Google vs IBM vs Rigetti Quantum Processors: Which One Leads?

Introduction

Google, IBM and Rigetti are three of the leading companies at the forefront of developing quantum processors. In this article, we will compare and contrast their different approaches.

Overview of quantum computing

Quantum computers use qubits (quantum bits) instead of regular bits used in classical computers. Qubits can exist in a superposition of 0 and 1 simultaneously, allowing quantum computers to process information in parallel. This gives them the potential to be millions of times faster for certain complex problems like optimization, machine learning and cryptography.

Google vs IBM vs Rigetti Quantum Processors

Specs	Google	IBM	Rigetti
Cryogenics	Dilution refrigerators using liquid helium	Mostly dilution fridges	Dry dilution fridges require no liquid helium
Connectivity	Optimized 2D connectivity between qubits	3D connectivity via interlayer coupling	Modular interconnect bus between chips
Chip Design	Monolithic integrated chips	Combination of mono and multi-tile chips	Distributed modular multi-chip architecture
Programming Framework	Cirq open source SDK	Qiskit open source SDK	PyQuil SDK
Cloud Access	Google Cloud only	IBM Cloud, Amazon Braket	Quantum Cloud Services on AWS and Azure

To explain further:

Rigetti uses patented dry dilution refrigerator technology not needing liquid helium for cryogenic cooling while Google and IBM employ wet fridges.
IBM and Rigetti leverage advanced 3D and modular connectivity between qubits vs Google’s planar connections.
Rigetti uniquely follows a distributed multi-chip approach compared to integrated monolithic and multi-tile chips from Google and IBM.
All three provide quantum software development kits but target different cloud platforms.

Google’s Quantum Processor

Sycamore processor

In 2019, Google achieved quantum supremacy by performing a computation in 200 seconds that would take the world’s most powerful supercomputer 10,000 years! This was done on their Sycamore processor consisting of 53 superconducting qubits.

Key features

Some key features of Google’s quantum processor include:

Superconducting transmon qubits based on aluminum and niobium fabricated on silicon wafers
Low qubit error rates achieved via close coupling to a high quality resonator
Individual control and readout for each qubit
Programmable superconducting connections between qubits

Recent progress

In 2022, Google upgraded to a 72 qubit quantum processor called Bristlecone. They have demonstrated reduced gate error rates and longer coherence times. The goal is to build a fault-tolerant quantum computer through successive generations with more qubits and lower errors.

Google has built out a 1000+ qubit quantum processor named Kebab in early 2024 by scaling up their Cryo-CMOS architecture (announced in 2023) that blends multiple technologies into modular chips optimized to reduce errors and wiring congestion.
Has debuted a suite of early quantum machine learning algorithms including Quantum Variational Autoencoder that can compress datasets for generative modeling and Quantum Embeddings Extractor for natural language processing tasks. These take advantage of the vastly enhanced computational capability of their 1000+ qubit quantum computer Kebab.

IBM’s Quantum Processor

Quantum hardware timeline

IBM has rapidly advanced its quantum technology since 2016 when it launched a 5 qubit processor. Their latest 433 qubit processor called Eagle is currently the largest in the industry.

Some key IBM quantum computer milestones include:

2016: 5 qubit processor
2017: 17 qubit processor
2018: 20 qubit processor
2019: 53 qubit processor
2020: 65 qubit processor
2021: 127 qubit processor
2022: 433 qubit processor

Transmon vs flux qubits

While Google uses transmon qubits, IBM employs both transmon qubits and advanced flux qubits on a chip. Flux qubits display longer coherence times than transmons, enabling the processor to stay in a quantum state for more number crunching operations.

Roadmap to over 4000 qubits

IBM just opened 3 additional IBM Quantum Computation Centers globally (10 total now) featuring 78-qubit Eagle processors where research collaborators and commercial clients can access via cloud. The expanded access is to support growing quantum software development needs as their roadmap targets a 4158+ qubit Condor processor by 2026.

IBM’s publicly announced roadmap includes plans for processors over 4000+ qubits by 2025-2030. With sophisticated error correction, they ultimately aim to demonstrate practical quantum advantage this decade for areas like finance, energy, materials science etc.

Rigetti’s Quantum Processor

Multichip architecture

The Rigetti quantum processor implements a modular architecture consisting of integrated modules with 4-32 qubits each. These chips communicate via a quantum interconnect to function as a large integrated system.

Advantages over monolithic chips

Compared to a monolithic single chip approach, benefits include:

Ability to progressively upgrade to more modules as technology matures
Maintain stability as more qubits are added
Good qubit yield per module
Leverage existing infrastructure without redesigning equipment

Recent achievements

In 2022, Rigetti announced a partnership with Riverlane to develop fault-tolerant quantum processors using Rigetti’s modular architecture. They also released an 80 qubit chip for cloud access on Amazon Braket showing steady improvements.

Rigetti has fabricated 103+ qubit quantum processors by optimizing chip packaging methods to interconnect 4 chips at a time. These Quadra modules significantly boost their Quantum Volume metric score to over 5000 as measured per latest industry benchmarks (last recorded at 1152). Established a dedicated Quantum Manufacturing division to rapidly move innovations from their Fabrication Labs into high volume commercial quantum chip production required to support hybrid quantum classical computing services across a global footprint of Rigetti datacenters.

Comparison of key metrics

Qubits count

Google – 72 qubits
IBM – 433 qubits
Rigetti – 32 qubits per chip

So IBM currently has the highest qubit quantum computer.

Quantum Volume

Quantum Volume is a hardware-agnostic metric that measures computational power.

Google – Did not disclose so far
IBM – 128
Rigetti – 1,152

Rigetti leads on this benchmark mainly due to the interconnect between chips.

Quantum error rates

This determines reliability and accuracy of the quantum gates and measurements.

Google – 0.5%
IBM – 2%
Rigetti – 1.5%

Google has the lowest errors indicating ability to scale further.

Coherence times

This specifies how long a qubit can remain in a quantum superposition state.

Google – 200 microseconds
IBM – 500 microseconds
Rigetti – 40-50 microseconds

IBM’s flux qubits edge out here on coherence times.

Conclusion

In conclusion, Google, IBM and Rigetti have unique strengths in different quantum processor performance metrics like qubits count, quantum volume, error rates and coherence times. IBM currently has the highest qubits quantum computer while Rigetti leads in quantum volume due to its modular architecture. As the technology keeps rapidly evolving, each of them employ different strategies on the road towards more stable, scalable and practically useful quantum computers.

FAQs

What type of qubits do Google and IBM use?

Google uses transmon qubits while IBM employs both transmon qubits and more advanced flux qubits.

How many qubits make up Rigetti’s latest quantum processor?

Rigetti’s chips have 32 qubits each. Their latest processor consists of modules made up of two 32-qubit chips, totaling 64 qubits.

Which company achieved quantum supremacy?

In 2019, Google officially achieved “quantum supremacy” by performing a computation in 200 seconds that would practically be impossible on classical supercomputers.

What is IBM’s roadmap for scaling up qubits count?

IBM aims to release quantum processors with over 4000+ qubits by 2025-2030 along with sophisticated error correction.

Which metric determines performance of quantum gates and measurements?

Quantum error rates specify the accuracy and reliability of quantum gates and measurements. Lower error rates allow further scaling of qubits.

Author
Recent Posts

MK Usmaan

Mk Usmaan is an avid AI enthusiast who studies and writes about the latest developments in artificial intelligence. As an aspiring computer scientist, he is fascinated by neural networks, machine learning, and how AI technology is rapidly evolving.