Theses›Quantum›Quantum memory

Quantum memory

Quantum memory becomes a fundable standalone product only as a telecom-native rare-earth/erbium module; otherwise value pools above it in the network/QPU layer (falsified if a standalone module raises $40M+ by end-2027)

  • Bear
  • Closing
  • $196M by 2035

The callEvery quantum repeater needs a memory, but IonQ and Cisco are building or buying the layer before its market arrives, so only a physics-scarce memory survives standalone.

D1D2R1R2C1foragainst D1D2R1R2C1foragainst
Fig. 1The balance. Drivers and catalysts against restraints, each block sized by its weight; catalysts hang dashed because they have not happened yet, sized by their odds. It tips toward the restraints.
  1. D1A quantum signal cannot be amplified, so every repeater must hold an entangled state until the next link is ready, and today's memories are bench-scale and at the wrong wavelength.
  2. D2Quantum computers start scaling by being networked, and networks start linking cities, which multiplies the repeaters that need a memory.
  3. R1IonQ bought Lightsynq for about $306.8M in stock and Cisco built its own room-temperature telecom entanglement chip; the buyers are making the part themselves.
  4. R2Demand is government and testbed money only; the first standalone memory sale went to a national research institute.
  5. C1The first multi-node repeater network orders 20 or more commercial memory units in one buy, making the module a merchant part.end-2028
Rests onthe beliefs underneath
Supported by · confidence
Threatened by · confidence
What would prove itdated tests, open and resolved
202720282029today1234predictionswatch 202720282029today1234predictionswatch
Fig. 2Each ring is a dated test: dashed while open, struck through when overdue, ticked or crossed once resolved.
  1. 1Value in quantum memory stays concentrated at the network-fabric / QPU-system layer, not the standalone memory module: no pure-play quantum-memory-module company announces a single funding round of $40M or more before 31 Dec 2027.Dec 2027 · open
  2. 2No company selling quantum memory as its main product raises a single round of $40M or more; one that does reopens the window.Dec 2027 · open
  3. 3A multi-node repeater network places a 20-plus unit memory order; DOE Q-NEXT, DARPA QuANET and the EU Quantum Internet Alliance are the buyers to read.Dec 2028 · open
  4. 4memQ or Qunnect shows recurring revenue that is not grants.Dec 2028 · open
Expressionslisted, private, technologies
  • XNDUXanadu Quantum Technologies Limited
    • speculative
  • IONQIonQ
    • speculative
     Owns Lightsynq's quantum memory and photonic interconnect technology and uses it inside its own trapped-ion computers and networking stack; it does not sell memories as a product.
  • CSCOCisco Systems
    • not screened
     Builds a prototype room-temperature, telecom-wavelength entanglement chip for quantum networks and invests in Qunnect; no storage memory product on sale.
Private · 4 private companies, 4 of them in the US; none in continental Europe.

Qunnect · memQ · Photon Queue · Lightsynq Technologies

3 more private companies tracked, not shown

Technologies
Who has already bet on itprivate rounds, from public sources

$55M raised by 4 companies, $42M of it in the last three years. Wave building: Money is arriving: recent rounds and tier-1 funds are in, but no runaway leader yet.

Backed by a top-tier venture firm: Photon Queue. Backed by a strategic investor: Qunnect, Lightsynq Technologies.

  • $23MQunnect
    • Founded 2017
     Series A extension, Jun 2025 $10M (Airbus Ventures, Cisco Investments, Quantonation) · Series A, 2022 $8M (Airbus Ventures, Quantonation)
  • $18MLightsynq Technologies
    • Founded 2022
     Series A, Nov 2024 $18M (Cerberus Ventures, Murata Electronics North America, IAG Capital Partners)
  • $10MmemQ
    • Founded 2021
     Series A, Mar 2026 $10M (Quantonation, Ocean Azul Partners, Yaletown Partners)
  • $4MPhoton Queue
    • Founded 2024
     Seed, Jul 2026 $4M (Playground Global) · State grant $1M (New Mexico Economic Development Department, Roadrunner Venture Studios)
What changedthe record, newest first
  • 23 Jun 2026Quantum memory is the branch of optical memory you cannot route around with electronics (quantum states do not survive conversion), which is exactly the structural moat classical optical memory lacks. That is what makes it the interesting merge.
  • 23 Jun 2026The strongest 2026 evidence in quantum memory is the integrators moving onto the layer themselves: a QPU maker bought a memory startup outright, and a networking incumbent built its own room-temperature telecom entanglement chip.
  • 22 Jun 2026The optical memory that is actually a funded product today is quantum memory for the quantum internet, and the only way to hold value is a telecom-native module physics keeps scarce.
2 earlier
  • 18 Jun 2026The version of "optical memory" that is actually a funded product today is quantum memory for the quantum internet, not the classical optical RAM people keep trying to build for switching.
  • 18 Jun 2026In quantum memory the big money and the strategic buyers are pooling at the network-fabric and quantum-computer layers, where the memory is just a component, while the companies selling the memory itself as a product still have only first government and testbed sales. The hardest component to build is not yet the one that captures the value.
Reading
Show the reading