(Chris Devlin) #1



to a different office; it’s moving because its
lease wasn’t renewed, but it will pay lower
rent. Money at Amprius isn’t spent on
creature comforts. It’s spent on science and
In a lab of the Sunnyvale office is Am-
prius’s crown jewel: a room-sized machine,
designed by Amprius and built in Europe
to its specifications, that applies a mix of
silane gas and other gases to a metal sub-
strate; the resulting chemical reaction cre-
ates the silicon nanowires. Visible through
a peephole in the machine about the diam-
eter of a silver dollar, the gas-application
process is a purple haze. Everything about
it is intricate and finicky: the composition
of the gases; the pressure and temperature
at which they’re shot in; the speed at which
the substrate moves along the conveyor belt
inside the machine.
Once the anode material comes out of
the machine, in a double-sided roll that’s
battleship gray, it packs about 200,000
silicon nanowires per square centimeter per
side. It’s cut and sent into a series of small
lab rooms, where workers in white coats
and blue surgical masks assemble batteries
essentially by hand. Amprius says the best
of these batteries have an energy density
about 60% higher than that of conventional
lithium-ion batteries. One downside is that
they don’t withstand as many discharges
and charges as conventional batteries—
something Amprius is working to improve.
Amprius’s cutting-edge batteries have
piqued the interest of the U.S. Army, which
is testing them for use in clothing that
soldiers might wear to power the devices
they use in the field. By far the batter-

Amprius, a decade-old startup with headquarters in Silicon Val-
ley, most of its operations in China, and investors in both coun-
tries that collectively have pumped about $140 million into the
company. They include Trident Capital and Kleiner Perkins, two
Silicon Valley venture capital firms; SAIF Partners, a Chinese
private-equity firm; and the Wuxi Industry Development Group, a
government-owned investment company in Wuxi, the Chinese city
in which Amprius has a sizable battery factory. Unlike many start-
ups, Amprius is already producing batteries and selling them to
prominent customers. Amprius had about $50 million in revenue
last year, says Kang Sun, the company’s chief executive. But its
technology remains buggy, and its future is hardly assured. “We’re
not out of the woods yet,” he says.
Sun is a tech-industry lifer. He favors coiffed hair, pressed shirts,
and straight talk. He grew up in China, earned a Ph.D. at Brown,
worked his way up to vice president at Honeywell, and then went
back to China to help build JA Solar, now one of the world’s largest
solar-panel makers. Today he lives near San Francisco, drives a
Tesla, and flies seemingly constantly around the world.
His current gig as head of a transpacific battery startup is, he
says, “the most difficult job I’ve had in my life.” Over the hours I
spent with him, one phrase kept popping out of his mouth, mut-
tered almost subconsciously, as if a mantra: “not easy.” As in: “Bat-
tery technology is not easy.”
The source of his lament: the maddening elusiveness of the
The anodes in most lithium-ion batteries are made of graphite, a
substance that’s cheap and plentiful. Amprius, like many other start-
ups, is trying to make anodes from silicon, which, gram for gram,
theoretically can hold 10 times as many lithium ions as graphite can.
“Theoretically” is a colossal caveat. Silicon’s upside as a lithium-ion
hoarder has a major downside too: When silicon is stuffed with lots
of lithium ions, it swells. That swelling can crack the anode material,
dramatically shortening a supposed super-battery’s life.
More than a decade ago, a Stanford materials-science professor,
Yi Cui, developed a new technique to avert silicon swelling in an
anode. It uses a structure of silicon that, at nanoscale, resembles a
single bristle of an upturned brush. Lab experiments proved that,
as each is stuffed with lithium ions, it has plenty of space to swell
without knocking into another bristle and cracking the anode. Am-
prius is the company created to commercialize the concept, known
as “silicon nanowire.”
Sun soon signed on as CEO, figuring he’d spend a few years
building Amprius and then flip it or take it public at a handsome
profit. A decade later, he’s still on the hot seat. “We have to scale up
30 times bigger,” he says. “Otherwise, we cannot make money.”


MPRIUS’S INTELLECTUAL HUB, in Sunnyvale, Calif., the heart
of Silicon Valley, is a bunker-like suite in an unremark-
able industrial park. The walls are scuffed, the furniture
looks rented even though it isn’t, and one day when I
visit, the floor under the men’s-room urinals is lined with card-
board sheets pocked with stains. This summer, Amprius is moving