ShowDaily Day3 - Flipbook - Page 3
sponsored by
upgrades, and proenvironmental impact targets
viding services that
through the Science Based Targets
help meet its goals.
initiative (SBTi) and engage in
Citing SEMI and
transparent reporting in line with the VLSI Research
recommendations of the Task Force
predictions, Dickon Climate-Related Financial Disclo- erson said the
sures (TCFD). And he said Applied’s semiconductor
suppliers would sign agreements
industry’s revenues
to adopt similar environmental
will approach a
initiatives by 2030, adding that six
trillion dollars by
suppliers are already on board the
2030, a growth rate
“Success 2030” plan.
double the historical
rate. If the capital
A more sustainable industry
investment intensity for wafer fab
Dickerson did not stop there, anequipment — WFE intensity — renouncing a “3 by 30” goal for
tains its recent five-year 11.5 percent
its manufacturing systems: on a
rate, then it is likely the semiconper-wafer basis, Applied targets a
ductor equipment industry overall
30-percent reduction in equivalent
will have annual revenues of $100
energy consumption along with a
billion to $150 billion by 2030.
30-percent reduction in chemical
But that growth must not conconsumption, and a 30 percent
tribute to global warming.
increase in throughput density —
“We must decouple our growth
which is the number of wafers profrom our environmental impact. If
cessed per square foot of cleanroom
we double or triple the size of our
space — by 2030. Those numbers are company, it would be irresponsible
set against a 2019 baseline.
to double or triple our environThe company’s “ecoUP” initiative
mental footprint. So we are making
includes improving the eco-pera commitment to new sustainability
formance of its existing and new
targets, which we believe are agsystems with hardware and software gressive but achievable,” he said.
Dickerson
continued from page 1
www.semiconductordigest.com
A more sustainable world
The prospects for such rapid growth
for the equipment and semiconductor manufacturing industries
are largely based on “the enormous
promise” of artificial intelligence to
accelerate research in areas such as
climate change, disease prevention
and public health. But Dickerson
pointed to the “Achilles heel” of
AI: the growing amount of power
consumed to train neural networks
with visual data.
Pointing out the rapid growth in
data “generated by machines and
consumed by machines,” Dickerson
said “AI and the data economy will
change our economy and our lives,
and the very foundation of competition will be transformed.”
The growth in data generation and
storage has the potential to impact
worldwide energy consumption, he
said, ratcheting up the energy consumed by data centers to 15 percent
of worldwide energy by 2030.
“Training neural networks is incredibly energy intensive when done
with the technology that is available
today,” Dickerson said, and bending
down that trend line will require
innovations in how AI data is processed, stored, and transmitted.
At a panel discussion following
Dickerson’s keynote, Continued on page 4
Wednesday, July 22 | 3
