Markets
Five More AI Companies Line Up Behind Aleph
Rival Lab, Chip Designer and Drone Makers Expected to File Before Spring; Lattice Seeks $4 Billion in Notes
VETERANS RECALL 1999, AND THE RAILWAYS
Halcyon Has Nearly Doubled Since March; Old Hands Note That All Times Are Subject to Change
NEW YORK, Oct. 12 — Aleph Superintelligence had not yet made its first trade on Monday morning when the bankers who sold its shares began talking about the next ones.
Five more artificial-intelligence companies are expected to file for initial public offerings before spring, according to bankers involved in Aleph’s record $62 billion sale, and some could be trading by then. They are Parallax, an Austin designer of AI chips; Moraine Labs of London, Aleph’s closest rival; Kestrel Systems of El Segundo, Calif., which builds autonomous systems and drones for defense; Wren Aerial, the Dallas drone-delivery company; and Sable Health, a Boston maker of clinical AI.
The bankers, who spoke on the condition that they not be named because none of the five has filed publicly, differ on the order of departure. They do not differ on the destination. “Every chief executive in this business was watching the screen at 11:52,” one said. “Three of them had called me by noon.”
In the trade the moment is called a window, but it is really a timetable: a list of departures, an expected order and a line of small type warning that all times are subject to change. The bankers’ expectations are set out below in that form.
Wren Aerial had the best-timed Monday of the five. At 12:01 a.m. a federal rule permitting routine drone flights beyond an operator’s line of sight took effect, and Wren’s 1,100 drones in Dallas–Fort Worth began flying under it.Part 108The Federal Aviation Administration rule for beyond-visual-line-of-sight flights below 400 feet, in managed corridors. One remote operator may supervise up to 40 aircraft. By 4 p.m. the region’s drone corridors had logged 1,412 flights without an injury, which is the kind of first chapter a prospectus writer would not dare invent.
The companies already on the board have done much of the persuading. Lattice Compute, which rents out graphics processors by the hour and went public in June at $54 a share, filed on Monday to sell $4 billion of convertible notes.Convertible NotesBonds the holder may exchange for stock if the share price rises past a set level. In return for that option, buyers accept a lower rate of interest. Its shares, at $88¼, are up 63 percent since the offering.LTCENasdaq881⁄4▼13⁄8−1.5%Lattice Compute
“When a company sells converts after a run like that, it is telling you two things,” said Paulette Nwosu, who trades convertible bonds for a fund in Greenwich, Conn. “It thinks the stock is going higher. And it would rather not sell any more of it at this price.”
Halcyon Robotics, a maker of warehouse and humanoid robots, has nearly doubled since its March debut at $72 and now trades at $141½.HALONYSE1411⁄2▲37⁄8+2.8%Halcyon Robotics Bankers say its chart has done more to fill the queue than any pitch book, because every board in the business has seen it, and most have it memorized.
Everyone on Wall Street is reaching for the same year this week, and nobody likes it less than the people selling the shares. The bankers’ answer to comparisons with 1999 is revenue: Aleph took in $21.3 billion in the 12 months through June, while many companies of 1999 took in next to nothing. The reply from older investors is that Aleph also lost $17.8 billion over the same 12 months, and that revenue has never yet stopped anyone from paying too much for a stock.
Investors old enough to remember 1999 and 2000 say the queue looks familiar, and that manias tend to build what they promise while ruining many of the people who pay for it.“In 1999 a young man told me I didn’t understand the internet,” said Bernard Kliegman, 81, who has managed money for a few dozen families from the same office on East 52nd Street since 1974. “He was right. Neither did he.”
The internet, as Mr. Kliegman points out, turned out fine. The Nasdaq composite index did not; it peaked in March 2000 and lost more than three-quarters of its value over the next two and a half years. Much of the fiber-optic cable laid in that boom sat dark for years, then carried the traffic the boom had promised, for owners who had bought it cheaply.
He reaches further back, too. Britain’s railway mania of the 1840s left the country laced with track and a great many of its investors poorer.The Railway ManiaA speculative boom in British railway shares in the mid-1840s. Parliament authorized thousands of miles of new line, much of it never built, and share prices collapsed before the decade was out. American railroads were overbuilt more than once in the 19th century, and many passed through receivership into new hands. Radio stocks were among the darlings of the 1920s and among the hardest hit after 1929.
“Every mania builds something,” Mr. Kliegman said. “The railroads got built. The radios got sold. The fiber got laid. The question is never whether the thing gets built. It’s who is holding the paper when it does.”
The question is never whether the thing gets built. It’s who is holding the paper when it does.
He has not bought any Aleph, and does not expect to buy any of the five behind it. “I have watched a lot of trains leave this station,” he said. “Most of them came back around a few years later, and the fares were a good deal cheaper.”