Quantum's quiet capital wave: Oratomic's $300m Series A and Singapore's QAI cohort

HorizonRéservé aux abonnés 11/07/2026 à 11h418Ajouter aux favoris

Arch and Khosla just led a $300m Series A into quantum startup Oratomic - a serious sum for pre-utility hardware. In parallel, QAI Ventures picked its first four Singapore quantum cohort startups. The 2026-2028 quantum thesis isn't gate count. It's who owns the tooling layer.

In plain terms

Quantum computing is not moving toward a "quantum ChatGPT" moment - but real capital is stacking. Arch Venture Partners and Khosla Ventures co-led a $300m Series A into Oratomic (Tech in Asia), a startup working on a specific atomic-modality quantum system. On the same news day, QAI Ventures backed four startups in its first Singapore quantum accelerator cohort. Two ends of the funding stack - mega-round and accelerator - moving on the same theme.

What's happening

$300m for a Series A in quantum hardware is the largest single check in the space this cycle, and reflects a specific thesis: the tooling and orchestration layer will monetize before the raw qubit count wins. Oratomic's pitch is not "beat IBM to a million qubits" - it's build the compilers, error-correction pipelines, and hybrid classical-quantum runtime that anyone with a fault-tolerant machine will need in 2027-2029.

QAI Ventures' Singapore cohort matters as a geography signal. Singapore has decided quantum is a national industrial policy: EDB funding, tax exemptions on quantum R&D, and academic tie-ups with NUS. Four startups is a small cohort - but signals commitment, not experiment.

Under the hood

The quantum stack under this capital:

Physical layer: multiple modalities still in play - superconducting (IBM, Google), neutral-atom (QuEra, Atom Computing, Oratomic), photonic (PsiQuantum), trapped-ion (IonQ, Quantinuum). No consensus winner.
Error correction: 2026 was the year of surface-code demonstrations. 2027 is expected to be the year of logical qubits at scale.
Applications: chemistry simulation and cryptanalysis remain the real economic use cases. AI-adjacent claims (quantum-accelerated ML) remain speculative.

So what

For decision-makers: quantum is not a 2026 revenue line. But the runway to the first commercial fault-tolerant workload has shortened from "decade" to "3-5 years" in serious estimates. That is short enough to fund tooling now. For builders: quantum compilers and error-correction libraries will trade at pre-AI-boom multiples if this thesis holds. For skeptics: watch Oratomic's first published benchmark. If it lands, the $300m becomes a floor, not a ceiling.