High-Efficiency Data Transfer

Default ARPA-E Project Image

OPEN 2015
Santa Barbara, California
Project Term:
04/07/2016 - 04/06/2019

Critical Need:

The explosive growth of the internet has been accompanied by a similar increase in the amount of data that is transmitted and processed. Much of the burden of this growth has been shouldered by datacenters, one of the largest and fastest growing consumers of electricity in the country. Datacenters in 2013 consumed an estimated 91 terawatt-hours of electricity, enough to power all the households in New York City twice over. This is about 2.5% of electricity produced in the United States, and with data traffic projected to increase 23% annually over the next five years, datacenters’ energy consumption is projected to double in about eight years. Increased bandwidth is needed to accommodate the anticipated growth in traffic. However, simply increasing the bandwidth with current communication systems would result in reduced efficiency and increased waste heat. Immediate solutions are needed to reduce the bandwidth bottleneck while increasing efficiency.

Project Innovation + Advantages:

The University of California, Santa Barbara (UCSB) will develop a new technology for optical communication links. Optical interconnects transfer data by carrying light through optical fibers, and offer higher bandwidths than copper with higher efficiency and, consequently, reduced heat losses. However, short-reach optical interconnects are not widely used because of their higher costs and larger device footprints. Production costs of these interconnects could be reduced by using silicon-based fabrication technologies, but silicon is not suited for fabricating lasers, a key ingredient. In contrast III-V semiconductors, are well-suited for fabricating highly efficient lasers, but at a high cost. The team plans to combine these components to create III-V lasers, grown on a silicon substrate, harnessing both the low cost of silicon and the superior laser of the III-V semiconductor. However, growing the III-V laser material directly on silicon is difficult due to incompatibilities in their crystal structures. The team aims to overcome this challenge by implementing nanostructures called "quantum dots" as the light producing material and by growing the structure on patterned silicon substrates to help contain potential defects.

Potential Impact:

If successful, innovations from this project could dramatically improve the cost and efficiency of optical interconnect technology and our ability to transmit greater amounts of data more efficiently.


Faster data communications will enhance national defense, security, and resiliency. 


Increasing the energy efficiency of data communications could significantly reduce CO2 emissions associated with powering datacenters.


Implementing the team’s technologies could save large amounts of power, and the increased performance and bandwidth of optical networks could improve economic productivity.


ARPA-E Program Director:
Dr. Michael Haney
Project Contact:
Dr. John Bowers
Press and General Inquiries Email:
Project Contact Email:

Related Projects

Release Date: