[Lecture Notes by Prof Yuen and possibly others 2000-2001] Akamai ------ Since early 2000 I have been keeping track of the share prices of several high tech companies, in particular Akamai, founded by Tom Leighton, a well known MIT professor in Computer Science. It is particularly intriguing for two reasons: first, I know several people in Akamai personally; its director of research, Bruce Maggs, provided me with access to computer facilities at Carnegie Mellon during visits to see my children; Charles Leiserson, who heads one of the research groups, was Shaw Professor in our Department in 95-6, and employed my PhD graduate Feng Ming Dong, who was working in our Department as Lecturer; second, I was curious to know whether its business model was based on something real or just hot air, in contrast to companies like Amazon or Yahoo whose income streams are much better defined even though profitability may not be. Would companies really want to pay to put their web contents on someone else's, supposedly much more accessible, servers, or will Akamai be able to, like the web hosting companies, generate alternative income while providing the service free or at least, at a loss? Does it really have superior algorithms to handle content caching much more efficiently than others, including all the competitors that would move in once the market is shown to be there? Sure it has some patents, but these days patents are often awarded first come first served, with priority to be decided later by litigation, rather than before the award; indeed there is already a court case between Akamai and Digital Island. Akamai lost US$40M in 1999, with revenue of just US$10M; yet its shares rocketed to $117, nearly five times the IPO price, on the first day of trading in October 1999. Its 2000 revenue was $90M, for a loss of $180M, and projected revenue for 2001 is $240M, loss at $140M, i.e., total forecasted expenditure is ~$400M. It has grown to over 1000 employees in less than three years. During my latest CMU visit in November 2000, it was still aggressively recruiting CS and ECE graduates. However its shares, hitting $300 in March 2000, had by then fallen to $50, but the management seemed to have taken this in its stride - 300 was ridiculous; 50 more realistic and the founders and venture capitalists were still well ahead, but 2001 is going to be shocking. Despite the Greenspan interest rate cut, which helped the market as a whole to stay more or less level in January/February, Akamai shares fell below $8 in mid March, significantly further than Amazon, Yahoo and other not yet profit making companies did. In April it saw the price down to $5.50, and announced a 14% staff cutback. The price recovered considerably since then, but declined again in August and September to reach new lows whereas the other two shares held up better. Compounding the problems, the cofounder of Akamai died in a terrorist plane crash in September. In October another layoff, of up to 25% of staff, was announced, though its shares rose with expectation that future losses would be less. The list of Insider Trading for Akamai makes interesting reading. Tom Leighton started selling in August 2000, when his share holding was around 4M; when he sold again two monthes later, the holding was $8M - he appears to have exercised additional share options with the money from the earlier sales, presumably expecting the price to recover from its recent falls; by exercising options when the price is low, you minimize your tax liability, since the difference between the exercise price and the market price on the day is considered to be your income and will be taxed, even if you have not actually sold the shares - and he has been selling almost continuously since then, presumably to raise cash to meet the tax liability. At the peak, his 8M share options would have been worth more than two billion; even today, he is on paper a very wealthy man. Others have done less well; Bruce Maggs sold all his shares, 30,000, at $50 in November 2000 (though he may be entitled to purchase more with options that have not yet been exercised, but the exercise prices are probably above the current market prices and so it would not be practical); others, with positions lower than director of research, would presumably have been allocated less shares. The drop in Akamai shares after the release of the financial report for 2000 and forecasts for 2001 in late January, seems particularly ominous. Previously, Akamai shares tended to bounce back strongly after each big drop, meaning that there are investors waiting to buy in at what appears to be bargain level each time; for nearly a month, this did not occur. There appears to be some doubt on the forecasted revenue figures as well as the values of the past investment in equipment: if all the servers Akamai bought and placed in locations around USA and the world do not generate revenue as expected, then their capital value in the books will eventually have to be written down, generating huge losses against future accounts. What if the business model proves to be unworkable? With over 1000 highly paid employees, thousands of servers, and high rents, Akamai is due to burn through its cash, obtained from the issue of shares (100M shares are outstanding in total) and bonds (a $300M convertible bond issue is on record), towards end of 2001, and will need to raise new capital from investors; in the current market conditions, and particularly in view of its recent share price drop, this could prove to be difficult. Whereas in the case of Amazon, also due to run out of cash this year, a buyout by a bargain hunter is feasible, in order to control its first class shopping website and huge customer base, it is much more difficult to envisage the same with Akamai, whose main attraction has always been the MIT/CMU professors and PhDs on its payroll, but they, unlike the Amazon website and customer database, are mobile, and are unlikely to stay around once the company changes hand. How do the highly qualified employees feel about Akamai today? Unlike the founders and venture capitalists, they do not have share options at low initial prices; they have given up, or at least suspended, promising careers in universities and research laboratories in order to take part in this great internet goldrush, and helped the early birds to become very rich without clear benefit for themselves. For all the talk about advanced caching techniques, it is not clear that the work most Akamai employees do can be called research and development, rather than straight systems programming. As the sun sets on Temptation Island, it remains to see one last effort to be made by the Pied Piper of Akamai to lead his followers to the promised land.