solyndra, Solexel and Netcrystal

Eric Wesoff: August 12, 2008, 1:00 AM

We read solar patents so you don’t have to…

While you were out carousing this past weekend, I was looking at some of Solyndra’s CIGS solar patents. Here’s what I found:

The patent for “Elongated PV Cells in Casings??? is authored by two former Solyndra employees, Ratson Morad, now with Daystar, and Benny Buller, now with CdTe thin-film leader First Solar (First Solar, incidentally, projects annual production of 1GW in 2009). The other listed inventors are Christian Gronet and Markus Beck, Chief Scientist at Solyndra. (And as I said yesterday: Why do patent holders and early employees leave a company like Solyndra?)

Here’s a link to a drawing of the cylindrical form factor from the patent document.  They also have some patent action in interconnecting and securing these very innovative and seriously non-standard solar units.

Since my evening was pretty much shot, I checked out patents by Solexel. Solexel is funded by Kleiner Perkins and Technology Partners. A while back, Daniel Englander and I did a little digging on the company and Kleiner responded by removing most mentions of the company from their Website. We looked into their technology but I think we got it a bit wrong.

These patents, invented by Mehrdad Moslehi describe a 3D thin-film solar cell (TFSC) fabricated with a reusable template sybstrate. (Drawings and photographs below.) And these filings describe a pyramidal 3D TFSC. Interesting stuff and it fits in with the MEMS theme we saw in their job postings.

Solexel’s advisory board includes Peter Peumans, an Assistant Professor of Electrical Engineering at Stanford, Deputy Director of the Center for Advanced Molecular Photovoltaics (CAMP) and an expert in solar cell modeling and characterization. Solexel has a CFO, Bob Komin, a post not usually filled in such an early stage firm.

Peumans is also co-founder and Chief Scientific Advisor of NetCrystal, a startup using a technology developed by Peumans’ group at Stanford. Bala Padmakumar is the CEO of NetCrystal, which is funded by Wellington Partners, Siemens, and X-Seed.

Bala is the principal investigator of a $99,000 Netcrystal SBIR Phase I project focused on the development of high-efficiency, lightweight, non-tracking, microconcentrator PV arrays based on stretched silicon. According to the SBIR document, “The stretchable silicon process can achieve accurate placement and electrical wiring of thousands of miniature solar cells in one parallel and potentially low-cost step.???

I’m scheduled to speak with Netcrystal’s CEO this week, providing he forgives me for the above link. We’ll get you more info then.