On an IBM Quantum Heron processor, a 156-qubit chip that loses coherence in microseconds, a small team in San Francisco has been running anomaly detection on financial datasets and reporting cleaner pattern recognition than the bare hardware delivers on its own [SiliconANGLE, Jan 2026]. The team is Haiqu. Its bet is that the quantum computers shipping right now, noisy, error-prone, and expensive, can be coaxed into doing real work if you wrap them in the right software.
In January 2026, Haiqu closed an $11 million seed round to build out that wrapper, which it calls a hardware-aware operating system for quantum applications [SiliconANGLE, Jan 2026]. The investor list is unusually broad for a deeptech seed: Primary Venture Partners, Qudit Investments, Alumni Ventures, Collaborative Fund, Silicon Roundabout Ventures, Toyota Ventures, and MaC Venture Capital all wrote checks [SiliconANGLE, Jan 2026]. The pitch is straightforward. Quantum hardware vendors are racing to add qubits. Customers want answers today. Haiqu sells the layer in between.
The bet
Founded in 2022, Haiqu builds what it describes as careful software for clumsy quantum computers, a phrase the company uses on its own LinkedIn page [LinkedIn]. The product is a hardware-aware software stack that lets developers run larger circuits on existing machines than the raw hardware would otherwise tolerate [SiliconANGLE, Jan 2026]. In practice that means circuit compression, error mitigation, and pre- and post-processing routines tuned to the quirks of specific chips. The company has demonstrated the approach with quantum preprocessing on IBM's Heron processor, applied to financial pattern recognition workloads [SiliconANGLE, Jan 2026].
The wedge is the gap between marketing-deck quantum and shipping quantum. IBM, IonQ, Quantinuum, Rigetti, and others publish qubit counts that climb each year, but the useful circuit depth on those machines remains shallow because of noise. A software layer that stretches what current hardware can credibly do is, if it works, valuable to every customer that has already bought time on a quantum cloud and wants to justify the spend.
Why it could be big
The near-term quantum software market sits in an awkward, interesting place. Hardware vendors keep promising fault tolerance is coming. Enterprise R&D teams at banks, pharma companies, and logistics firms have budget lines for quantum pilots and limited patience for results. A vendor-neutral software layer that improves output quality on whatever chip the customer has access to is, on paper, the kind of pick-and-shovel position that has worked in past compute transitions.
The investor mix reflects that thesis. Toyota Ventures has a documented interest in quantum applications relevant to materials science and mobility. Collaborative Fund and MaC Venture Capital bring consumer and growth-stage pattern recognition. Primary Venture Partners and Silicon Roundabout Ventures lean technical. Qudit Investments is a quantum-specialist firm. The collective signal is that several distinct theses, enterprise pilots, sector-specific applications, and infrastructure plays, all route through the same software layer.
Seed round (Jan 2026) | 11 | $M
The team
Haiqu was co-founded by Richard Givhan, an engineer trained at Stanford who serves as chief executive, and Mykola Maksymenko, a quantum researcher whose career has spanned the Max Planck Society and the Weizmann Institute of Science and who serves as chief technology officer [TechFundingNews] [FundBat] [InVenture]. Maksymenko was previously head of R&D at SoftServe Inc, the Ukrainian-American IT services firm, a background that helps explain Haiqu's depth bench of researchers based partly in Ukraine [Venture Views].
That bench is not small for a seed-stage deeptech company. Haiqu reports more than 20 researchers on staff [Venture Views]. Team members surfaced publicly include Vladyslav L., who works on quantum machine learning problems [LinkedIn], Oleksa Hryniv, a junior research engineer [LinkedIn], Volodymyr Serhieiev [LinkedIn], and Maxence Grandadam, who was an award winner at the IBM Quantum Developer Conference 2025 [SiliconANGLE, Jan 2026]. For a company building compiler-adjacent infrastructure for a hardware platform that changes every six months, the density of working physicists and ML researchers on staff is the asset that matters most.
What the bears say
The most credible concern about Haiqu is the one that hangs over the entire near-term quantum software category: if fault-tolerant quantum hardware arrives on the timeline that IBM and others now project, the value of error-mitigation and circuit-compression layers shrinks because the hardware itself stops being clumsy. Competitors like 1QBit have lived inside that uncertainty for years. The bull answer, and the one the seed investors are clearly underwriting, is that the bridge period is not a quarter or two. It is most of this decade. Enterprise customers running pilots on Heron-class chips today need useful output today, and the software layer that delivers it accumulates customer relationships, benchmark data, and integration depth that carries forward into whatever hardware generation comes next [SiliconANGLE, Jan 2026].
There is also the question of distribution. Haiqu sells into a small number of sophisticated buyers, most of them already in direct contact with hardware vendors. The company has not disclosed named enterprise customers. The investor syndicate suggests at least some warm introductions through Toyota Ventures and the broader fund network, but converting pilot users into paying API customers is the work of the next 18 months.
What to watch
Three things will tell the story over the next year. First, named customer disclosures: a single Fortune 500 logo running Haiqu in production on a quantum cloud would meaningfully change the conversation. Second, published benchmarks: the IBM Heron financial workload result is suggestive, and a steady cadence of peer-reviewed or vendor-co-signed performance numbers across more chip architectures (IonQ, Quantinuum, Rigetti) would establish the hardware-neutral claim. Third, the next round. Seed deeptech rounds of this size and syndicate composition typically expect a Series A inside 18 to 24 months, which would put Haiqu back in the market sometime in 2027.
The quantum software category has been waiting for a company that treats today's noisy hardware as a customer problem rather than a temporary embarrassment. Is Haiqu the team that turns that posture into recurring revenue, or does the bridge period close faster than the bridge-builders expect?