1 of 21

Smart Current

التيار الذكي

BLACKOUTS ARE NOT JUST AN ANNOYANCE

2 of 21

3 of 21

Middle East North Africa is the 2nd highest Geo-economic region in both frequency and duration of power outages

SDG affected

strongly

unaffected

medium

16.45

Average Blackouts

per Month

avg. 4.9 hours

“Characterized by unreliability and inefficiency” (De Gruyter)

>700Bد.إ

Yearly economic loss

4 of 21

Electrical grids

Quantum Topological Data Analysis

Regulate power-flows

Graph neural network

“Classical features”

Quantum enhanced

It has been shown that optimal power flows lead to

60% reduction in blackout risk.

Most blackouts are avoidable.

JEPE

1

1

5 of 21

Electrical grids

Quantum Topological Data Analysis

Regulate power-flows

Graph neural network

“Classical features”

Quantum enhanced

We provide automated real-time powerflow corrections to avoid and mitigate blackouts.

However, optimization methods are too slow to respond in real time.

6 of 21

Overview

Electrical grids

Quantum Topological Data Analysis

Regulate power-flows

Graph neural network

“Classical features”

Quantum enhanced

Online blackout response dashboard

Open-source quantum circuit implementations

Blackout response and prevention machine learning models

smartcurrent.com

7 of 21

Smart Currents

8 of 21

Smart current – Our innovation, our mission

Electrical grids

Quantum Topological Data Analysis

Regulate power-flows

Graph neural network

“Classical features”

Quantum enhanced

Electrical grids

Quantum Topological Data Analysis

Regulate power-flows

Graph neural network

“Classical features”

Quantum enhanced

Enhanced ML approach

using power grid geometry

Train a graph neural network with emphasis on neighbor connectivity

9 of 21

Representing Graphs as Topological Surfaces

10 of 21

Quantum Speedup of the Genus/Betti number approximation

Classical-Complexity

Quantum-Complexity

Super-Polynomial

Polynomial

11 of 21

Electrical grids

Quantum Topological Data Analysis

Regulate power-flows

Graph neural network

“Classical features”

Quantum enhanced

Our end-to-end quantum-inspired models for power flows can be found on Github.

We’ve also created the first open-sourced implementations of two circuits.

Quantum TDA Boundary Operator

Quantum Orthogonal Neural Network

12 of 21

Hardware Verification

Noisy simulation

IBM Guadalupe

Candice, 30s

13 of 21

Business

Electrical grids

Quantum Topological Data Analysis

Regulate power-flows

Graph neural network

“Classical features”

Quantum enhanced

Market (MENA)

Yearly investments in power 80+Bد.إ

Value proposition

Efficiency improvements and risk minimization for grid operators

No direct competitors, indirect competitors can be partners

14 of 21

Roadmap

6 months

Training

Validation on DR POWER dataset (over 1TB)

Cost ~د.إ20,000

Infrastructure

AWS Sagemaker API

IBM Quantum Cost ~د.إ150,000

Business

Create partnership with one key MENA government

Personnel

Yearly Cost ~د600,000

3+ years

Training

Addition of government-specific databases

Cost ~د.إ120,000

Infrastructure

Govt. On-premises, IBM Quantum, AWS

Cost ~د.إ1,500,000

Research

Investigate other topological properties

15 of 21

16 of 21

Sohum Thakkar,

United States

Henok Daniel, Eritrea

Yaffa Jaradat, Palestine

Elias Huber, Switzerland

Allen Baranov,

USA

Omar Al Khazali,

Jordan

Kannan Vijayadharan,

India

Candice Kang, China

Haya Fuad,

Syria

Stephen Lang, USA

17 of 21

Sohum Thakkar,

United States 🇺🇸

Henok Daniel, Eritrea 🇪🇷

Yaffa Jaradat, Palestine 🇵🇸

Elias Huber, Switzerland 🇨🇭

Allen Baranov,

USA 🇺🇸

Omar Al Khazali,

Jordan 🇯🇴

Kannan Vijayadharan,

India

Candice Kang, China 🇨🇳

Haya Fuad,

Syria 🇸🇾

Stephen Lang, USA 🇺🇸

🇮🇳

18 of 21

Overview slides

19 of 21

Value Proposition Breakdown

Problem – Blackouts

  • > 200$B economic loss in MENA
  • Up to 60% risk reduction with better flow-management
  • Classical GNNs can help, but not capture all topological features

Solution – Business value in MENA

1

1

2

JEPE

2

Bring the power of quantum to the manage energy grids

Adapt to the need of our partners

Collaboration across MENA for a better future

Market: Yearly investments in power 80+Bد.إ

Client proposition: Efficiency improvements and risk minimization for grid operators

No direct competitors, indirect competitors can be partners

3

Frost & Sullivan

3

6 months (training & partnership)

3+ years (Deployment & Research)

Phase 1

Cost~د.إ500,000

Phase 2

Cost~د.إ20,000,000

20 of 21

Business | How our innovative solution creates value

Competitors

1

Value now – Implement our classical graph neural network and further develop our quantum enhancements in cooperation with industry partners

2

Leverage the power of quantum computers as hardware matures

  • Topological data analysis
  • Quantum enhanced neural networks

Bring the power of quantum to the manage energy grids

Collaboration across MENA for a better future

Our approach based on graph neural networks and close collaboration with grid-managers allow tailored use-cases beyond grid-management: Grid planning and maintenance, efficiency optimization, stability evaluation

Adapt to the need of our partners

Collaborate across MENA with governments and non-government entities. Establish local quantum competences.

Competitor

Description

Voltus

  • Focus on energy distribution
  • Technology platfrom
  • Focus on profit maximization

SPINE LLC

  • Consultancy
  • Broad support beyond grid management

Grid-operators

  • Could implement inhouse management software

While potential competitors exist …

  • None provide our quantum expertise
  • Use optimization, with slow turnaround
  • Few of them in the MENA region

Different business focus and all of them could be business partners!

Improve and save the lives of ordinary people

21 of 21

How did we test?

We can run some part of the algorithm classically.

Classical runtime

TDA

Compound GNNs

Dim 1

Dim 2

Dim 3

Dim 4

Dim 5

Dim 6

Dim 1

Dim 2

Dim 3

Dim 4

Dim 5

Dim 6

Feasible zone ☺️

Need quantum 🙄