What will it take for investors to succeed?

A recent report published by the MIT Energy Initiative, entitled Venture Capital and Cleantech: The Wrong Model for Clean Energy Innovation, published July 2016, has stirred up a lot of renewed interest in clean energy investment models. It was followed by an article “Software Ate Cleantech: Now What?” written by one of the study’s authors. The study and article concluded that early clean technology investments failed for five key reasons:

  1. Long development cycles. Developing breakthrough materials, such as those for next-generation solar panels, can take 20 to 30 years.
  2. Significant capital requirements. To reach scale, venture capital money was used to build factories, sometimes before R&D was completed.
  3. Business model failures. Many companies faced expensive customer acquisition and long sales cycles.
  4. No sales premium. Energy companies were selling into established commodity markets where price (of electricity or of the solar panel itself) was the main driver.
  5. Few exit opportunities. Acquirers in the utility and industrial sectors were not eager to take risks or place a premium on a startup’s growth.”

Unfortunately, the study itself is deeply flawed in several regards:

1. It compares a teenager to two adults. The study compares compare investment results between CleanTech, BioTech and InfoTech over the period between 2006 and 2011, a period in which CleanTech investment was still nascent, but the other two categories were already over 30 years into the venture investment cycles. An appropriate comparison would have been to take the early hardware centric years of InfoTech (i.e. 1976 through 1981) or similar formative years (1986 through 1991) for the BioTech sector and compare them to CleanTech between 2006 and 2011. For those of us investing in IT and BT in those years, the results were far, far closer to what we have seen for CleanTech than the comparative periods used.

2. It ignores the development cycle of industries. The study either fails to comprehend or it just ignores the development cycle that new industries go through. First comes the hardware and capital centric phase (think personal computers, disc drives, monitors, and memory). Then comes the first overlay of application software that improves the functionality of the hardware. Then comes the networking phase in which hardware and software are connected to each other and function in unison. Last comes the “cloud” phase in which functions move to where they best operate in a seamlessly orchestrated system that ties together all of the foregoing elements. Go back and look at just how successful the overall VC community was in building personal computers and semiconductor hardware — not so much., by today’s standards.

3. It ignores the relevance of markets. In the early phases of an industry you don’t build innovative parts or software, you build the entire machine and the factory needed to produce it. At that point in the market, there is neither a supply chain nor a distribution channel nor a ready audience of buyers. That was as true of the auto industry and the personal computer industry as it was of the early phases of building wind turbines and solar panels. Until you had an active automotive market, you couldn’t make money designing anti-lock brakes or automatic transmissions. Until you had a vibrant computer market, it was hard to make money on subcomponents. But as these industries matured, they enabled an ever broader and ever more global range of innovators and entrepreneurs to build subcomponents that propel the industry forward. We are only now approaching that phase of CleanTech.

4. It ignores the differences between replacing incumbency and venturing into whitespace. One reason that computing, software and the Internet have been so successful for the venture capital model is these industries played into open whitespaces between existing industries rather than trying to compete with or replace existing industries head on. Biotechnology has been a far less fruitful pursuit for the venture model because it was producing drugs that, by and large, competed directly with those of the incumbent pharmaceutical industry.

5. It ignores the difference between consumer products and B-to-B products. Computers, software and drugs are largely purchased by us as individual consumers. Energy is largely purchased by utilities or large corporate buyers. Almost by definition, consumer markets change more rapidly than do B-to-B markets. Is it any surprise then that the three most successful CleanTech companies to date (Tesla, SolarCity and Nest) were all three consumer oriented businesses?

So did the Study get anything right? Can anything be gleaned from the study’s conclusions?

1. Long Development Cycles. Both BioTech and CleanTech suffer from long timelines, and InfoTech hardware did so much more than today’s software and internet-apps based businesses. By definition, investing in technologies that take 20-plus years to mature out of 10-year lived funds is difficult, if not impossible. Most of us have known that for some time, but LPs have been slow to adopt alternative investment vehicles. Until we do, this remains a valid constraint on producing desired investment results. Solving it means either accelerating the speed of inventive iteration (which is coming rapidly in energy) or changing the investment vehicle (which is being explored by several of the new initiatives).

Bottom line: Development cycles are shortening, technology is becoming more modular, contract manufacturing is more available, and iterative speed is improving — all of which mean this restraint is less important today.

2. Significant Capital Requirements. This was far more true of the early stages of InfoTech and will have been more true of the early phases of CleanTech. But today we have solar factories, wind factories and battery factories, so today’s clean energy innovations can improve processes in these factories without having to build an entirely new factory. Second, as more solar panels, wind turbines and batteries are deployed, a growing number of companies are finding software, communications and Internet tools that manage these hardware technologies more effectively — meaning the next generation of improvements are already less capital intensive, have far shorter development timelines and look more like the InfoTech investments of the mid-1990’s. Only those trying to build entirely new infrastructure or hardware modalities today must go through building out, proving and then scaling entirely new factories.

Bottom line: Because companies can develop drop-in improvements (think about 1366’s innovative way to produce silicon wafers, or Amprius and others work on anodes rather than complete batteries), capital requirements are falling. Players like Jabil, Foxconn and Flextronics have established large scale clean energy outsourced manufacturing capabilities. Existing hardware is being made more efficient and more effective through software. So the capital requirements are falling.

3. Business Model Failures. InfoTech and biotech primarily produce customer products. Energy has historically been bought by utilities and charged through to ratepayers. By definition, big-ticket B-to-B sales cycles are far longer than retail consumer sales cycles. But that was as true for companies selling nuclear plants, coal plants and gas turbines as it has been for wind and solar. The learning here that is relevant is the transition InfoTech made from mainframes to minicomputers to PCs and now back to servers comprising thousands of microprocessors. That is very much where the solar and battery industries are headed and their growth is increasingly being driven by retail customers (and large corporate buyers).

Bottom line: So here too the business model is evolving as it did in InfoTech, toward shorter sales cycles and more consumer-focused products. It isn’t happening overnight but the trend is highly parallel to that of IT and that should improve investment returns.

4. No Sales Premium. This is an excuse based on too few years of experience. Ask the oil and gas fracking companies whether they are getting paid a premium on their “technology-based” energy sales. The answer is no; and yet fracking is very much the case of a new energy technology rapidly overtaking the old, even at commodity prices. Now fracking relied heavily on government funding of its underlying technology and was 25 plus years in development when it “burst upon the scene” in 2005. But there is absolutely no reason the ongoing cost/price declines in wind, solar, batteries and electric vehicles shouldn’t continue to take market share from energy sources that cannot innovate at the same pace.

Bottom line: All industries commoditize. Over time software and business model innovations outstrip hardware innovations and reap greater returns per dollar of capital invested. It is all part of how industries change and absorb innovation. What is surprising is how short our memories are. Most InfoTech entrepreneurs (and the authors of the MIT study) weren’t in business when the hardware phase of the InfoTech industry was being built. But without those predecessors who carried the heavy loads, the current It industry would not exist. So too, the energy industry of our future is being built on the shoulders of those early CleanTech failures.

5. Few Exit Opportunities. Exits reflect success. So it is hard to complain there are few exits when almost nothing is finished. Even Tesla, the most celebrated “CleanTech” success company is far from being fully grown up. Practically speaking, this criticism is calling the game at half time.

Bottom line: There will be big clean energy winners and they will affect the pricing and premiums paid for others in the sector. What is true is that technology is trying to disrupt an industry with a 100-year change cycle history, we have in less than 50 years reduced those change cycles by half. If we can now reduce that 50 years by half again, then the future of clean energy will be bright.

Unfortunately, the study and articles published since then by its authors drew further conclusions poorly supported:

1. “We hope that our findings will dispel the notion that the climate change can be solved by simply replacing VC money with patient capital.” We’re not sure who asserted that patience alone would solve CleanTech’s problems, but the authors have done little to indicate precisely what they think will address their concerns.

2. “New private and public funding sources may be able to better support revolutionary technologies.” New money won’t fare any better than old without the benefit of the forces already at work we have discussed above. Maybe more money would help, but there is no indication that the total volume of money is today greater than it was in 2005–2010, if anything it is smaller. Patience may help, but patience requires money to keep companies alive and it kills IRRs. So something more than what the authors have offered will be required.

3. “A new wave of public and private support will be required to reboot investment in cleantech:” Although this conclusion is otherwise wholly unsupported, it is a truism in this sense: If the public support drives CleanTech prices below conventional energy prices, then CleanTech can win on pure cost and a market will exist (for as long as the subsidy exists). Germany, Spain, et al have proven this point, but so what. Basically the authors conclude that if you want CleanTech to grow you have to subsidize it heavily and wait — that is not a helpful or very insightful conclusion. They say virtually nothing about what new money should do to create different results — other than a) have a lot of money and b) have a lot of patience…

4. “But capital is only part of the equation. Cleantech companies require much of the same support as other types startups: help with developing sales strategies, hiring the right key employees, marketing and fundraising. But they also require some things a SaaS company doesn’t: lab space, manufacturing expertise, and — particularly important when selling to risk-averse adopters — opportunities to test and demonstrate their technology.” Most veteran clean energy entrepreneurs will tell you this last criterion, the ability to test, deploy and obtain “bankability” for their products was and remains the most critical factor in achieving ultimate success. Yet, of all the new efforts and initiatives being discussed in the clean energy arena, virtually none are focused on this critical stage.

5. “To succeed, cleantech founders will have to run lean. First, they will have to take advantage of incubators and accelerators and will need to fund R&D through public grant funding. Later, they will need to rely on debt for projects (such as Black Coral and Generate Capital) and manufacturing (including private working capital revolving funds and public sources such as OPIC) as an alternative to equity financing. Throughout the process, cleantech founders should be in close contact with corporate and strategic partners who will be not only the likely customer for the product, but also potential acquirers of the company.” Here too the implication is that the firtst generation of cleantech entrepreneurs and investors were profligate. They were not, they simply had to build the entire ecosystem for themselves. Today’s entrepreneurs have the fortune of turning to a more mature industry and to a far greater number of different players for support. Not using that support is stupid. But to assume it was always there is just false.

Interestingly, one of the commentators on these articles is Jonathan Silver, the former Executive Director of the U.S. Department of Energy’s Loan Programs Office. Jonathan makes several meaningful additions to the dialog:

While the difficult history of early investments in clean energy is well known, the extrapolation to the ideas in this piece are curious.
First, the field is actually “clean economy”, not “clean tech” and includes energy efficiency, remediation, sustainability and other “footprint”plays. …
Second, only power related investments suffer from most constraints identified here. Power generation investments are capital intensive, illiquid and generate bounded returns. They always have and perhaps the vc’s who got into generation should have known better. But, many parts of the clean economy can generate outsized returns. The paucity of investments the authors identify is driven much more by the echo chamber between GPs and LPs, which goes hot and cold, often at the wrong time, about a sector. The clean economy is growing faster than almost any other set of industries in the world.
Finally, I wish we could rid ourselves of the bias towards innovation as the only way to solve problems. Most solutions are incremental in nature. Most do not require that entire industries be disrupted. If we spent the next ten years aggressively deploying the renewables technologies we already have, we would move much, much closer to our decarbonization goals. It is arrogant to believe that only disrupters [can solve] these problems.”

Another commentator correctly writes:

The vast majority of all the new generation in the whole US is zero carbon. China is actually reducing its coal use decades ahead of when they said they would and deploying more wind & solar each year than most countries have total generating capacity. In general, around the world, wind and solar are both expanding exponentially. Electricity is finally coming to the poorest billion because of solar + batteries + LEDs in a way that 60 years of “electrification efforts” have utterly failed to do. EVs have been slow to catch on but are now racing ahead so that we can clean up both electricity generation and transportation.”

Stephan Dolezalek is one of the founders of Resourcient.org, an organization focused on revitalizing investment in the clean energy industry and continuing the quest to battle climate change.