Carl W Nelson
Carl Nelson Consulting, 1325 18th St NW Unit 803, Washington, DC 20036
e-mail:
carl@carl-nelson.com (updated Jun 2013)
ABSTRACT
Small young companies can commercialize photonics. Many are already doing so and many more are getting started with a small government subsidy from Small Business Innovation Research. Capital is gushing, even begging for good investment opportunities, for companies ready to commit themselves to commercial success.
Why should big companies have all the fun and profit? Little companies, too, can compete for photonics fun and profit, as long as they are not too anxious to make the profit right away. One may even grow into the next generation's Intel.
Small entrepreneurial companies, more agile than the giants of their time, created America's information technology industry. They started industries with little help from government which favored the giants' mainframe mindset. The computer giants saw their own markets collapse as mainframe market fell from 100% in 1977 to 1% in 1987, says Gilder1 ,as upstart new companies brought in the PC. Microsoft, Intel, and Apple became household names after Xerox sowed and abandoned the seeds of new industries: laser printer, desk-top, and mouse. A few giants (Motorola, Hewlett-Packard, AT&T) then invested the massive capital to furnish the new industry with an infrastructure. And even though a few giants do pioneering research in photonics, their hurdle rates and business size threshold discourage them from bringing the innovations to market.
A few photonics high tech innovators recently have been helped by a subsidy that the most entrepreneurial used to penetrate commercial markets as well as give a direct benefit to the sponsoring government agency. Although the entrepreneurs provided the inspiration, the Small Business Innovation Research (SBIR) program furnished free early development money. The companies highlighted in this paper, who got their SBIR from the Ballistic Missile Defense Organization (BMDO), have shown that Yes, Small Companies Can Commercialize Photonics.
Many companies have found a success path through BMDO's SBIR. Success, of course, means different things at different stages of industry development. Photonics still has a tenuous hold on information technology mainly in fiber optic telecommunication. And until the economics shifts (much cheaper photonics gizmos), electronics will be hard to displace. Gilder's paradigm of "waste bandwidth and save watts" will open opportunities for the nimblest of these companies.
Ortel Corp (Alhambra, CA) recently declared another nice quarterly profit and acquired an interest in a Chinese enterprise which already distributes Ortel products. And in Beijing exhibition in November 1996 it revealed a line of wireless microcell and repeater products. Ortel, the creature of CalTech's Amnon Yariv, had a little SBIR money and healthy profits before going public in 1995. Since then it has maintained its profits and market capitalization was $300M(illion) in mid-December 1996. In perspective, Ortel's SBIR was dwarfed by its profits as a source of R&D finance.
SDL Inc (San Jose, CA) started winning SBIRs as soon as it qualified by reorganizing as an independent company in 1991. Its laser diodes and other products have led to growth that doubled its employees and yielded multi-million dollar profits after going public in 1995. Its market cap is $250M in mid-December 1996.
,STRONG>Cree Research (Durham, NC ) started with SBIR's providing about 20 percent of the initial capital for its silicon carbide process.Cree turned that investment into a fledgling blue Light Emitting Diode after which it went public (1993) and found itself in a world-wide blue LED war for what Hewlett-Packard estimates is a $2B market (and even more for blue lasers). In 1996 Toshiba announced a "more advanced version" of Nichia Chemical Industries's GaN-based laser2. DOD helped with $3.9M in contracts for improvement to get to "world-class high-power and high-frequency transistors". Shin-Etsu Handotai and Sumitomo paid $2.7M for licenses to Cree's LED process for which Cree will still supply the bare SiC.3 Still, war is expensive and Cree runs near zero profit on recurring sales.
Micracor Inc (Acton, MA), a venture-backed spin-off from MIT Lincoln Lab, is moving toward production for several markets. Founder Aram Mooradian says Micracor can produce 500 mW of circularly symmetric, diffraction limited output directly from a semiconductor laser in wavelengths 980 nm to 1010 nm with over 20% electrical-to-optical conversion efficiency. The low divergence (10 mradian) output needed only a simple, small focusing lens to couple the output into a single mode fiber with 90% efficiency. BMDO SBIR supplemented the starting capital and then matched funds from prospective customers for a commercially convincing demonstration.
The micro-machined opto-electronics resonator by IntelliSense Corp (Wilmington, MA) should lead to a tiny gyroscope, less than 5 cc volume (like a pencil eraser for those who still remember pencils) and ten grams weight, that would last 100,000 hours (about 12 years, which is important for military munitions that sit patiently on shelves waiting for a war). GEC-Marconi Electronic Systems Corp teamed with founder Fariborz Maseeh for the SBIR and beyond. Market studies have estimated a $2B+ market for micro-machined inertial sensors in 2000.
paint a="" few="" lines="" on="" surface="" overcoat="" them="" magic="" polymer,="" activate="" circuit="" in="" silicon="" with="" less="" than="" 4v="" and="" voila!-="" the="" polymer="" shines.="">UNIAX Corp (Santa Barbara, CA) owns the shiny polymer with which it reported the Light Emitting Electrochemical Cell in Science in 1995. UNIAX sees this set-up pushing aside two presently pursued technologies: light emitting epitaxial layers like AlGaAs/GaAs, and porous silicon. Since then it has attracted well over $1M beyond the SBIR subsidy plus a new $1M from a large chemical company and a friendly note in Business Week.
Founder Sadeg Faris of Reveo Inc (Hawthorne, NY), an escapee from an American corporate giant, started a company that grew to 80 employees, and then he started another (this time profitable) that is growing rapidly with $3M of BMDO SBIR in optical storage and displays exploiting polarization. It now has 45 employees and has spun out a subsidiary, Vrex Inc, to mass-market a $69 3-D stereo display product for TV sets.
Vixel Corp (Broomfield, CO) started as four people in 1990, won a BMDO SBIR for a VCSEL technology, raised several millions including $11M in late 1995 (from a consortium of venture capital, an electronics firm, a university, and a well-known musician), and now claims to sell the only commercially available VCSEL product. Each of the two succeeding SBIRs had more matching funds than the one before.
GEMFIRE (Palo Alto, CA) is a separately capitalized company to exploit Deacon Research's polymeric thin panels for low-cost direct-view, 100-inch displays for a $100B market. Laser light is guided to the display pixels in a thin film on a plastic substrate, using an array of optical waveguide switches. The display will be rugged, flexible, paper-thin and lightweight, bright, efficient, wide viewing angle, and even transparent. (Passengers on a United 777 can watch a neighbor's video.) GEMFIRE will make display systems for high cost, low volume use and then license the technology to a few strong partner companies. (Many SBIR companies dream of profit-laden high-end products, if they've thought about the market structure.) Deacon had several BMDO SBIRs with a growing match ratio. Market acceptance will depend on unit price which must substantially undercut its closest competitor even with a technical advantage of wide viewing angle and large area in a lighted room.
Optivision (Palo Alto, CA), started by three academics and now grown to 100 employees, sells compression and networking products. It claims first real-time MPEG compression at 10% of the previous price. At its first BMDO SBIR in 1987 it had six employees.
Northeast Photosciences (Hollis, NH) helped found a separately capitalized manufacturing company, HOLOS Corp (Fitzwilliam, NH), to produce products from Northeast's hologram technology. New developments for laser beam filters marketed by Ealing Electro-Optics and optical metrology will provide subsequent products. BMDO SBIR contributed about $2M.
SBN Corp (a pseudonym) was started by an entrepreneur who left a top-twenty SBIR winner. Its five SBIRs (three from BMDO) and private investment are moving it to stability if it can drive down the cost of its sophisticated photonic devices. It will market a line of products in 1997, especially tunable lasers and filters for Wavelength Division Multiplexing devices.
The only firm of the top twenty SBIR winners (at least $10M each) to grow its market cap and become a big success story from six employees in 1987 is Advanced Technology Materials Inc (Danbury, CT). Its mid-December 1996 market cap is $180M.
Other companies pursuing photonic products with BMDO SBIR and matching private finance are Optigain (Peace Dale, RI), Optical Concepts (Lompoc, CA) now a subsidiary of WL Gore, TACAN Corp (Carlsbad, CA), Berkeley Optics (Berkeley, CA), NZ Applied Technology (Woburn, MA), Holoplex (Pasadena, CA), and Templex Technology (Eugene, OR).
New BMDO wagers on photonics include fiber-to-fiber connection, optical logic gate, wavelength division multiplexing, VCSELs, all-optical router, asynchronous logic, tunable diode lasers, E-O switch, blue laser, MEMS, fiber lasers, inter-chip connection, gallium nitride films, wavelength tolerant laser receiver, transparent transistors, anti-reflection coatings, conducting polymer films, optical clock signals, and multilayer mass storage.
Few of these firms and products are yet profitable. Their investors are still living on hope while information technology economics move their way. SBIR has carried and will carry the concepts until the product and company can compete in the open market. Some of these companies move from BMDO SBIR infancy support to other agency support in a prolonging of the development and of profitability. And the more the SBIR, the worse the future profit prospects.
Small R&D companies, photonic or otherwise, can compete for $1B(illion) per year from the SBIR program. A project funded typically gets under $1M over three years in two phases: Phase 1 under $100,000 to study the feasibility of a new concept and Phase 2 normally under $1M to prototype it. After SBIR the government retains only a right to royalty-free use for government purposes. No repayment, no equity. BMDO applied a flexible funding formula to its SBIR, especially since 1991, of conditioning Phase 2 funding on private sector investment. BMDO's contributions to each ranged up to $2.5M..
Why does SBIR exist? The political favorite - market failure, the idea that the free market will not develop high risk new technology in small business because the Return On Investment is too low and /or the risk too high. Congress in 1982, and again in 1992, accepted the small business complaint of being underused in federal R&D and consequently being unable to help America compete in the high tech global market. Unfortunately, except for a few pockets of commercial success, the situation has not much improved after thirteen years of SBIR. The small businesses get about the same fraction of federal R&D as they did in 1982 and no evidence has yet been presented to show that the beneficiaries have fared any better than inventive companies not getting SBIR. (The beneficiaries who have learned to win SBIR awards would leap to disagree although two words should squash their claim - Bill Gates.)
Some companies find themselves trapped between BMDO's commercial SBIR competition and the ordinary SBIR programs who prefer predictable R&D results. If the companies cater to the other agencies' demands they will never bring a product to market. If they start down BMDO's road, they may fail to attract the capital. On balance, they are better served to discover the dead-end early.
The BMDO SBIR can make a bolder claim - that the beneficiaries have commercialized their new technologies and that the capital was as efficiently spent as possible with government.
Most government programs, even those aiming for economic gains, have trouble finding credible metrics beyond job creation (in the Member's district and before the next election) which lasts only as long as the money. Permanent value comes from net growth in permanent jobs. (Note that net jobs is more elusive to measure than politicians will admit as jobs in new technology kill jobs in any replaced technology.) Although no decent evaluation has ever been done of SBIR, BMDO's SBIR can offer some econometrics. (The General Accounting Office in 1991 did survey SBIR beneficiaries in a pseudo-evaluation with no control population and no independent confirmation of the data supplied.)
Thirteen firms with at least one BMDO Phase 2 award went public between 1992 and 1996. Seven won their first Phase 2 from BMDO when their median size was eight employees. Their total market cap grew from about $800M at IPO to $1800M in late 1996. (Market cap is the number of outstanding shares times the NASDAQ trading price.)
Ten more firms say they will go public to raise about $150M "when market conditions are right", an easier claim to make than to actuate. More firms will soon reach the point where their capital needs (and the itchiness of their investors to cash out) will push them toward the public market. Although going public wasn't BMDO's goal for any firm, it serves BMDO's interests by accrediting the technology and the firm. Being public also admits market cap as an SBIR evaluation measure and imposes a business discipline on the firms.
Private capital investment, a convincing indicator of future economic activity, is pouring into BMDO SBIR projects. By late 1996 $100M was committed, spent, or planned for 100 projects for $100M of Phase 2 SBIR starts since 1992. Figure 1 shows the accelerating private spending 1993-1996. Few projects now get an unmatched subsidy, whereas all got it before 1991. (Most other SBIR awards across government do not require matching.)
Such tallies must understate actual investment since collecting exhaustive data on private capital would have free-market government intrude on private business data. For SBIR competition purposes, the government need only discover enough capital commitment to decide which proposers and ideas have the best commercial prospects. The firms actually moving towards a post-SBIR market are perforce deploying more capital than government can count.
Many of these SBIR companies found the matching money because BMDO created the incentive - get it or get gone. If the capital had not come forth, BMDO would have had to climb down and pass out the money as free subsidies. Competition prevailed, however, and BMDO's strategy worked to prove that photonics companies can find capital.
"We're a market-driven company", say many companies to convince government and themselves of their devotion to commercialization and therefore worthy of SBIR. Easy to say. Typical SBIR companies lack the time imperative of market-driven innovators who must catch a market wave before the opportunity evaporates, especially in 1990s information technology. Says Idealab creator Gross of the Internet, "If a company can't go from concept to launch in nine months, it's not going to make it."4 The market-driven firms must productize quicker than government programs can feed them money, especially the SBIR two-step. Most SBIR companies like the better mousetrap myth with a market magically appearing on the day development ends. Not so easy. (Note: about 5000 mouse traps have been patented.)
To attract true market-driven companies the Department of Defense (DOD) started a Fast Track SBIR sub-program wherein DOD matches third party cash at a rate that depends on the company's SBIR maturity. The youngest SBIR companies get 4:1 matching, the maturest get 1:1, and all else get 2:1. Youngest means fewer than ten employees and no previous Phase 2 awards. Maturest means more than five Phase 2 awards. The motivating idea was to speed those concepts whose prospects relied on catching an early market. One estimate by a Hewlett-Packard cognescenti says that six months sooner means 30 percent more total profit and six months later 30 percent less total profit and that one month sooner to market would cover the entire engineering and development costs. Third party cash provides a commercial validation that government cannot provide and separates story tellers from entrepreneurs. The DOD Program Manager, Jon Baron, says that Fast Track works to "smoke out the entrepreneurs".
Qualified Fast Track applicants get an immediate infusion of $40,000 to keep the project moving while DOD evaluates the proposal and they get "the Department's highest priority" for approval of Phase 2 (whatever that phrase means in practice). As long as Fast Track proposers are a small fraction of the total, the DOD agencies can afford to go along without sacrificing their short-term, predictable, and military R&D objectives. Few DOD SBIR deciders accept Congress's commercial premises of SBIR.
While few government projects have a time-to-market urgency, except when the bullets are flying, government can benefit from getting the products it likes into a competitive market. The alternative requires government to pay the full development cost. In that regard, government will have to decide which it wants - speed and cheap products or order and dear products. Fast Track is its first recognition that DOD has a choice. The established paradigm closely controls R&D details and pretends that the resulting products are an undefined "low-cost, high performance".
DOD's first data on Fast Track showed that it received 19 applications in late 1996 which is fewer than ten percent of Phase 2 applications. BMDO had the most projects of four participating DOD agencies, 45 percent of the proposals in a program with just ten percent of DOD's SBIR money. Most applications (13) were in the highest matching ratio (4:1) and none were in the 1:1 ratio. More money, more interest; that's just economics. But the 4:1 ratio were available only for companies who had never won any Phase 2 and had fewer than ten employees. So, the experienced players who could not or would not attract 1:1 matching would rather chance the mob at the trough.
Some venture investors will not invest in Fast Track because the company started its capital raising too late and because the combination of young technology and immature management present too much business risk. (Technologists have to understand that investors prefer first rate management with second rate technology to the reverse.) If photonics applicants want Fast Track help, they will have to line up the deals before Phase 1. If the shy investors are worried about risk, they could take heart from Garnsey's finding of a "15% failure rate per year in catering, compared with only 5% for the technology-based firms".5 They should also realize that consciously taking on smart risk remains the best way to succeed in investing. "We embrace risk because the opposite is timidity, which is what most of the world is good at. Always remember, dear reader, that it wasn't the Wise who discovered the New World".6
A full evaluation of Fast Track, of course, must wait the long years of a gestation period and the uncertainty of linking market results to any seed stage factor. Since Fast Track is founded on the premise of grabbing quick-moving markets, the idea will work only if government picks market-driven companies and speeds the investment..
"What we had briefly was a crazy situation in which someone would quit a job at Silicon Graphics, fill out a napkin, take it to a VC, get a million, and go public within a year - before they had a viable business - at a market cap of half a billion dollars"7.
The Fast Trackers and others who claim they cannot find money do not live in the 1996 world of information technology. Investment is booming in America, "a money bomb has hit our industry", and will continue to boom until the bubble bursts. With the information technology market in pell-mell advance, capital will eagerly seek opportunities to exploit the 1995 Telecommunications Act in a way that advances small businesses against the regulated monopolies and the exploding demand for both software and hardware with ample share for photonics. The links among video, photonics, software, and Internet will be giant and bound to expand beyond any estimates made today. The investment evidence is compelling:
1) Returns. Annual returns on private venture investments were 62% in 19958 and four of the top 21 venture firms with 1993 Information Technology IPOs had post-IPO gains over 100%9and IPOs in the Standard and Poor's New Issues Index nearly doubled in 199510. "$10,000 invested in [companies valued under $25M] on Dec 31, 1951 soared to over $29M [through 1994], achieving a compound growth rate of over 20% for the 43 years 11 whereas large companies had a 10.5% growth rate. In the last four years 2300 IPOs have come to market12. "The go-go nineties are puttng all previous venture cycles to shame."13 "You pick the right start-up, the return beats any other form of legal investment and probably most illegal investments"14. Table 1 shows the growth of, and returns to, venture money.15 1990 1991 1992 1993 1994 New Funds 15 10 18 31 27 New Money ($B) 1.8 1.2 2.5 2.5 3.8 Returns (%) 34.4 20.4 22.3 32.9 51.4 Table 1. Venture Trends 1990-1994
2) Volume. VC firms invested $5B in private companies in the first half of 1996 in the average five-year old venture-backed firm rose 65% to $11.6M in 1994 up from $7M in 198516. Many mutual funds investing in small companies are so oversubscribed that they closed their doors to new investors. Even Japan investors are coming, since they can only earn about 2% in the home bond market.17 Venture capitalists sank $366M into New England companies in the second quarter18. (Maybe sank is too negative a verb.) and a record $761MM into Silicon Valley19 in the second quarter of 1996.
3. Growth. Of the $7200M in private equity financing by venture backed companies 47% went to information technology;20 of Inc's fastest growing 100 firms (1991-1995) half are information technology; the most profitable IPO ever was a company with a powerful new switch for networks of PCs and workstations21. Some also, of course, cashed out their earlier investments in twenty IPOs raising $890M. Michael Murphy, editor of California Technology Stock Newsletter says, "A drop in technology stocks is a dumb reason to stop VC funding", since "technology represents 15% of the nation's economy, and notes that "half the world's population has never made a phone call"22. The high tech tariff treaty of December 1996 promises to reduce tariffs, and therefore expand trade volume, on about $40B per year in present US exports of semiconductors and telecommunications and $89B in computers software and electronics23.
Photonics companies, too, can get some of the $1B a year. They can compete well in those few SBIR programs seeking high-tech market-driven innovation. (Note: Market driven does not mean market-blathering which many SBIR applicants practice on market-indifferent government. Market-driven also means a healthy knowledge and respect for the economics of the photonics industry.)
Chances of winning vary a lot. The raw statistics, that only ten percent of proposals are funded for Phase I tell a misleading story. The National Science Foundation and the US Air Force, for example, see vastly different worlds yet claim the same criteria. BMDO funds about 30% of proposals. To keep plenty of wiggle-room, the government keeps its SBIR literature vague which forces it to face hordes of uncompetitive proposals which in turn encourages the small business lobby to seek more money by citing an "intense competition".
Proposers need a view of the agency and a post-SBIR market strategy since marketable photonics products need a lot more than $1M. The market-driven can avoid from those agencies who want incremental and predictable improvement on an inflexible and restricted budget. Since the feds don't see commercialization the way market-driven companies do, and since the feds don't know markets, and since the feds cannot judge marketability claims, many simply use the companies' commercial fantasies as a rationalization to fund government technology. But the agency's wants will rarely serve the company's market prospects. A credible fantasy may thus actually win an award and lose a market.
The small photonics companies can either fall into the SBIR trap or they can take the message from the companies helped by BMDO's SBIR that Yes, Small Companies Can Commercialize Photonics. And even for fun and profit.
Note: Author managed the BMDO SBIR program 1987-1996.
1 G Gilder, "The Gilder
Paradigm", Wired, December 1996, p225
2 The Economist, Sept 21, 1996
3 PR Newswire, Oct 8, 1996
4 E Matson, "He Turns Ideas into Companies - at Net Speed", Fast
Company, Dec-Jan97, p34-36
5 E Garnsey, "Pennies from Heaven", New Scientist, 24Aug96
6 The Motley Fool Investment Guide
7 J Burke, The Red Herring, Sept 1996
8 JF Geer, "The Venture Capital Boom", Financial World, Nov 18, 1996
9 Source: VentureOne Corp, quoted by ZA Herlick, "Venture Capital
Roundup", The Red Herring, Farch 94, Issue 9
10 D Lohse, "Small Isues' Big Gains May Hold for Long Haul", Wall
Street Journal, April 29, 1996, p C1
11 C Willis, Worth, Sept 96
12 "No tech, No takers", Inc, May 1996, p45
13 Financial World, Nov 18, 1996
14 M Moritz, quoted by S Kaufman, "Good Ideas get early help from
venture incubators", San Jose Mercury News, Feb 19, 1996
15 data from Venture Economics Investor Services, cited in Financial
World, Nov 18, 1996
16 Source: National Venture Capital Association, quoted by M Selz, Wall
Street Journal, May 2, 1996, pB2
17 Forbes, Oct 21, 1996
18 The Boston Globe, Aug 21, 1996
19 The San Jose Mercury News
20 DT Gleba, "No End in Sght for the Info Technology Boom", Upside, May
1996, p82
21 S Tully, "How to make $400,000,000 in just one minute ...", Fortune,
May 27, 1996, p85-92
22 Wired, Jan 97
23 S Schiesel, "Long Road for Global Technology Trade", The New York
Times, Dec 21, 1996