接上篇，计划认真地把这本书做些摘录，供以后常回来翻翻。
3 Von Sequoia & Tom KPCB
Rather than merely identifying entrepreneurs and monitoring them, as Rock had done, the new venture capitalists actively shaped them: they told company founders whom to hire, how to sell, and how to structure their research. And to ensure that their instructions were implemented, the new venture capitalists came up with a second innovation: rather than organizing one large fundraising, they doled out capital in tranches, with each cautious infusion calibrated to support the company until it reached an agreed milestone. If the 1950s had revealed the power of liberation capital, and if the 1960s had brought the equity-only, time-limited venture fund, the advances of the 1970s were twofold: hands-on activism and stage-by-stage finance.
50年代-70年代的VC四大发明，50年代ARD和仙童的故事揭示了自由资本的神奇，60年代Davis & Rock Fund演化出了设定期限的纯股权VC基金，区别于ARD最初的开放式基金设置以及SBIC尝试，模式上只是重点关注创始人；70年代则是VC基金的运作方式进化，进化出了更加细节的辅导公司和分批注入资金的尝试。这些做法虽然一直延续到了今天，但不少人用起来还非常顾虑。
Valentine is said to have remarked that underperforming company founders should be put into a cell with Charlie Manson. He rose through the ranks at Fairchild Semiconductor and later at its rival National Semiconductor, developing a sideline investing his own money, including in the Rock-Palevsky bonanza, SDS.
Don 的故事和很多VC大佬一样，都是自己有了成功经历，痛恨失败才开始将VC作为事业。
It took Valentine a year and a half to raise $5 million for his first fund. But in the end he succeeded by tapping pools of capital that enjoyed charitable status: the universities and endowments that escaped not only regulation but also capital-gains tax. The Ford Foundation came in first, later to be joined by Yale, Vanderbilt, and eventually Harvard; ironically, the Ivy League investment bosses showed a greater open-mindedness about a gruff Fordham graduate than many alumni could muster. In so doing, the endowments set in motion one of the great virtuous cycles of the American system. Venture capitalists backed knowledge-intensive startups, and some of the profits flowed to research institutions that generated more knowledge. To this day, the conference rooms at Valentine’s old firm are named after their main limited partners: Harvard, MIT, Stanford, and so on.
Don的募资也并不顺利，18个月才募到了500w刀，这就是红杉基金的第一期，1974年。幸运的是受益于捐赠和慈善基金的政策，进来的LP都是这个类型，特别是大学的捐赠基金其实无意启动了这个伟大的循环：捐赠基金投资于VC——> VC投资到知识密集的创业公司——> 利润再反哺学校等科研部门，以生产更好的知识。Don还把这些LP的名字挂在了会议室的门上，非常好的做法。
Unlike Arthur Rock, he was a hands-on business operator by background, and in his years as a semiconductor salesman he had learned how to translate products into profits. You had to go with the version of your invention that would earn you the fattest margin, and you had to open sales channels to as many customers as possible. In the case of Atari, this meant capitalizing on one of Bushnell’s many semi-formed epiphanies: if Pong could be sold to families rather than bars, the market could be expanded enormously.
Von 比较Hands-on的方法论放今天依然很有用，追求高毛利、打开销售渠道，尽量To C！多么伟大的洞见，基本是一个企业家非常朴素的思路。
In late 1974, a few weeks after his visit to Atari’s factory, Valentine made a decision. He was not going to invest yet: Atari was too chaotic. But he was not going to walk away: the potential was too terrific. Instead, he would get involved cautiously, in stages, and he would start by rolling up his sleeves and writing an Atari business plan. If all went well—if Bushnell embraced his strategy, and if the plan attracted interest from other venture capitalists—then Valentine would invest. He would risk his money, in other words, only when Atari had been at least partially de-risked. Activism and gradualism would thus combine to make a hot tub culture backable.
Atari项目能看出来Von的谨慎和远见，太乱了暂时不投，又能看到未来的巨大前景，并不走开，相反Von采取逐步投资、撸起袖子亲自下场帮忙写BP、找人一起跟投的策略。这样的做法不成功都没有道理。
The tally of initial public offerings plummeted from more than one thousand in 1969 to just fifteen in 1974, and the S&P 500 returned more or less nothing over this period.The collapse all but wiped out the nascent hedge-fund business, and likewise a headline in Forbes wondered, “Has the bear market killed venture capital?” After attracting $171 million in new funds in 1969, venture capitalists raised only $57 million in 1974 and a mere $10 million the year after.
从石油危机开始的这五年的动荡堪比当下了。没想到美国当年也是69年前后每年能有1000个IPO，相比之下中国几百个并不算多；但到1974年就跌到15个了，跌掉了98.5%… VC的募资金额也从1.7亿美元跌到了1000w，跌掉了94%。这样巨大的收缩比国内的近几年严重多了。
By the middle of March, Sears had placed an order for seventy-five thousand Home Pong machines. Atari now had what Valentine had been waiting for: a promising new product and a powerful distributor.At the beginning of June 1975, Valentine duly invested. He bought 62,500 shares for $62,500, making what would now be termed a “seed investment” in Atari. But it was only the start. Twelve months later, in the summer of 1976, Valentine confronted his next challenge. Atari’s bell-bottomed engineers had come up with a fresh idea: a console that could play not only Pong but multiple games of the owner’s choosing. To capitalize on this breakthrough, Atari was going to need a far larger capital infusion—perhaps as much as $50 million. There was no way the venture capitalists of the era could mobilize that kind of cash, and the stock market was all but closed; in 1976 only thirty-four companies managed to go public.For Atari to develop its multi-game console, Valentine would have to come up with another way of raising capital.By the end of the day, a starstruck Bushnell had agreed to sell Atari for $28 million. For Valentine and his fledgling fund, it was a satisfying exit. Sequoia notched up a useful 3x return, demonstrating the value of the new investment methods.
1974年设立的红杉基金一期，Atari几乎是第一个项目了，Von跟着等了半年多到75年6月才投了个6w的种子，仅是基金规模的1%，一年后又放了近100w进去，基金的20%规模了。再过一年面临公司需要的巨大投入投不起，只好卖掉了，2800w的价格创造了3倍的回报。考虑当时的极端市场环境，已经相当不错了。75年全市场才募到了1000w，放100w进去必然是当时最大的VC项目了。
In 1973, Bill Draper’s Sutter Hill Ventures struck a landmark deal with Qume. What made the deal special was that Sutter Hill imposed a condition: Qume’s founding engineer had to dump his underpowered chief executive and allow the venture guys to bring in a star graduate of Harvard Business School. When the company took off, the CEO’s stock options generated a huge payout. Sutter Hill went on to repeat the Qume formula over and over, liberating up-and-coming big-company executives and distinguishing West Coast venture capital from the tamer version in the East.
Sutter Hill创造的Qume公式后来非常常用：换个更有能力的CEO，给一大笔股票期权，更快成功。
Hillman committed up to $5 million, provided that the partners could raise a matching sum from others. Kleiner and Perkins proceeded to raise $1 million from Rockefeller University, nearly that much from two insurance companies, and a bit more from wealthy individuals and trusts. By the first week of December 1972, they had put together a fund of $8.4 million.
KPCB的起步比红杉要早两年，1972年的募资环境还没完全恶化，KP一期募到了840w。
Kleiner and Perkins set up shop in a new low-slung office park at 3000 Sand Hill Road, becoming the first partnership to occupy what was to be the epicenter of the venture industry. Their timing was poor: they were launching their fund on the eve of the first oil shock, and their first few investments performed as poorly as the economy. They backed a plausible semiconductor startup, but it was run into the ground by inexperienced managers. By the end of 1974, Kleiner Perkins had shelled out $2.5 million for nine investments. Kleiner and Perkins felt sufficiently glum to rethink their strategy. The new Kleiner Perkins formula doubled down on activism. Rather than funding outside entrepreneurs, the partners would incubate startups in-house, kicking ideas around with junior associates.
KP虽募资不错，也选了个好办公地，无意间成为了沙丘路上的第一家VC。但赶上市场环境恶化，还是出师不利，72年底设立的基金到74年底的2年投资期，仅投出去250w，不到基金的1/3，9个项目几乎死了5个。这让他们重新思考自己的投资策略，形成了新模式：从投资外部企业转向更多的内部孵化项目，更加积极参与企业运作，才有后来的巨大成功。
In an inversion of Atari, the technical risks were daunting, but the market risks were negligible. Then, having cleared that hurdle, Perkins invested $50,000 in the project. It was a token amount, roughly equivalent to Valentine’s initial seed investment in Atari. If the project hit a wall, the Kleiner Perkins fund would lose less than 1 percent of its capital. Brook Byers, who joined Kleiner Perkins a few years later as a young partner, reflected on the lessons that KP drew from this experience. By focusing exclusively on the “white-hot” risks in a project, you could find out whether the venture was likely to work while risking as little capital as possible. With that, Perkins resolved to finance Tandem’s Series A round without sharing the risk with other partnerships. Shoveling his chips onto the table, he invested $1 million in Tandem in early 1975, receiving 40 percent of the equity. If Tandem had not worked out, there might never have been a second Kleiner Perkins fund.
Tom对风险的看法最值得学习也极其经典，仅仅关注其最紧要的风险“White-hot ”risk就可做投资判断。Atari主要是技术风险，市场风险可忽略不计，5w投资虽然不多，但经验很关键。74年底的反思后，75年初对Tandem A轮这笔100w/40%股份就是用新方法孤注一掷了，如果失败就Game Over了。
Perkins’s law: “market risk is inversely proportional to technical risk,” because if you solve a truly difficult technical problem, you will face minimal competition. Thanks to the high barrier to entry, Tandem’s profit margin remained juicy even as its sales soared. By 1984, Tandem had generated a bit over 100x on KP’s $1.45 million investment. The $150 million profit dwarfed the combined $10 million return that KP made on all of its first nine investments. But precisely because the technical challenges were so formidable, the barriers to entry in this business would be high, and Genentech would be able to extract fat margins if it succeeded. It was another illustration of Perkins’s law.
市场风险和技术风险成反比，技术难度越高/技术风险越大的，就没有竞争，市场风险反而小，还能收获高毛利，实现大成。反之技术没门槛的，竞争激烈市场风险就很大，这个观察和总结非常到位。KP把重点投入到高技术风险的项目的孵化上，Tendem和Grentech都是如此，类似于今天的硬科技。Tendem这个项目在十年后退出的时候实现了100x回报，收益达1.5亿，而此前的9个项目才实现了1000w的利润。所以Perkins Law的要义其实是：重点关注高技术风险依然值得投入的项目，搞定了就改变世界那种！
The next day Perkins met Swanson again and laid out a suggestion. The science was captivating, but the $500,000 cost of proving it was prohibitively high, given the uncertainty. So Perkins proposed to repeat the formula he had developed for Tandem: identify the white-hot risks, then find the cheapest way of going after them. Swanson should cut the cost of his experiments by not hiring scientists or setting up a lab. Instead, he should contract the early work to existing laboratories. Perkins agreed to invest the new sum that Swanson needed: a mere $100,000. In exchange for this modest commitment, representing just over 1 percent of the Kleiner Perkins fund, he acquired fully one-quarter of Genentech’s stock. There was nothing unfair about this: Swanson had tried to shop the deal elsewhere and had found no takers. Having parted with a quarter of their company for a mere $100,000 in 1976, Boyer and Swanson sold a 26 percent stake the next year for $850,000, and in 1978 they sold just 8.9 percent for $950,000. As of 1984, the fourteen investments in the first fund showed a combined profit of $208 million; of that, fully 95 percent came from Tandem and Genentech.
基因泰克的案例也是KP关注最紧要风险的成功实践，不只是找到核心的技术风险，还要找到最便宜的解决方案，然后投一小笔钱进去试水，1-2%的基金规模试水，成功了就继续下注，失败了损失也非常可控。这其实是极好的策略。最初在1976年这笔10w/25%股份的交易估值约40w，相当于今天的人民币2000w估值投了500w进去，2年之后公司估值就翻了25倍，达到了1000w，相当于5亿RMB。这个策略其实是非常值得我们借鉴的策略。KP第一期基金到1984年结束，大约12年之后，已经实现了整体超过20x的基金回报，2个项目的百倍回报贡献了基金整体回报的95%。
4 Apple, Angels Markkula & Montagu
Even though Valentine was perfectly suited to be Apple’s first investor, his initial reaction to Jobs and Wozniak was skeptical. Jobs “was trying to be the embodiment of the counterculture,” Still, Bushnell and McKenna had told him that these guys were worth an audience. Because he valued his network, Valentine went through the motions of asking what Apple was up to. Markkula decided to put his energy behind Apple. He became an adviser to the Steves, writing their business plan, serving as marketing chief and company chairman, arranging a bank credit line, and ultimately investing $91,000 of his own capital in exchange for 26 percent of the company.And so, almost on a whim, Venrock committed $300,000 for 10 percent of Apple. By valuing the company at $3 million.
Apple的故事还是蛮令人深思的，今天的3w亿美元巨兽的种子轮融资估值仅30w，增值了1000w倍。Von虽然经验丰富，特别是阅人上，还是对Jobs的反文化个性不够认同，决定等等看；这也成就了Markula，他其实是当时同样的撸起袖子下场干的套路，亲自写BP、帮忙融资，一年后公司A轮估值就翻了10倍，融到了洛克菲勒家族Venrock的30w。
The image of the moving train represented a new gloss on stage-by-stage investing. In the case of Atari or Genentech, follow-on venture capitalists wrote checks once the white-hot risks had been neutralized. In the case of Apple, VCs were being told that they should invest simply because others were investing. However circular this logic, it was by no means crazy. The whispering grapevine was sending a message: Apple would be a winner. In the face of that social proof, the objective truth about the skills of Apple’s managers or the quality of its products might be secondary. If Apple was attracting funding, and if its reputation was soaring thanks to well-connected backers, its chances of hiring the best people and securing the best distribution channels were improving, too. Circular logic could be sound logic.
At a quarter to seven that evening, Mike Scott appeared again. “Mr. Montagu, you are really a fortunate guy,” he said. Steve Wozniak had decided to buy a house. To raise the cash, he wanted to sell some of his own equity. Montagu asked how much stock Wozniak was selling.“Four hundred and fifty thousand dollars,” came the answer. It was more stock by far than Venrock or Valentine had laid hands on. A giddy Montagu called Kramlich again. “Dick, I wouldn’t be here without you!” he said, offering to split the allocation. Kramlich never told Rock that he had acquired a large slice of Apple through this roundabout route, and for years he kept quiet about it. Two of the biggest winners in this saga were the improbable duo of Anthony Montagu and Dick Kramlich, proving that sheer luck can sometimes matter more than anything.
Apple的故事中跟风投资的现象已经出现了，并且这种循环逻辑实际十分有道理，把市场上的钱拿完，买来各种资源，必然能成功。这已经是当下独家兽的竞争策略之一了——扎堆儿投。Montagu的坚持和执着也十分难得，功夫不负有心人，最终还是投了进去，也还没忘把朋友带上。沃兹尼亚当年45w的房子放今天是2000w人民币，也不便宜，但15%的股份是3万亿人民币了。
The low capital-gains tax and the change to the prudent-man rule rounded out a policy mix that was extraordinarily favorable to venture investors. Venture-backed firms could go public without showing a history of profits. Employee stock options were taxed only when they were finally exercised, not when they were initially granted. Limited partnerships were exempt from tax, and they protected investing partners from lawsuits. No other country was so friendly to the venture industry.
In the five years from 1973 to 1977, the venture industry had raised an average of $42 million annually. In the next five years, it averaged more than twenty times that—fully $940 million annually. With the return of the hot IPO market following Apple’s debut, established VC operators began to generate extraordinary profits: annual returns of between 30 and 50 percent became commonplace. Having drummed up $5 million for his first fund, Don Valentine raised a second fund of $21 million in 1979, followed by a $44 million fund in 1981.
美国的VC政策很早就理顺了，资本利得税、LP结构、期权在执行时才纳税、没利润也能IPO这几点结合一起让行业发扬光大。国内其他都差不多了，唯一缺乏的是上市制度这条，依然过于严格和非市场化。石油危机之后的五年，恢复还是非常快的，VC募资翻了近20倍达到快10亿美元，也是之前69年高点的5倍了。同时的红杉74年的一期基金500w，5年后的二期基金2100w，7年后的三期基金达到了4400w。
5 3Com & Cisco
But Silicon Valley’s venture-capital machine, now equipped with a full tool kit as well as a dense network of players, was about to deliver two simultaneous lessons. First, that it could fight off the challenge of Japan, whose formidable semiconductor manufacturers threatened the Valley’s core industry. Second, that it could at last eclipse its long-standing American rival, the technology hub centered on Boston.
80年代的VC发展壮大，不仅受益于前辈们在70年代发展出来的技术工具：阅人、下场干活、分阶段投资等，也开始受益于投资人和企业家达成的网络效应，一方面共同对抗当时来自日本半导体产业的威胁，另一方面也对美国的大企业构成威胁。
Hierarchical organizations can be good at coordinating people when the objectives are clear: think of an army. But when it comes to commercializing applied science, the Valley’s culture of “coopetition” has proved more creative than the self-contained, vertically integrated corporations of Boston or Japan. Large companies bottle up ideas and often waste them.
合作是这些网络效应产生的源泉和基础。公司间的合作比企业内部的垂直网络要更灵活，更有优势。
By emphasizing the porous boundaries between Valley startups, she was probing the quality of the relationships within a cluster and suggesting why some clusters pull ahead of others. A cluster that is dominated by large, self-contained, secretive companies will be characterized by tight relationships among insiders at each firm, but few links between professionals at one firm and similar professionals at another one. In contrast, a cluster that consists of transient startups will feature fewer deep bonds among colleagues, but it will be enriched by myriad, looser external connections.
What created these conditions in the Valley? There are two familiar answers. First, California law prevents employers from tying up employees in non-compete agreements; talent is free to go where it pleases, unlike in most states, including Massachusetts. Second, Stanford has been generous in allowing professors to take sabbaticals to work on startups, and this permissiveness has fostered ties between academia and business; in contrast, MIT professors have risked losing tenure if they spent too much time on side projects.
产业集群是松关系还是紧关系还非常关键。紧关系是围绕大企业的，松关系则是小企业之间的，硅谷的成功主要是依赖松关系类型的。
Without Tom Perkins’s loose ties to his old friends at Hewlett-Packard, Tandem Computer would never have been conjured into being. Without Nolan Bushnell’s loose ties to Don Valentine, and without Valentine’s loose ties to Mike Markkula, Apple might never have become a real business.
For many of these venture novices, building their networks was not just one thing that they did. Rather, it was the thing—the key to getting established in business. Bill Younger, who joined Sutter Hill in 1981, set himself the task of taking the smartest people in his Rolodex to lunch; at the end of every meal he’d ask, “Who is the absolutely best guy you’ve worked with?” Younger then made it his mission to meet that best guy—it was almost never a woman—and toward the end of the meeting he would repeat the question: “Who is the absolutely best guy out there?”After a year of moving from one best guy to the next one, Younger had a list of about eighty superstars, and he cultivated each of them methodically. He would send one luminary a technical article that might be relevant to his research; he would call another to mention that an old colleague was asking after him. In this way, Younger spun a web of loose connections that would form the basis for productive startups when the right opportunities beckoned.
松关系带来了Tendem、Atrai、Apple的成功，共同成就了创始人和其背后的投资人。此后，建立和融入这种网络其实成了VC最重要的工作了，Bill的尝试今天都依然值得学习。
One Friday in 1981, Fairchild’s former chief executive Wilfred Corrigan circulated a business plan for a new semiconductor company, LSI Logic. By the following Tuesday, Kleiner Perkins and two co-investors had stumped up $2.3 million; the only reason it took so long was because Monday was a holiday.
LSI的投资看起来太神奇了，周五发出BP，周二就到位了230w。要不是因为周一是假期，就周一到了。
Seeking to reduce risk, Boston VCs often provided “developmental capital” to businesses that already had a proven product and some early sales; this was a whole lot safer than betting on fledgling startups. Howard Cox, who embarked on a career at Greylock in 1971, boasted that he lost money on only two of his forty investments. “I did not back companies where the product might fail.
不同于硅谷的套路，波士顿的VC更保守，仅支持已经有产品和初步销售的发展阶段的公司，而不是初创阶段的公司，这其实也是一个不错的策略，只是没有那么大的收益。1971年成立的Greylock就是这个风格，40个项目只失败了2个，胜率不错。
Metcalfe set out to raise money for 3Com in September 1980. He attracted offers in no time. The Mayfield Fund, set up by Rock’s former partner Tommy Davis, proposed to value 3Com at $2 million, or $7 per share. Dick Kramlich of New Enterprise Associates, another former Rock partner, put together a syndicate willing to pay a valuation of $3.7 million, or $13 per share; this was before 3Com had done anything. But Metcalfe was determined to get more. Declaring that his company was worth $6 million and that his shares should fetch $20 a piece, he set out to beat the venture guys at their own game.
3Com的融资看起来相当顺利，80年代已经很多投资人了，踊跃报价，公司估值从200w居然一路提到了600w，按今天的人民币看差不多3亿估值的天使轮了。
At some point in each conversation, the venture guy would launch into a lecture on the three reasons startups failed: the excessive ego of the founder, too little focus on the most promising products, and too little capital. Having recognized this mantra, Metcalfe started to preempt it.
Metcalfe的理由其实挺充足，创始人、产品和资金是公司挂掉的三大原因，他现在只缺资金了。公司成功似乎不在话下，于是估值就不断涨。
The good news for Metcalfe was that Boston venture capitalists were properly impressed by the hiring of Krause. They liked to invest in ready-made teams. 3Com now had a top-flight inventor and a top-flight manager. Pretty soon, Fidelity Ventures, the VC branch of the storied Boston money house, announced that it would finance 3Com at $21 per share. At last, Metcalfe had a valuation that exceeded his $20 target. Eventually, after a month of frustration, Metcalfe concluded that the promised $21 investment was a mirage: it vanished every time he neared it. A startup’s scarcest asset is time. Venture capitalists in Boston turned out to be champions at wasting it.
With that, Silicon Valley’s network delivered the deal in a matter of minutes. Melchor picked up the phone and spoke with Mayfield and Kramlich, and soon it was agreed that Melchor’s fund would commit $450,000, Mayfield and Dick Kramlich would be in for $300,000 each, and another $50,000 would come from small investors with connections to 3Com. There were no futile protection clauses in the fine print, no eleventh-hour conditions, and no need for Metcalfe to hustle for dollars at the expense of time with his company. On Friday, February 27, 1981, 3Com received a check for $1.1 million in exchange for a third of its equity. If the cash had not arrived that day, 3Com would have missed payroll.
Metcalfe had failed to get the $20 per share that he had wanted. But he did have the pleasure of calling his Boston tormentors one last time. They would not be investing in 3Com, he informed them.“Why?” came the aggrieved response. “We supported you when no one else would.” “No,” Metcalfe shot back. “You lied to me when no one else would.” 3Com went public in 1984, generating a 15x return for its early investors. But this success was a small part of a larger phenomenon
3Com融资的故事其实非常经典，波士顿VC给出了非常慷慨的价格，却也要求非常细节的尽调，条件不断加码，交易迟迟不能完成。创始人及时止损，没有被拖下去，转而回到硅谷VC，接受了略低的价格，迅速完成交易。资金如果晚到一天，公司就发不出来工资了。这次融资从80年9月一直到81年2月才做完，耗时近半年。融资完成时创始人还不忘diss下波士顿的VC们。3年后公司就上市了，早期投资者们回报15x。
In 1986, Lerner and Bosack quit Stanford to work full time on Cisco. They were joined by three other former Stanford employees, and they began to sell homemade versions of the multi-protocol router. Money was tight, and the founders sought out venture capital, showing up at networking events and pitching dozens of investors. But their efforts came to naught. For one thing, the venture boom had cooled. A surfeit of capital had depressed returns, and the previous year private venture partnerships had raised $2.4 billion, down from a bit over $3 billion in each of the two previous years. For another thing, the multi-protocol router could not be protected by patent: Stanford claimed ownership of Cisco’s inter-networking breakthrough. Then there was the question of the founders themselves. Bosack was by turns silent and given to relentless logical soliloquies; Lerner developed a habit of saying “Control-D” when she tired of his algorithmic monologues.
By the start of 1987, Cisco had made enough progress to recruit a couple of extra people. But without venture backing, the company could not attain escape velocity. In the absence of experienced guidance, Lerner and Bosack hired cheaply and eccentrically. A former naval officer with no startup experience arrived to serve as vice president for finance. A new chief executive took it upon himself to veto a sale of routers to a lab with military ties, explaining that if Cisco’s gear malfunctioned, it might trigger World War III, which was more responsibility than he could handle.
Spurned by investors, the Cisco team soldiered on tenaciously. They kept the lights on by maxing out credit cards and deferring salaries; Lerner took a side job to help pay the bills, and one co-founder made a personal loan to the company.
思科的故事也很精彩。86年创业赶上资本有点降温了，虽然公司已经开始卖路由器了，融资还是很难，也和创始人的古怪性格有关系。两位只能胡乱请人进来，用人不慎真是最大的教训了，几乎一片混乱。只能不断的刷信用卡、迟付工资甚至打零工来支持公司。
But then Cisco’s fortunes turned—in classic Silicon Valley fashion. The antimilitary chief executive knew a lawyer, and the lawyer had a partner named Ed Leonard, and Leonard happened to work with people in the venture industry. In a different corner of the world economy, this might have been irrelevant: a lawyer like Leonard would not have bothered senior venture guys on behalf of some random half acquaintance. But Silicon Valley rainmakers were unlike rainmakers elsewhere: they positively wanted to be bothered. Making and taking introductions was their stock-in-trade. If Leonard introduced them to a long-shot entrepreneur, it could only boost his standing. However, the way Valentine saw things, none of this mattered. Hewlett-Packard had testified that engineers were “tearing the hinges off the doors to get the products,”
思科的融资故事是最好的“松散链接”的典型，不靠谱的CEO碰巧认识个律师，律师正好有个同事，同事正好和VC行业的人共过事。VC这边Don还挺重视，通过HP的关系发现客户真是非常认可。
At the end of 1987, Sequoia duly invested $2.5 million for a third of Cisco. On the face of it, these terms were fairly generous: six years earlier, 3Com had sold a third of itself for just $1.1 million, although it had been at an earlier stage in its development. But Valentine had understood Cisco’s weaknesses and structured the deal accordingly. Fully one-third of Cisco’s stock was set aside for existing managers and future employees, enabling Valentine to recruit a clean slate of executives who would take over the leadership from the founders.Lerner and Bosack retained the remaining third, but the majority of their stake was converted into nonvoting stock options, giving Valentine control of board decisions.To make a success out of Morgridge, Valentine gave him founder-type incentives. He loaded him up with stock options so that he would pocket around 6 percent of Cisco’s success; this gave him more skin in the game than some CEO-founders.
87年末的思科终于成功融资250w，给了红杉33%的股份，公司还预留了33%股份给未来团队，创始人只拿走了33%还不带投票权。找了个CEO的期权代价大约是6%。
The most thoughtful new entrant was Accel Capital, the first venture partnership to position itself as a specialist in particular technologies. By accumulating deep expertise in software and telecoms, Accel aimed to have the inside track on which entrepreneurs to back and how to guide them to a healthy exit. At the same time, Accel embraced an approach that it came to call “the prepared mind.” Rather than looking anywhere and everywhere for the next big thing, the partnership carried out management-consultant-style studies on the technologies and business models that seemed to hold promise.
6 Accel & GO, UUNET
Accel是首个专注行业的VC基金，主要参与软件和电信项目。Prepared Mind 也是其核心的逻辑方法，就是要有研究驱动的深刻洞见，基金雇了一批管理顾问来支持研究。
This tension between the planners and the improvisers tested the industry’s identity, as we shall see presently. To raise money for their pen-computer plan, Kapor took Kaplan to meet John Doerr, the high priest of the Valley’s improvisers. Doerr and his friend Vinod Khosla set the pace at KP, and they were out to back truly revolutionary startups that could spawn entirely new industries. Magnetic and messianic, Doerr in particular became the go-to investor for fearless founders, who loved him for championing their visions even more passionately than they did. He had “the emotional commitment of a priest and the energy of a racehorse,” one entrepreneur marveled. “The number of different things about which John Doerr has said, this is the greatest thing ever, is a big number,” a rival investor noted, with a mixture of respect and cynicism. Both men were traveling over the next days, but they juggled their schedules to arrange a layover at St. Louis airport. Doerr met Kaplan at the gate, and they hammered out a deal: together, Kleiner Perkins, Mitch Kapor, and Vinod Khosla would purchase a third of Kaplan’s project for $1.5 million. Doerr would become chairman. Kapor and Khosla would serve as board directors. What would he call his company? Doerr asked him.“GO, all caps. As in GO forth, GO for it, GO for the gold.”“After Doerr’s intervention, Kaplan was able to raise $6 million in a matter of days; it was more than the $5 million he had targeted. He soldiered on until 1993, raising money periodically with Doerr’s help but failing to realize his vision of building a pen-operated computer. In the end he sold GO at a fire-sale price to a division of AT&T. His backers were left with almost nothing. GO showed me the downside for entrepreneurs of Kleiner’s approach. If it couldn’t be a home run, Kleiner did not care if the company struck out. It was like, go big or go home. . . . There is an arrogance to the KP approach. All that ego about changing the world. Kapor was correct that Doerr’s style invited trouble. Around the same time as the GO fiasco, Doerr and Khosla launched a next-generation laptop company called Dynabook Technologies that burned through $37 million of investors’ capital before closing. Doerr also trumpeted a string of technological prospects that turned out to be busts: human gene screening, anti-aging drugs, designer chemicals.He seemed to have forgotten the old Tom Perkins dictum: when you invest in a company facing a technical challenge, the first thing you do is take the white-hot risks off the table.
KPCB和GO的故事很有意义，明星创始人和明星投资人也无法避免的失败。从构思出GO的思路，到融资，都没那么难，KPCB甚至是在机场和创始人见面，达成了150w买1/3股份的安排，起步就融了500w，这已经是当年KP第一期基金的规模了。后续融资的故事也颇为传奇，几次不行，最终还是大佬Doerr出面轻松解决问题，但最终还是卖掉了，几乎所有人都颗粒无收。这也是KP的策略，要么做大，要么作死。类似的失败经历还有几次，不只是在电脑领域，还有基因检查、抗衰老药和化合物设计等领域。这些失败，回顾起来主要是背离了Tom在创办KP时坚持的原则：先把最紧要的风险搞定！
To underscore its strategy of specialization, Accel’s second fund, raised in 1985, was aimed exclusively at telecoms. The offering document proclaimed that “in an information-based economy virtually every electronic system will communicate with other systems”; the market for modems, networking, video sharing, and other telecom applications would be enormous. Incumbent VC rivals were impressed. Kleiner Perkins invested $2 million in Accel Telecom. Around the partnership’s tenth birthday, a tally showed that of the forty-five Accel investments that had experienced an exit, only seven had lost money.
Accel’s specialist approach made it particularly adept at identifying what venture capitalists call “adjacent possibilities.” By embedding themselves in their respective sectors, sitting on boards of portfolio companies, and blending their direct observations with management-consultant-style analyses, Accel partners could anticipate the next logical advance in a technology. “Every deal should lead to the next deal,” was another Accel saying. Swartz in particular liked to invest in successive iterations in a single product.
Accel’s performance in its first few funds left no doubt that it was onto something.The specialist telecom fund multiplied its capital 3.7 times, generating an annualized return that was more than twice as high as the median venture fund of its vintage. Taking the first five funds together, Accel generated an even better performance: the average multiple was 8 times capital. And yet the striking thing about Accel was that despite the partners’ firm intention not to chase hubristic grand slams, it was grand slams that dominated performance. Accel Telecom more than conformed to the so-called 80/20 rule: a whopping 95 percent of its profits came from the top 20 percent of its investments. Other early Accel funds exhibited similar power-law effects. In the firm’s first five funds, the top 20 percent of the investments accounted for never less than 85 percent of the profits, and the average was 92 percent. In short, the power law was inexorable. Even a methodical, anti-Kleiner, prepared-mind partnership could not escape it.
Accel是专业基金的开拓者，第二期基金专门投电信，连当时很牛的KP也做了LP参与进来。这期基金的表现也相当不错，10周年的时候52个项目仅有7个亏钱，45个赚钱，真是非常稳的表现。Accel也证明了“产业连接型投资”的可能性，一个项目自然地延伸到下个项目，一投一串，非常好。Accel也再次实践了幂律定律，20%最好的项目产生了整个基金90%以上的回报。
VCs as individuals can stumble sideways into lucky fortunes: chance and serendipity and the mere fact of being in the venture game can matter more than diligence or foresight. At the same time, venture capital as a system is a formidable engine of progress—more so than is frequently acknowledged.
In July 1990 a young Tennessee senator named Al Gore laid out a public-sector vision for an “information superhighway.” Rather than operating on existing telephone lines, as the internet did, Gore’s superhighway envisioned brand-new fiber-optic pipes that would turn household TVs into interactive terminals. Kapor experienced another one of his epiphanies. Gore’s government-led fiber-optic superhighway was still dominating the headlines. But the way Kapor saw things, it would be prohibitively disruptive and expensive. Rather than ripping up the ground to lay fiber-optic cable, it would be cheaper by far to build out the copper-wire-based internet.”
Adams wrestled with his dual instincts. If Kapor had been a standard investor, he would have turned him away. But, in his idealism and political outlook, Kapor felt like a kindred spirit. After some further thought, Adams accepted Kapor’s offer. If Kapor’s call was a stroke of luck, Accel’s deliberative, prepared-mind process was grinding along in the background. By now Accel had gotten wind of the internet’s potential through three different channels. Kapor had called; Gooding was following the scent; McLean had caught a glimpse of the voracious demand for online connections. The question was whether Accel would convert these hints into an investment. Luck and a good telecom team had put the opportunity of UUNET clearly in view. But Accel was still not seeing it. As often happens in the venture world, it took a nudge from a competitor to change Accel’s attitude. In February 1993, a telecom company called Metropolitan Fiber Systems made a play for UUNET.
Barris realized he was onto something. Because of his GE background, he knew what sorts of online services big customers would pay for. Because of his internet background, Adams knew how to provide them efficiently. By combining their knowledge, the pair of them could mint money
At NEA, Barris was by now keen to work with Adams, so he quickly agreed to match Jarve’s higher valuation. Delighted, Adams told Barris that he would do the deal exclusively with NEA; after all, Barris was the investor best positioned to help UUNET. But, in an example of VCs acting to protect their reputations and networks, Barris declined the opportunity to squeeze his rivals out. He had been brought to the dance by Arthur Patterson, and he refused to double-cross him. Eventually Accel agreed to the new price. In October 1993, the three partnerships stumped up a total of $1.5 million.
In the course of straightening out UUNET’s accounting, a bookkeeper came across a box of unpaid invoices. UUNET had lost track of debts for routers and other gear that amounted to a horrifying $750,000. The sum canceled out half of the capital that UUNET had just raised. Weeks after its $1.5 million Series A, the company was almost out of money.
UUNET的故事也很有意思。Kapor并没有因为GO的失败受太大影响，依然在寻找机会。90年戈尔副总统提出了信息高速公路的计划，要用光纤，但Kapor的直觉是光纤太贵了，铜缆才OK。这个判断让他说服了UUNET的创始人得以投资和加入进去。之后A轮融资的过程也是一系列的巧合：巧合的电话、负责人正好看这个领域，即便如此专业基金Accel都没看明白，直到竞争对手出现。NEA本来可以独吞的项目，也保持了其风度，采取了合投。即便如此，融资一搞完就发现欠了别人75w，150w的融资瞬间只剩一半了…
Back in 1993, a coin toss might have decided who backed UUNET: the corporate capitalists at Metropolitan Fiber or Accel’s venture capitalists. Now Accel stood to be fabulously rewarded. UUNET’s flotation gave the company a valuation of $900 million, and then, in a wonderful full-circle ending, Metropolitan Fiber made its second appearance, buying control of UUNET at a valuation of $2 billion. Through luck more than brilliance, Accel found itself pocketing 54 times its original stake, a profit of $188 million. Menlo received a similar return. NEA made even more because it held on to its position longer. Venture capitalists as individuals had made plenty of errors. But venture capital as a system helped UUNET to spread the internet to millions.
3年后UUNET和投资人就迎来了收获。20亿美元的收购对价让Accel赚了54倍，Kapor更是巨额回报。
When Mosaic Communications, by then renamed Netscape, went public the following year, the young Marc Andreessen owned just 3 percent—the same share that Clark had owned in Silicon Graphics.
那个时代的创始人真不容易，Marc居然只有3%的股份，估计还不如CEO的6%把。
Mosaic also marked a new stage in the evolution of the power law. Venture-capital returns are dominated by grand slams partly because of the dynamics of startups: most young businesses fail, but the ones that gain traction can grow exponentially. This is true of fashion brands or hotel chains as well as technology companies. But tech-focused venture portfolios are dominated by the power law for an additional reason: tech startups are founded upon technologies that may themselves progress exponentially. With that sort of wind at a tech startup’s back, no wonder profits could grow exponentially.”
Khosla duly visited the founders at their office on the corner of El Camino and Castro in Mountain View. He liked to think of venture bets as financial options. You could never lose more than your initial stake, but the upside was unbounded. Given what the power law meant for startups, what Moore’s law meant for computing power, and what Metcalfe’s law meant for networks—and given how each law compounded the effect of the others—Mosaic Communications was one of those options you just had to have. After the meeting, Khosla called Doerr. “We should just do it,” he told him. A few days later, Clark and Andreessen returned to pitch to the full Kleiner Perkins investment committee. There had been no Accel-style prepared-mind planning, but that did not matter: it took the Kleiner Perkins partners all of forty-five minutes to approve the investment. “We knew it was a high price,” said one partner, “especially with what seemed like a twelve-year-old as the technology guru behind it.” But everybody around the table remembered another one of Tom Perkins’s dictums: you succeed in venture capital by backing the right deals, not by haggling over valuations.
网景的投资真是幂律定律的最好阐述，受益于摩尔定律、Metcalfe定律，网景是双重加成，所以即便KP没有Accel那样深度的研究，但还是非常迅速给出了投资的建议，仅仅就是无法拒绝。即便价格很高，Doerr还是非常清楚，好的投资关键是项目，不是价格。