Frustrated with interconnect delays? Batch computing offers a solution


By Phillip Ng, Vice President of Corporate Development, Soluna

Our energy grid is rigid. It was built a century ago for a different country and a different energy economy. Integrating this system with the new intermittent generation was difficult, to say the least. This results in frequent periods of renewables oversaturation, but still relying on carbon-producing assets to meet peak demand.

Our noble and necessary goal of decarbonizing the grid will only be achieved if we make renewable sources economically viable and if we modernize both the grid and the demand side according to the renewable energy production profile.

In addition to intermittent renewable energy generation, solar and wind, unlike fossil fuels, cannot yet be reliably stored for later deployment.

Result: Demand must become more flexible to resolve this mismatch.

A higher supply than demand means a reduction, ie a waste of energy. In 2021, approximately 14.9 TWh of otherwise viable renewable energy has been curtailed. This equates to $610 million in lost revenue, or enough energy to power the city of Chicago for a year.

This problem is getting worse as renewables increase their share of the next-generation market. Additionally, unless we manage to create more flexible demand, marginal solar and wind power will not help increase net capacity on the grid:

Wind penetration is the X axis; Utility to the grid measured by capacity credit is the Y axis. As penetration increases, marginal utility decreases. Source: Lori Bird, M Milligan and Debra Lew. 2013. Integrating Variable Renewables: Challenges and Solutions. Technical report. NREL.

Based on our analysis, looking at IPP production data, we estimate that up to 40% of the energy generated by individual solar and wind farms is wasted. This creates massive economic pressure on these power producers. It also prevents the construction of new projects in areas rich in wind and solar resources.

Batteries and transmission offer long-term solutions to erasure, but are not sufficient to meet this challenge for several reasons:

  • Batteries are expensive and depend on a supply chain of rare minerals that is both bottlenecked and dry. They have a short lifespan and are difficult to dispose of responsibly.
  • Transmission updates are expensive, time-consuming, and hampered by the lack of a unified national network in the United States
  • Both technologies are still emerging, promising for future use, but require development and are not immediately deployable at this time.

There is a third option: batchable computing.

Batch computing, or batch processing, is computing that can be processed in batches at a time, ideal for large data sets of repetitive inputs that are not time-sensitive. The calculation process began to be used on a large scale at the end of the 19th century, when Herman Hollerithfuture co-founder of IBM, invented a tabulating machine that used batch processing to count census results.

The beauty of batchable computing is that it is interruptible, allowing it to run on intermittent renewable energy. Collocating data centers that run batch computing with renewable power plants creates cover for Independent Power Producers (IPPs) who otherwise only sell power to the grid. When supply exceeds demand and the grid cannot absorb more megawatts, IPPs on the Soluna grid can sell their excess power to the data center.

This solution greatly alleviated the reduction problem.

Case study: 150 MW wind farm

Soluna partners with IPPs to deliver an integrated solution that marries network decarbonization goals and critical IT advancements. Soluna buys the energy that would otherwise be wasted by these IPPs and uses that energy to power compact data centers that run batch computations.

Currently, we operate two locations in Kentucky: Project Mary and Project Sophia (for Marie Curie and Sophie Wilson – all our sites are named after women scientists who have left an indelible mark on history).

We recently opened our new location in Texas, Dorothy Project (named after Dorothy Vaughan, a NASA program mathematician who sent the first American satellites into space).

Batch computing is immediately deployable and scalable, providing a reduction solution that’s ready now. To find out more, read our white paper on reduction mitigation and network stabilization through batch computing.


Phillip leads development activities for Soluna.

Previous "Athletic" Vols Hit Practice Field For First Time This Preseason
Next Commonwealth Games 2022 Day 5 Highlights: India wins two more Golds; Medal tally rises to 13