Breaking 1,000 Quantum Bits by 2025, Rigetti Unveils New Quantum Computing Roadmap

At last week's Investor Day, Rigetti Computing described a detailed product roadmap - covering all aspects of enterprise activity [1]; and announced new partnerships that will help them achieve their "Quantum Advantage " goal.

 

"As a pioneer in the quantum space, Rigetti is committed to delivering performance at scale, with the goal of becoming the industry standard," said Chad Rigetti, the company's founder and CEO [2]: "We are making strategic investments in quantum hardware, software and partners. We believe this will enable us to make progress toward quantum dominance."

 

"In addition, we are announcing several important partnerships, including, among others, a partnership with Bluefors to develop new modular dilution coolers to support our planned 336-qubit, 1,000+-qubit and 4,000+-qubit quantum processing units. In addition to this, we are also making our current 80Q Aspen-M-2 and 40Q Aspen-11 systems publicly available on Microsoft's Azure Quantum. Together with Amazon Braket, Rigetti quantum computers are now available on two of the largest public cloud platforms in the world."

 

32d15009fa461f9a0d1ae84808671218

 

01Massive Performance Boost: Breaking 1000 Quantum Bits by 2025

 

1)       Update the hardware roadmap again

 

dc894478fc51daeec6067a402eb99032

 

Rigetti Computing Roadmap

 

In May 2022, Rigetti provided a brief description of the QPU roadmap and now they have provided some additional information about the roadmap. In terms of quantum bit count, their next processor - codenamed Ankaa - will have 84 quantum bits in a single chip and is expected to be available in early 2023. The chip will use their fourth-generation architecture, which uses tunable couplers and a square lattice topology, and will offer significant improvements in gate fidelity and quantum bit connectivity compared to third-generation QPUs. The tunable coupler will allow better isolation between adjacent quantum bits and will also enable Rigetti to improve the processor's gate latency; the tunable coupler was first demonstrated by Google in its Sycamore processor, and Rigetti says it aims to achieve 99.5 percent, or better, gate fidelity with this new processor, which would be a significant improvement.

 

The new square lattice architecture will improve the connectivity of the average quantum bit and allow each quantum bit to interact with four adjacent quantum bits, an improvement over the 2 or 3 quantum bits in the current Rigetti third-generation architecture. This will make it easier to embed specific quantum programs in the processor and reduce the number of SWAP gates needed to move quantum bit information. It may also make it easier to implement various error-correcting codes: surface codes, for example, which will work better with a square lattice topology.

 

Following the launch of the 84-quantum-bit processor, Rigetti plans to launch a 336-quantum-bit processor later in 2023, codenamed Lyra. Lyra will consist of four Ankaa chips: they will be assembled on a carrier substrate. The technology will take advantage of a key technique that Rigetti has long been working on: one that allows information about entangled quantum bits to be exchanged between chips with no loss in speed or fidelity compared to connections between different quantum bits on the same chip.

 

After launching Lyra with 336 quantum bits, they are planning to launch a QPU with 1,000+ quantum bits in about 2025 and a QPU with 4,000+ quantum bits in about 2027.

 

Both QPUs consist of multiple 84-qubit chips, but they will assemble more chips on carrier substrates, or will reach dozens; to achieve a higher number of quantum bits, the 4000+ QPU will utilize multiple chillers, which will likely be a distributed quantum computer. This is different from IBM, which has developed the world's largest dilution cooler, the Goldeneye, for 4000+ quantum bits.

 

Although Rigetti does not expect to increase the number of quantum bits on the Ankaa chip in the meantime, they will likely continue to refine it in subsequent revisions to provide continuous improvements in gate fidelity, coherence time and other parameters. They will also need to implement larger dilution chillers, improve flex I/O, and provide next-generation control systems for these higher capacities.

 

2) Expansion of their own fab

 

Rigetti's manufacturing strategy is very different from most of their other quantum startup competitors. From their early days, they decided to acquire their own fab to build their chips, rather than using university clean rooms or trying to partner with a large semiconductor foundry like GlobalFoundries. To that end, they opened their own manufacturing facility in Fremont in May 2017, called Fab-1. Using their facility, they were able to complete their design in 5-15 weeks; whereas if they had used an outside semiconductor foundry, it would have taken an estimated 22-40+ weeks.

 

Recently, Rigetti has been manufacturing devices not only for themselves, but also for Fermilab, the U.S. Air Force Research Laboratory (AFRL), and Purdue University. Their main purpose in doing so is not to generate significant new revenue by acting as a foundry, but to gain engineering feedback from their organization that will help them with their own processing and processor development. rigetti is a major commercial partner in the Center for Superconducting Quantum Materials and Systems (SQMS), led by Fermilab.

 

In order to increase Fab-1's capacity, Rigetti is starting a Fab-1 expansion project in early 2022: an additional 5,000 square feet, which will almost double its size. The company expects the project to be completed by the end of 2022, which will allow them to handle the larger volumes of wafers needed to develop the large QPU mentioned above.

 

02New Partnerships

 

1) Collaboration with Bluefors

 

Larger multi-chip processors will require larger dilution chillers with greater cooling capacity and modular designs to facilitate multi-chiller implementation. To achieve this, Rigetti has formed a partnership with Bluefors to develop a new chiller design called KIDE, which is also used in the IBM Quantum System Two. The company expects to receive the first prototype of KIDE in early 2023, followed by delivery in late 2023, to support Lyra's 336-quantum-bit multi-chip design.

 

b42963c89cdefee7ca2ed4c354a4bf1b

Bluefors KIDE's dilution chiller model

 

2) Collaboration with Yestech on error mitigation tools

 

An important requirement for achieving quantum advantage is to provide processors with the lowest possible error rate. While this has a lot to do with hardware architecture at the chip level, and implementing and controlling electronic circuits, software can also play a big role. in May 2021, Yestech acquired a Canadian startup called Quantum Benchmark, which developed error mitigation software, True- Q. As part of the investor day announcement, Rigel acquired a Canadian startup called Quantum Benchmark, which developed error mitigation software, True- Q. Q. As part of the investor day announcement, Rigetti announced that they will integrate True-Q software into their QCS platform: they will offer this error mitigation capability to users as a built-in feature.

 

3) Collaboration with Ampere and Nvidia

 

Rigetti was also an early proponent of hybrid classical/quantum processing, providing effective solutions to difficult computational problems. in February 2022, they announced a partnership with classical microprocessor startup Ampere to integrate Ampere's extremely high-performance chips into QCS: these Ampere processors should provide a more efficient solution to the classical part of the algorithm than any other company using more traditional microprocessor chips and provide Rigetti's end users using its QCS with advantages and faster overall uptime. rigetti has now started another program with nvidia to develop a hybrid GPU-QPU workflow for climate modeling applications; the goal is to apply quantum machine learning in a high-performance hybrid workflow technology, and build on Rigetti's recent research on weather modeling applications.

 

Reference links.

[1]https://quantumcomputingreport.com/rigetti-announces-detailed-product-roadmap-and-new-partnerships/#

[2]https://www.globenewswire.com/news-release/2022/09/16/2517634/0/en/Rigetti-Announces-New-Partnerships-Provides-Business-Updates-at-Inaugural-Investor-Day.html

2022-09-20