let’s talk.

redefine reality.

In 1964 the Soviet astronomer N. S. Kardashev, writing on how civilizations might transmit information across space, classified them in three types by the energy they harness: a planet’s, a star’s, a galaxy’s.

  • Type I, a planet. “close to the level presently attained on the earth”: Earth’s own level when he wrote, about 4 × 1019 ergs per second, or 4 × 1012 watts.
  • Type II, a star. “capable of harnessing the energy radiated by its own star”, for example with a “Dyson sphere”: about 4 × 1033 ergs per second, or 4 × 1026 watts.
  • Type III, a galaxy. “in possession of energy on the scale of its own galaxy”: about 4 × 1044 ergs per second, or 4 × 1037 watts.

Each step is a leap: from Type I to Type II the energy grows about a hundred trillion times, and from Type II to Type III about a hundred billion times.N. S. Kardashev, “Transmission of Information by Extraterrestrial Civilizations”, Soviet Astronomy 8, 217 (1964) (opens in a new tab)

We are working toward one complete system that integrates conventional computers, AI, machine learning and quantum technology.

Along the way, we are closing gaps in quantum technology by developing simulation software for superconducting components and error-correction technology.

We hold our own technology to the standard we apply to others’ claims. The system must be reliable, parallel, programmable, debuggable and predictable.

five components.

  1. 01see an example

    assess.

    Independent due diligence on quantum technologies and their claims.

  2. 02

    verify.

    Results verified under a stated model with the rules fixed in advance: the model sets the number.

  3. 03

    advise.

    Quantum readiness: use cases, timing and risk, encryption included.

  4. 04

    connect.

    Inventions from universities and laboratories, matched with capital and companies.

  5. 05

    inform.

    Briefings, education and ongoing monitoring of quantum technology for decision-makers.

$1.3T–$2.7T

at stake for industry from quantum computing by 2035, by an approximate estimate.

McKinsey & Company, Quantum Technology Monitor 2026 (opens in a new tab)

$400B–$600B

possible value of quantum computing use cases in finance by 2035, by an approximate estimate.

McKinsey & Company, Quantum Technology Monitor 2026 (opens in a new tab)

“Private investments increasingly hinge on credible paths to error correction and revenue from emerging use cases.”

McKinsey & Company, Quantum Technology Monitor 2026 (opens in a new tab)

let’s talk.

An objective, verifiable account of what quantum technology delivers.

send message

Finance: collateral optimization, fraud detection and derivative pricing, and the move to quantum-safe security.

Insurance: mortality forecasting (the model fit) and reinsurance allocation, shown at small scale on a quantum processor; risk assessment and pricing, proposed.

Quantitative investment: portfolio allocation and asset modeling, derivative pricing and hedging.

Finance and quantitative investment: McKinsey & Company, Quantum Technology Monitor 2026 (opens in a new tab). Insurance: H.-Y. Liu et al., Science China Information Sciences 68, 202501 (2025) (opens in a new tab).

how it fits together.

  1. 01

    classical first.

    Classical computers do most of the work; the quantum processor takes the parts of a problem where it adds value.

    McKinsey & Company, Quantum Technology Monitor 2026 (opens in a new tab)
  2. 02

    one system.

    CPUs, GPUs and quantum processors work side by side as co-processors, managed by one software layer.

    M. Mohseni et al., arXiv:2411.10406 (2026) (opens in a new tab); McKinsey & Company (opens in a new tab)
  3. 03

    AI runs quantum.

    Machine learning already helps quantum computers correct their errors: a neutral-atom processor reached below-threshold error correction with machine-learning decoding.

    D. Bluvstein et al., Nature 649, 39 (2026) (opens in a new tab)
  4. 04

    quantum for AI, later.

    In a hypothetical later stage, quantum computers take on AI tasks that are extremely difficult classically, such as model training.

    McKinsey & Company, Quantum Technology Monitor 2026 (opens in a new tab)
  5. 05

    the limit.

    For machine learning on classical data, moving the data in and out can cancel the quantum speed-up.

    J. Preskill, Quantum 2, 79 (2018) (opens in a new tab); M. Mohseni et al. (2026) (opens in a new tab)