Monday, December 19, 2011

Warren Buffet Goes Long Solar Energy

"Who's to say that your mind won't change
Yesterday's excuses just aren't relevant today
It may take time it may take a strong will
But we don't have to swallow such a bitter pill
Better to have tried
Better to have learned
Better to take that risk"

- In My Eyes, "Take the Risk"

For the second time in two weeks, MidAmerican Energy Holdings, a subsidiary of Warren Buffett's Berkshire Hathaway, has acquired an interest in a large photovoltaic solar farm project when it announced the acquisition of a 49% stake in the Agua Caliente Solar Project, a 290 MW photovoltaic project in Yuma County, Arizona from NRG Energy. The first deal, announced on December 9, was the acquisition of the Topaz Solar Farm, a 550MW photovoltaic project in San Luis Obispo County, California, from First Solar. Both projects are being built by First Solar, and both will sell their power to Pacific Gas and Electric through a 25 year power purchase agreement (PPA), which will help California achieve its mandate of generating 33% of its power from renewable sources by 2020. The major difference between the two projects was that the Topaz Solar Farm was not backed by a U.S. Dept. of Energy (DOE) loan as First Solar was unable to close the loan guarantee by the September 30, 2011 deadline, whereas Agua Caliente Solar Project is backed by a $967 million DOE loan guarantee.

Given the horrible performance of photovoltaic (PV) stocks this year, with the universe of stocks down about 75% year to date, many people are asking what does Warren Buffett see that's so attractive in solar technology that the stock market doesn't see?

To start with, MidAmerican acquired a solar PV farm, not a PV cell or module maker: falling PV panel prices are good for the power producers and bad for the PV panel manufacturers. Since both the Agua Caliente and Topaz projects negotiated their PPAs before the collapse in PV panel prices, the falling prices would be a benefit to the owner of both projects. Second, it is likely that MidAmerican was able to negotiate transaction terms that were very favorable. First Solar has been struggling financially and has been seeking buyers or investors in its large scale projects for the last several months, and had previously sold its 660 MW Desert Sunlight project to NextEra Energy Resources and General Electric, its 280 MW Antelope Valley project to Exelon Generation, and the Agua Caliente project to NRG Energy in August. When First Solar lost the $1.9 billion DOE loan backing on the Topaz Solar Farm project, it needed to find an investor quickly to keep the project on schedule and minimize its own financial impact. The estimated cost of the Topaz project was over $2 billion, and it would have been difficult for First Solar to finance given the cratering of its stock price and the fact that it's profitability is under pressure given the brutal price pressure in the PV panel market. Although the financial terms were not disclosed, one would logically assume that a smart buyer and an increasingly motivated seller would negotiate a deal that had some favorable features for the buyer.

By closing the deal in 2011, MidAmerican gets to take full advantage of existing tax credits and accelerated depreciation that the project qualifies for, including the U.S. Treasury Grant Rebate program which returns 30% of the system's cost to the owner regardless of the earnings and tax liabilities, and which was set to expire at the end of 2011. That represents at least a $600 million benefit to MidAmerican which significantly improves the IRR of the investment. Given the facility's guaranteed energy output of at least 1,066 GW-hrs annually, PG&E's favorable PPA terms, and the very significant tax credits and accelerated depreciation benefits, a rough estimate of the unlevered IRR for the Topaz project indicates it is >15% for a very low risk endeavor, a heck of lot more than MidAmerican could get from any other equivalent risk investment.

A final factor may be that investment tax credits for wind power are set to expire at the end of 2012, and MidAmerican owns several wind power facilities, whereas solar power ITC's will not expire until 2016. Entering the solar power market today will provide a new tranche of ITC's with a longer horizon to replace the expiring wind power subsidies at the end of next year.

In time, I strongly believe that photovoltaic electricity generation will continue to grow as a percent of total electricity production as its costs continue to drop and approach grid parity with other technologies. It will exist in both centralized (e.g. solar farms) and distributed (e.g. rooftop) configurations, and costs will decline as PV cell efficiency improves, secondary efficiency enhancement products appear (for example total internal reflection distributed optics systems), more efficient power conversion technologies are developed (e.g. microinverters), PV panels are integrated with home structures and building materials, and economies of scale develop as the installed base grows. Other ecosystem improvements must come along with it, such as a more flexible, intelligent grid to handle varying load conditions as solar electricity becomes a larger factor in total power generation, but these areas are attracting significant investment capital and are moving rapidly as well.

Solar electricity will survive and become an integral part of our long term, clean renewable energy strategy. The market factors that are causing the pain felt by PV panel makers will eventually resolve themselves - PV panels are a cyclical commodity business very much like DRAMS, as I have pointed out. There will eventually be a shakeout, as a cyclical commodity industry cannot support too many competitors, and market share will consolidate among the top 4 or 5 major PV panel producers. But the industry will continue to grow and will provide opportunities for innovative downstream companies, which is where the best investment opportunities exist. That is where Warren Buffett chose to invest.

Warren Buffet is famous for recognizing undervalued companies with significant intrinsic value that can generate above-market returns over time. It looks like MidAmerican's recent foray into the PV solar power generation market is an opportunistic move to capitalize on some motivated sellers enhanced by some attractive cash flow generation with very little technical, execution or market risk. It is a validation of the PV solar industry as a source of utility scale power generation, at least with the existing state and federal financial incentives, but it is not a statement of any sort about the current state of the PV module industry, which is still struggling with crushing excess capacity of polysilicon and PV modules. However, Buffett is also known for investing in companies with long term sustainable value - it is unlikely that MidAmerican bought into the solar farm business as a short term financial opportunity driven by outsized tax incentives - it is likely that underpinning the opportunistic investment is a long term belief in the viability of solar electricity as a sustainable growth, profitable business. And that's good news for the solar energy industry.

Sunday, November 06, 2011

Where is the Value in the Solar Cell industry? (Hint: It's not the cells)

A few weeks ago it was announced that First Solar's CEO, Rob Gillette, left the company abruptly. This is never a good sign. It usually means that either 1) something really negative was unearthed at the company, or 2) the board and the CEO had a major falling out that was so serious that the board was forced to dismiss him immediately, which is usually over something that fits into 1) First Solar (Nasdaq: FSLR) dropped 25%  in response, and is down 75% from its high of $170 earlier this year. First Solar has been under pressure for the last several months because of the free fall in PV panel prices this year, which have dropped over 40% due to plummeting polysilicon prices and excess inventories of PV panels being dumped in the market. First Solar, which makes CdTe thin film PV panels, which operate at lower efficiency (13-14%) than silicon PV cells (20-22%) but are less expensive to manufacture and install, are directly affected by price changes in the silicon PV market, and has had to cut the prices of its panels and seen its profit margins decline just like the silicon PV companies.

The massive price declines in the PV panel industry after the huge increases during the polysilicon shortages of 2007 - 2008 got me thinking about the relative value in the PV marketplace and in the supply chain. The analogy that struck me as that PV cells themselves are like DRAMs in the semiconductor industry; the PV panels are like memory modules, while the PV panel array is more like an electronics system. You could compare the PV array to a PC, or a disk drive array, or a solid state storage device, it really doesn't matter. The point is, the PV cell is a commodity product whose price is largely dependent on supply-demand imbalances and there is very little value derived from intellectual property. Sure, some companies might be able to extract a small premium for slightly better efficiency, or some novel form factor that makes the panels easier to fabricate and install, just like DRAM companies that offered a special organization, or were the first out with a particular package or module configuration would get a small premium in the DRAM market, but there is a lot of historical data on DRAM prices (and DRAM company profit margins) that shows how volatile and cyclical they are, and the volatility is driven mainly by supply-demand imbalances.

If you take the analogy one step further, think about who made the most profit in the electronic systems that incorporate DRAMs as their major memory component. Putting aside Intel's effective monopoly in the microprocessor market, because there is no analog to Intel in the PV industry, the companies that have captured the highest profit margins in the electronic systems markets have been the companies that added intelligence, communications, and specialized processing capabilities to increase the performance and functionality of the systems. It was the companies that developed PCI Express, SATA, InfiniBand, and FiberChannel, that developed WiFi and LTE wireless networking, that developed 10G and 40G Ethernet controllers, or 3D graphics accelerators. I think that the companies that will eventually capture the most value and have the most successful business models in the PV industry will be the companies that add intelligence and improved performance to PV arrays.

One of the areas where companies are adding value to the PV industry are microinverters and panel optimization devices, which change the architecture of the PV array from serially connected (think Christmas tree light strands) to parallel connected, which can compensate for variations in individual panel output caused by clouds, dirt, temperature, or other environmental variations, to optimize power harvesting and operate the array at is maximum power point (MPP) even as the operating environment changes. The economic payback derives from the enhanced power output from the PV array over time, which more than pays back the incremental cost of the additional panel electronics. Companies that provide microinverter solutions include Enphase, a pioneer of microinverters that is in registration for its IPO, and several venture-backed private companies such as SolarBridge, Enecsys, Involar, Direct Grid, and ArrayPower. Microinverters convert each panel's DC power output to grid quality AC power, compared to the legacy method of using a single inverter to convert the DC power output of the entire PV array to AC power at the connection to the grid, known as a grid-tie inverter. Performing the DC-AC conversion at the individual panel level provides greater flexibility in optimizing the power output of the entire array even if some of the panels are not operating at peak levels. Some microinverter companies like Enecsys incorporate wireless and wired communications into their solution to provide data collection from individual panels, enabling remote monitoring and troubleshooting to add further value to the array owner or operator.

There are several venture-backed private companies developing panel optimization solutions including SolarEdge, Tigo Energy, Azuray, and eIQ Energy. Panel optimizers do not perform DC-AC conversion at the panel, but rather perform DC voltage conversion to an optimum DC level so that the grid-tie inverter can efficiently convert the array DC output to grid-compatible AC power. Panel optimizers track and optimize the output from each panel in the array and thereby adaptively optimize the maximum power point (MPP) of the entire array, even with rapidly changing conditions like variable cloud cover. The panel optimizer is integrated into the panel or installed at the panel, and often features remote panel monitoring and communications to enable the array operator to monitor, troubleshoot, and control panel performance in real time.

These companies are generally early stage except for Enphase, and are establishing the market for their products. But if their solutions provide the performance and cost benefits they claim, they should be able to increase their penetration of the PV market and command pricing and margins that reflect the uniqueness and benefits their products provide.

Another area of added value for PV panels are performance enhancement products such as tracking systems and lenses for PV concentrator systems. Several companies are developing novel low cost concentrator optics that can be integrated into a PV panel and create higher power output levels or comparable output levels using fewer PV cells, thereby lowering the levelized cost of energy (LCOE). Some of the companies developing these products are Solaria, Banyan Energy, Skyline Solar, SolFocus, Energy Innovations, Morgan Solar, Covalent Solar, Prism Solar, and Stellaris. Again, the opportunity is to add value through increased power output and lower cost of power by integrating an innovative feature onto the PV panel or PV array that adds a relatively small incremental cost.

The point is that there are products being developed that can improve the efficiency and power output from a PV array that are separate from the PV array. They are symbiotic with the PV array, and may actually benefit from falling PV panel prices since that ultimately helps drive PV electricity closer to parity with grid electricity and expand the PV installed base. As long as the products provide significant benefits and can be improved to maintain competitive differentiation, they can capture value and remain somewhat insulated from PV panel price swings. The innovator can extract reasonable and defensible profit margins from the PV array "system" even if the cost of the PV panels themselves continue to decline. From an investor standpoint, these are the types of solutions that are worth considering and have the potential to generate outsized returns. These may be the companies that become the next Linear Technology of the PV industry rather than the next Micron.

Monday, October 17, 2011

Intel Leaving the DTV Market

"Change it had to come
We knew it all along
We were liberated from the fall that's all
But the world looks just the same
And history ain't changed
'Cause the banners, they all flown in the last war"

- The Who, "Won't Get Fooled Again"

Intel's recent announcement that it is leaving the digital TV (DTV) market is the second major company exiting the DTV business in the last month - Broadcom recently announced it is shutting down its DTV and Blu-ray business units as well. There really hasn't been much good news in the DTV business for the last few years. Trident Microystems, once a leader in the category, ranked 3rd in global DTV SoC sales in Q1 2011 with only 7% of total revenue according to iSuppli (MStar was #1 with 37%, Mediatek #2 with 9%), and that was after acquiring the DTV and STB businesses from both Micronas and NXP; Trident recently announced a 20% headcount reduction to reduce expense amid a continued decline in its business. STMicro acquired Genesis Microchip in 2008, which was also a former category leader; STMicro ranked 5th with 2% of total sales.

The trend is obvious: DTV chipsets, which are extremely complex SoCs containing multi-standard HD digital audio/video decoders, video image enhancement processors, DTV demodulators, high speed/high precision video A/D converters, and many different interfaces, have become a commodity and market share is consolidating in Taiwan, which has built an industry on taking over high volume markets by competing on price. The consumer, of course, benefits from this: a recent scan of advertised prices for 55 inch, LED backlit, 120Hz/240Hz 1080p HDTVs showed prices in the $1500 - $2000 range for premium brands like Sony, Samsung and Sharp. It's truly amazing how much value and functionality is available for so little, but that has been the calling card of the semiconductor industry - giving its customers more and more for less and less.

What may also be driving companies like Intel and Broadcom out of the DTV market is the TV's diminishing role in the future vision of the all digital, connected, anytime, anyplace media experience. The large screen, high definition flat panel HDTV will still be the anchor viewing device that sits in the family room with the requisite couch and bowl of popcorn for social or family viewing events like sports, movies, award shows or the continually devolving array of reality shows, and it will increasingly be connected to the Internet, but it will remain primarily a passive viewing device. Many attempts and painful lessons have been learned in trying to force interactive TV down the unwilling throats of the viewing public. No, people generally don't want to interact with their TVs, they want to be entertained by them.

Intel said they would shift their focus to the IP set-top box (STB) and gateway media processor business, which is being folded into its tablet organization, which is focused on tablets, smartphones, and ultrabooks. Gateways and next-generation STB's are becoming more intelligent, more complex, and are where the action will be in the connected digital home. And where change and complexity are the greatest is where the opportunity is for companies to develop new solutions that can be sold at acceptable margin.

Consumers, content providers, and network operators are all scrambling to protect their existing content, networks, and advertising and PayTV revenue while making their programming content available to viewers on any platform at any time. TV Everywhere, introduced last year by Comcast and Time Warner, is essentially an effort to provide content from the cable and satellite operators' portfolio of highly rated channels and programming content through the Internet to Web-connected devices like notebook PCs, tablets and smartphones. The catch is that you must be a subscriber to the cable or satellite provider's network in order to access the content. "Free" Web content providers like Hulu and Clicker are limited in the content that they have access to, and based on some of Hulu's well know financial difficulties seem to be inevitably heading toward moving their content behind a pay wall. Premium content, which wins high viewer ratings, generates high advertising rates, and draws subscribers, isn't cheap to create and isn't going to be free to watch, at least not for much longer.

The network operators and content providers have a mutual beneficial arrangement that provides the incentive to remain joined at the hip: the network operators need to have a complete package of highly rated channels to offer in order to attract and retain subscribers; the channel owners receive about $32 billion dollars annually in affiliate fees paid to them by the network operators for the privilege of offering their channels and programming content. That's a pretty strong reason to keep their alliance intact even while providing the content through time and space shifted Internet distribution.

So what does this have to do with semiconductors and Intel and Broadcom exiting the DTV business? Semiconductor companies are struggling to generate revenue growth and preserve profit margins in a business that is increasingly seeing compressed product cycles (and profit cycles) as these extremely complex SoC's are becoming commoditized faster and faster. The next major evolution in the DTV business is likely to be in the STB or broadband gateway, where HDTV content will be received from both broadcast and Internet sources and distributed over home IP networks on different physical media (WiFi, coax, powerlline, CAT5 wiring) and stored and viewed on any type of platform: DTV's, STB's, DVR's, PC's, tablets, smartphones, gaming consoles, etc. By the way, there will be multiple viewing platforms in the home, and the PayTV content can only be viewed if you are a subscriber, and the viewing statistics need to be captured for rating systems and advertising rates, so it will be necessary to provide absolutely watertight security on these multiple streams of content. This may be the tie in to Intel focusing its efforts in STB and gateway media processors and moving the business into the group that is responsible for tablet, smartphone and Ultrabook technology.

The next generation of gateway's and STB's will need to be much more intelligent, provide multiple HDTV stream processing, and offer much more stringent security than today's systems. This capability will come from the semiconductor SoC suppliers, and this is the opportunity they will be maneuvering to win. The distributed digital media era is upon us. Amid all the change, one thing is constant - the major production studios and network operators will maintain control of the content. As the Who wrote it in their song, "Meet the new boss, same as the old boss."