Quantum computing has long promised to solve problems that classical machines struggle with, from modelling complex molecules to optimising large systems and even applications in materials and drug discovery. Yet, despite decades of research, most quantum computers remain confined to labs, dependent on extreme cooling, fragile hardware, and tightly controlled environments.
As the field matures, many stakeholders now feel that if these machines are to move beyond demonstrations and into real-world infrastructure, they will need to scale, connect and operate within practical limits.
Enter photonic quantum computing. Instead of relying on superconducting circuits cooled to near absolute zero in cryogenic chambers, photonic systems use light particles as qubits.
Supporters of the approach argue that photons—central to fibre-optic communication networks—may offer a more natural route to scale and interconnect quantum systems.
Quanfluence is working on just that. The Bengaluru-based startup is building a photonic quantum computer and chips from the ground up. Its co-founder and CEO, Sujoy Chakravarthy, is a semiconductor entrepreneur turned quantum hardware founder, who believes that the path to usable quantum systems lies in rethinking the architecture itself.
His decision to turn to photons reflects frustration with the current state of the field. “There is no established method of doing things,” he said, pointing to the variety of competing approaches, from superconducting qubits to trapped ions and silicon spins.
Despite years of progress, large-scale machines remain elusive. “Scaling was the biggest limitation in all of these methods,” Chakravarthy said, “which is why you do not have a million-qubit machine today.”
For Quanfluence, photonics offered the least constrained route forward. “The qubits are already photons,” he said. “So it’s easier to scale.”
Why Photonics Looks Different
A key advantage, according to Chakravarthy, lies in operating conditions. Unlike superconducting systems that require extensive cryogenic cooling, photonic machines can run largely at room temperature.
That distinction matters as data centres face growing energy demands. They also enhance connectivity. Today’s computing infrastructure relies on optical fibre to link machines. Quantum systems, however, cannot simply copy data between nodes.
In his view, photonics fits naturally into that future. “Today, when you think of how machines are connected in a data centre, it’s through photonic links,” Chakravarthy said. “So what you have to do is put some kind of transduction where you go from one machine to the other, and my imagination of that transduction is somehow photonics,” he added.
Quanfluence’s work is deeply hardware-led. The company builds core components, such as measurement systems and single-photon detectors, in-house, and most of them operate at room temperature.
These are used in communication, quantum random number generation and quantum key distribution. “Some of these parts today are commercially available from Quanfluence,” he said.
While these products generate revenue, they are not the end goal. “Our main business is that all of these go into that machine we are building,” Chakravarthy said.
The most challenging piece remains the resource state generator, which he described as having the lowest technological readiness. Resource state generators form the fundamental blocks of fusion-based quantum computing as they create highly entangled quantum states.
Bridging the Gap Before the Big Machine
Alongside its long-term quantum computer effort, Quanfluence has built a separate system aimed at today’s optimisation problems, which can be tackled without full quantum hardware.
Quanfluence developed what Chakravarthy calls a quantum-inspired machine based on the Ising model of quantum magnetism. “You can load an energy function onto that, and you get the same answer,” he said.
The system does not require cryogenic cooling and fits into a standard data centre rack. “We actually have customers who are using that,” Chakravarthy said. While it is not a quantum computer, he sees it as a practical bridge. “It fills the gap to a quantum computer.”
The speed gains can be significant for large problems. Chakravarthy cited vehicle routing as an example. “Some of these things [calculations] take 15–20 minutes,” he said. “A machine like this could do that in about 20 seconds.”
Accuracy is not absolute, but that trade-off is often acceptable. “You can’t really use something like this if you’re looking for a 100% accuracy solution,” he said. “If you get to 90%, it’s good. If you get to 95%, it’s great.”
What About the Chip?
Alongside system-level hardware, Quanfluence is building photonic chips that underpin its quantum architecture. Chakravarthy stressed that quantum hardware must ultimately move away from large optical tables.
“Finally, a big tabletop setup has to scale into a chip,” he said. The challenge is that photonic chip design does not yet benefit from the mature electronic design automation flows used in conventional semiconductors.
Much of today’s photonic chip work still starts from first-principles physics. Engineers design structures based on how light is expected to behave, simulate them using specialised physics tools, and only then move towards fabrication.
“Even the simulation tools are not the standard tools,” Chakravarthy said, noting that photonic simulations often rely on physics-based solvers rather than mainstream chip design software. This makes iteration slower and mistakes expensive.
Fabrication happens at external foundries, primarily outside India, with typical turnaround times of around six months per run. “A mistake can cost you a million dollars and one year of time,” he said.
To manage this risk, Quanfluence relies on extensive lab setups that allow the team to validate optical behaviour, measurement systems and integration workflows before committing designs to silicon.
While the intellectual property remains entirely in-house, the supply chain remains global, with critical components such as lasers, diodes and specialised fibres still sourced from the US and Europe.
For Chakravarthy, this chip work is not ancillary. It is central to converting quantum physics into engineering. Without reliable photonic chips, scaling qubits, integrating systems and reducing form factors remain theoretical goals rather than deployable outcomes.
The Road Ahead
Quanfluence does not follow a linear qubit roadmap.
“Getting my first qubit is a challenge,” Chakravarthy said. The company has publicly committed to reaching qubits by 2029, after which scaling would accelerate. “The year after that, about a 1,000-qubit machine.”
Chakravarthy’s roots lie in semiconductors, with a previous company acquired in 2018. Today, Quanfluence employs around 20 people, many of whom have a background in advanced physics.
He claims that government support has helped offset some of that risk. “We have a grant from the Department of Telecom (under the Digital Communication Innovation Square scheme),” he said, adding that the National Quantum Mission has been useful as well.
Yet, the biggest hurdle remains cultural rather than technical. “Believing that you cannot design greenfield technology out of India,” Chakravarthy pondered.
He does not see quantum as a replacement for AI or classical computing. “They are perfectly complementary technologies,” he said. Quantum systems, he argues, will tackle problems classical machines cannot.
When Quanfluence’s photonic quantum computer finally arrives, it will not resemble the dramatic machines often associated with the field. “It looks like a data centre rack. Nothing so exciting,” he joked.
The post This Bengaluru Startup is Betting on Light to Build India’s Quantum Future appeared first on Analytics India Magazine.