ICM HPQC News Flash - August 2026
- 12 hours ago
- 16 min read
By Oliver Campbell │ ICM HPQC Fund │ August 2026
“Written by humans, please don’t blame the robots for our typos”
Five by Five – “Loud and Clear”
HPQC Fund 1 has closed: Find attached our 12 month summary letter to investors as well as an up-to-date performance data, and a summary of the year for our portfolio holdings alongside fund and industry commentary.
HPQC Fund 2 launches: Thesis remains intact with strong return proof-points. Very exciting time to launch with lots of opportunities emerging. Find links to the non-NDA VDR room, have a look around, get in touch, ask questions, say hello.
Teach-ins: We will be hosting teach-ins for our new Economics of Quantum model as well as our differentiated view on Compute in Space; a critically nuanced
spaceahem sector. Please reach out if you are keen to learn more.ESG – Excel, Surveys . . . Gut Feel?: HPQC Principal Oliver Campbell argues that ESG frameworks which focus on what can easily be measured while staying silent on what cannot will systematically overweight energy and emissions relative to harder and possibly more consequential issues, leading to unforeseeable investment outcomes. Read the full article here.
Around the traps: Stay up to date with industry and company info. This month’s industry news flow reflects the Gordian Knot of finance and physics and how two interrelated sides of the industry see the same problem differently.
HPQC Fund 1 has closed: Letter to Investors
On June 30th the fund turned one year old, and is now closed to further investment. It has been a busy twelve months, and we thank you for your support along the way. HPQC Fund 1 stands almost fully invested across five companies spanning the compute stack. The past year has brought new investments into Diraq (spins in silicon, quantum computing), Mixx Technologies (co-packaged optics and silicon photonics) and, most recently, neuromorphic chipset supplier SpiNNcloud. These sit alongside existing investments in Salience Labs (ultra-fast all-optical circuit switches using silicon photonics) and Q-CTRL (AI-powered infrastructure software for quantum computing and quantum sensing). We have kept to our plan of allocating swiftly to the high-quality projects we had been tracking for some time.
Thank you to all of you who have hit us up with questions and shared insights over the last 12 months. Keep them coming! When we started the fund we had ambitions that we were building not just a great returns engine, but a community of fellow tech junkies who would help us improve and thrive on every axis. That ambition is being realised, and we are grateful for your part in it.
Below you will find an overview of the fund across the last 12 months, and a bit of prognostication as to where we think we are headed next. You should be receiving a bit of correspondence from us across the next few weeks – please keep your eyes out. If you feel like you have missed any of these let us know:
Formal fund administration reports for your records
Invitations to lunch and breakfast events in September
Information about HPQC Fund-2
To manage both the increased demand for our insights and products in the ASEAN region we have hired Oliver Campbell who is based at our HQ in Singapore. And to lead our global project origination efforts we have taken on Charlie Kidd. Their bios follow below. Other new team members include new directors on our tech advisory board – check out our website for their impressive bios.
If you would like more details on Fund 1, or how to get into Fund 2; or if you’d like to tell your friends all about us, or even just book some time for a chat, I would welcome your email.
With respect,
Matt Gould
Portfolio Manager HPQC
Meet HPQC at SEMICON Taiwan
Charlie and Matt will be at SEMICON Taiwan next week. If you’re attending and would like to connect, get in touch with Charlie to arrange a chat.
The Thesis, Restated
The past 60 years of compute scaling — more transistors, a falling cost per transistor, ever-higher manufacturing yields — are running up against physical limits. Simply pouring more capacity (capex) into infrastructure with structural bottlenecks does not relieve the pressure; it exposes it. The next gains in compute must therefore come from elsewhere in the stack: packaging, interconnect, photonics, quantum hardware and control software, and data-centre power and cooling, among other layers. This is the ‘Compute Gap’, and it is where HPQC invests.
HPQC Today - Forward View
The portfolio is active, and demand for its companies’ products — the natural consequence of the Compute Gap our models describe — is building at pace. This is, in turn, driving portfolio companies to raise further capital to meet that demand. As of August 2026, three companies have new rounds expected to close in the near term:
Salience Labs: a Series B targeting completion by the end of Q3 CY2026, at roughly 3x the valuation at which HPQC led the Series A.
Q-CTRL: a capital raise set to close with term sheets already in hand, at a USD1.025bn valuation — putting the company on a path to become the fund’s first unicorn.
Diraq: our standout quantum holding has indicated at board level its intention to pursue a Series B over the next nine months, following the U.S. Government’s own investment in the business.
These valuation uplifts across the portfolio are set out in full in the NAV (net asset value) Summary for July 2026, and in the accompanying Guidance Summary, which incorporates the upcoming capital raises in a forward view to December 31st 2026.
We have two new team members:
Charlie Kidd
Head of Origination, ICM HPQC Fund
Charlie Kidd joined ICM in 2026 as Head of Origination for the ICM HPQC Fund, covering Europe and Asia, and is based in London. He is responsible for building and managing the fund's early-stage deal pipeline across the compute infrastructure stack, sourcing founders in silicon, photonics, quantum and adjacent hardware before they reach the broader market. Prior to joining ICM, Charlie spent four years as a Senior Strategy Advisor at the UK Department of Health and Social Care, where he led high-profile work on the COVID-19 pandemic response and adult social care reform. Charlie holds a Master's in Public Policy from Harvard University and a Bachelor of Science in Natural Sciences from Durham University.
Oliver Campbell
Investment Principal, ICM HPQC Fund
Oliver Campbell joined ICM in 2026 as an Investment Principal for ICM’s HPQC fund and is based in Singapore. He previously spent twelve years in Singapore as a portfolio manager and investment analyst on Stewart Investors’ Worldwide, Global Emerging Markets and Asian equity strategies. Earlier in his career, Oliver spent eight years as an equity analyst covering the tech and telco sectors in Taiwan, Hong Kong and Singapore after starting his career as an industrial trade journalist across global emerging markets from Pakistan to Turkey, Thailand, South Africa and others. Oliver holds a BA (Hons) in Chinese and Management from the University of Durham (UK), a Diploma from Renmin University (Beijing, PRC) and an MSc in Finance (Economic Policy) from the School of Oriental and African Studies (SOAS), University of London.
HPQC Companies
By any measure, this has been a vintage year for operational proof points. Diraq’s foundry-scale fabrication demonstration, Q-CTRL’s claim to practical quantum advantage, Salience’s move into pre-production manufacturing, Mixx’s acquisition-led push into Indian manufacturing, and SpiNNcloud’s largest-ever deployment all landed within a five-month window — a clustering of milestones that speaks as much to the sector’s maturation as to any single company’s progress.
The past year has been an exercise in de-risking manufacturability rather than chasing research headlines. In February 2026, Australia’s National Reconstruction Fund Corporation made a strategic AUD20m investment explicitly designed to anchor quantum manufacturing onshore, rather than simply fund research, taking Diraq’s total venture funding past US$100 million.
Three months later came a Letter of Intent with the U.S. Department of Commerce for up to USD38m in proposed CHIPS Act funding to support domestic scaling of fault-tolerant silicon processors, alongside a Palo Alto office expansion that doubled the local team.
Most significant of all: in July 2026, Diraq and imec fabricated and demonstrated an eight-qubit linear array on a standard 300mm CMOS foundry line, published in Nature Communications. This is not incremental — it is direct evidence that the “billions of qubits on existing semiconductor infrastructure” thesis holds at foundry scale, not merely in the laboratory.
Q-CTRL’s year has been defined by the shift from technical promise to commercial proof. In May 2026 the company reported a 3,000x simulation speed-up on a commercially relevant materials-science problem on the IBM Quantum Platform, cutting a Fermi-Hubbard model simulation from over 100 hours to two minutes — what it describes, not unreasonably, as the first evidence of practical quantum advantage.
That claim has been reinforced by widening hardware integrations, including autonomous calibration software for rack-mounted quantum systems and native integration of its Fire Opal software with IonQ’s hardware. The positioning that emerges is of a software layer on which multiple hardware vendors depend, rather than a wager on any single platform.
Salience has moved from demonstration to manufacturing readiness this year. In February 2026, it partnered with Tower Semiconductor to shift its photonic-integrated-circuit optical switches from development to pre-production for at-scale AI data centre deployment, followed in March by the launch of its 32-port all-optical switch — the first in a planned family that scales to 64 and 128 ports to meet growing AI data centre capacity needs.
A parallel collaboration with Keysight Technologies to build the industry’s first dedicated optical circuit switch testing environment lends credibility via a recognised test-and-measurement partner.
The company has grown past 50 staff across the US, Taiwan and India, with its first-generation HBxIO (a silicon-integrated optical engine and co-packaged optics (CPO) platform) engine delivering 25.6 terabits per second of bidirectional bandwidth.
Most recently, in July 2026, Mixx acquired Bangalore-based Sophic Silicon Technologies for analog mixed-signal design IP integral to its HBxIO platform and struck a manufacturing collaboration with Kaynes Semicon to build a dedicated test-and-assembly facility in India — the first instance of an international deep-tech company anchoring both the design and manufacturing layers of a next-generation AI silicon platform domestically in India.
Management’s own view is that AI interconnect remains sufficiently nascent that no single technology will dominate, even as recent M&A activity drives up the valuations of Mixx peers. We anticipate Mixx announcing a Series B raise before year end 2026.
SpiNNcloud’s year has centred on proving deployability at scale. Its SpiNNaker2 architecture moved from prototype to a commercially available supercomputing platform, with the full-scale system capable of simulating up to 5 billion neurons in real time across 5 million cores.
The flagship proof point is a deployment to Leipzig University comprising 656,640 cores across roughly 4,320 chips, simulating a minimum of 10.5 billion neurons to support drug-discovery research — its largest system deployed to date.
Commercial traction was tracking in the double-digit millions of euros through 2025, including engagements with US national laboratories and European supercomputing centres. We will be providing more information as we close the transaction with them this month.
The Past Year for the Industry
It has been a red-letter year for investment across the compute stack. The listed space has, once again, dominated the mainstream headlines, but the more instructive story of the past twelve months is the industry’s growing preoccupation with the binding constraints imposed by physics and infrastructure. These fall into several themes, each with different implications for the portfolio.
Power, not chipset availability, is now the binding constraint
Every layer above it — nuclear deals, grid queues, even decisions on memory allocation — traces back to a single question: can sufficient power get to the building? This carries two implications for HPQC. The first is that our existing holdings sit on both sides of the demand equation: photonics and fabrics, of the kind supplied by Mixx Technologies and Salience Labs, and neuromorphic chipsets, of the kind supplied by SpiNNcloud, all raise compute capacity while lowering power draw per unit of output. The second is that HPQC is watching the supply side closely, with small modular reactors and high-voltage direct-current standards for data centres among the opportunities currently under review.
Photonics is where the next moat will be built, or broken
Whoever wins the coming standardisation fight will command real pricing power over the next cluster generation — precisely the ground on which HPQC is positioned via Mixx Technologies and Salience Labs.
Quantum’s narrative has matured: from “someday” to “which decade” to “which hybrid architecture”
At Computex in May and June 2026, the liveliest conversations among server suppliers concerned quantum’s integration into the data centre. The bridge between GPU and QPU — Nvidia’s NVQLink (an open, high-speed interconnect architecture that connects quantum processors to classical AI supercomputers for real-time error correction) among the more credible efforts — now looks a more plausible near-term story than standalone claims of quantum advantage. M&A (mergers and acquisitions) activity is, meanwhile, concentrating attention on the modality question, and increasingly on the spins-in-silicon, lowest-qubit-unit-cost thesis that underpins our investment in Diraq. That IBM has decided to supplement its Superconducting qubit team with a Silicon Spin team is a huge endorsement of the modality we have been championing for almost five years now (not to say we told you so!)
Memory has flipped from footnote to headline constraint
A year ago the story was GPU (graphical processing unit) allocation; today it is HBM (high bandwidth memory) allocation, with the knock-on effect of data centre demand now visible in consumer markets, in the form of shortages and higher prices. The performance of listed memory names tells much of this story already; we continue to watch memory and packaging technology closely.
Outlook
There is no shortage of commentary questioning the sustainability of hyperscaler capital expenditure, and equally little sign of it slowing. There is, at the same time, a growing conversation about sovereign compute — France, Japan, the UAE, India and Canada among the governments seeking to build domestic capacity in pursuit of data residency, geopolitical resilience and AI independence.
Notwithstanding the macro picture, HPQC’s focus remains bottom-up: companies that help alleviate the physical bottlenecks in compute. We believe that investing in high-quality businesses which take cost out of compute offers a longer runway for growth, with lower revenue volatility, than exposure to headline capex figures. This emphasis on low unit cost and high throughput reflects our background in infrastructure investing and is a theme we will pursue with equal discipline in HPQC Fund 2.
Risks:
Being right about the bottleneck doesn't guarantee that any one company captures the value, and questions about standardisation remain open.
None of the four funding rounds noted above are fully complete at the time of writing. We've described them as expected to close before year-end, but a lead pulling back or terms shifting in diligence could push timing into 2027 or alter pricing.
Particularly relevant to Diraq, but not exclusively so, is that the modality thesis could move again. There have been strides in other quantum modalities over the past 12 months that challenge spins-in-silicon cost advantage; even recently, however, acquisitions in the spins-in-silicon space have brought the industry's attention back to the modality, with IBM investing strongly to build its own spins-in-silicon team – a move which should reflect positively on Diraq’s strong trajectory. The HPQC Fund, along with IBM now, retains high conviction that spins-in-silicon will be the major commercial modality for quantum computing at scale.
While we are at least one step removed from capex risk, we are not fully immune to it — a hyperscaler slowdown could hit us through delayed customer decisions and harder fundraising for our companies, even without direct capex exposure.
From the fund perspective, in just one year we have had zero exits and zero DPI (distributions to paid-in capital). This is normal for the stage, and every valuation uplift referenced above remains unrealised on a cash basis.
2. HPQC Fund 2 Launch:
The compute bottleneck thesis that anchored HPQC Fund 1 is not just intact, it's intensifying. The constraints we identified early (power, interconnect, qubit and photonic economics) remain the binding variables on the AI compute buildout, and the market is only now starting to catch on. HPQC Fund 1 is the proof: portfolio companies built on unit-cost fundamentals rather than TAM narratives are converting technological differentiation into commercial traction, and the returns are starting to show it.
What's changed is the opportunity set, not the thesis: the infrastructure layer beneath compute is deeper and more investable today than it was when we started, with more credible teams solving harder sub-constraints at better entry points. Fund 2 exists because the thesis worked and the window to back the next generation of picks-and-shovels compute infrastructure is open now, and we, alongside our Technical Advisory Committee, Chaired by Dr Bill Jeffrey, will continue to fish precisely with “rods not nets” by identifying specific winners in the stack, cross-referencing them with industry and academia and backing them with conviction.
Links to Fund 2 VDR here.
Please get in touch with any questions - we would be more than happy to discuss the Fund, the industry or companies with you.
3. Teach-ins: Economics of Quantum & Compute in Space
We will be hosting teach-ins for our new and updated Economics of Quantum model as well as our differentiated view on Compute in Space. Get in touch if you are keen to learn more. In the meantime, here is some background reading on this rapidly evolving . . . space, ahem sorry, sector:
4. ESG – Beyond the Scorecard
By Oliver Campbell │ ICM HPQC Fund │ August 2026
ICYMI, Oliver Campbell’s recent perspective on ESG is worth a read. Rather than treating ESG as a box-ticking exercise, Oliver explores why conventional scoring can miss what matters most when investing in early-stage compute and quantum technology – from governance and dual-use technologies to Scope 4 emissions, Jevons Paradox and the pace of technological change.
Read the full article here: ESG – Excel, Surveys, and . . . Gutfeel?
Around the Traps: The Latest News from The Sector
In compute, finance and physics are tied in a gordian knot and yet financial analysts and engineers rarely seem to converge despite describing the same buildout. Industry news flow this month shows how analysts are worried about how AI capex is financed: debt, leases, deteriorating cash flow, all the while treating compute as one undifferentiated input. Whereas engineers worry about whether the hardware can actually be used: interconnect, power, utilisation below 1% of peak, while treating the capital as a given. Neither one prices the other's variable, even though the physical bottleneck is a major financial one (and vice versa). HPQC sits at that junction: rather than a TAM-sized bet on “AI compute” or a pure engineering view with no capital lens, the fund disaggregates the buildout into specific physical constraints and lets unit-cost economics and the Technical Advisory Committee judge which ones are durable and still mispriced.
Below are some of the articles that caught our eye this past month:
The Compute & Capex Debate
“AI Doesn't Need to Be a Bubble to Break Things” — Full Stack Capitalist
Sidesteps the bubble/no-bubble argument entirely. Its claim: the AI buildout is a genuine economic phenomenon rather than a policy failure, because it's crowding out capital and physical inputs toward the highest-return sector in a supply-constrained system; everyone else absorbs the opportunity cost whether they had a vote in it or not. Cites Goldman's projection that global AI investment crosses $1 trillion in 2026, reaching 2.8% of US GDP by 2028. Practical read: treat elevated input costs and equipment lead times as structural for the next two to three years, independent of whether AI itself pays off. fullstackcapitalist.co
“The AI Capex Bubble: Bigger Than You Think, Funded by You Don't Know Who” — Pine Tree Macro Research
The financing-mechanics piece. Documents hyperscaler capex roughly tripling 2024–2026, forcing Alphabet and Meta to halt buybacks. Sharpest data point: Alphabet went free-cash-flow negative in Q2 2026 for the first time, with long-term debt more than doubling to $98bn in H1; Amazon's long-term debt jumped 81% to $119bn in Q1 alone. Argues off-balance-sheet lease commitments understate real exposure, and the equity-to-debt financing shift is accelerating against deteriorating cash flow, not in spite of it. Subsequent reporting puts a number on that understatement: Alphabet, Microsoft, Amazon, Meta, and Oracle are estimated to be carrying roughly $1.65T in off-balance-sheet debt combined, with the comparison to Enron's accounting now made explicitly rather than implied. pinetreemacroresearch.substack.com · techradar.com
Quantum Commercialisation Check
QED-C / Center for Quantum Networks — Quantum Networking Roadmap
Identifies critical gaps and technologies needed for widespread commercialisation, detailing 10 high-impact use cases of which only two are currently supported. Short-distance applications reach commercial maturity within 5 years; long-distance distributed quantum computing is roughly a decade away. Aimed at funders and technologists rather than investors — useful as a calibration point against more aggressive commercialisation claims. quantumcomputingreport.com
Q-CTRL, “Making Quantum Useful”
Written by Q-CTRL's own product leadership. Core claim: commercial quantum computing is entering a phase where progress is measured by customer value delivered rather than purely technical milestones, building on an earlier argument that a narrow focus on the quantum stack itself has been holding commercialisation back. q-ctrl.com
The Physical Layer
“8 Trends Shaping the Future of HPC in 2026” — Bull / ISC26 recap
Drawn from ISC High Performance 2026. Central data point: Jack Dongarra's closing keynote observation that the world's top supercomputers routinely achieve less than 1% of theoretical peak performance in practice, because processors spend most of their time waiting for data to move between memory, storage, and compute. Argues the field is shifting benchmarks away from peak FLOPS toward time-to-solution, energy per trusted result, and reproducibility — essentially HPQC's bottleneck thesis, arrived at independently by the HPC engineering community. The demand side looks no more solid than the supply side: Wood Mackenzie estimates over two-thirds of requested US data centre electricity for AI will never actually materialise, phantom projects inflating the capacity numbers everyone is underwriting against. bull.com · bloomberg.com
Tom's Hardware — the US-China photonics fight
The FCC is drafting a measure to block imports of new Chinese optical transceiver models, aiming to publish the rule before year-end 2026 — complicated by the fact that the US remains China's biggest photonics customer despite China's current dominance of the supply chain. tomshardware.com
Portfolio Newsflow
Diraq
On Aug 14, Diraq published a blog post arguing the quantum industry's hiring model — historically built around doctoral-level candidates — can't support future workforce scaling, pointing to undergraduate engineering, technical training, and roles like cryogenic engineering as alternative pathways. In mid-July, Diraq also announced publication of research on eight-qubit operation of a 300mm SiMOS foundry-fabricated device, alongside a U.S. expansion announcement.
Q-CTRL
In the week ending Aug 15, Q-CTRL broke the previous experimental ceiling for the Quantum Fourier Transform, running the algorithm across 100 qubits on IBM Heron hardware using a convolutional compilation strategy. Process fidelity hit 11.4% at 50 qubits, with the target frequency remaining distinguishable all the way to 100 qubits — written up alongside an accompanying arXiv paper.
Mixx Technologies
On Jul 23, Mixx announced the acquisition of Sophic Silicon Technologies, a Bengaluru-based semiconductor IP design firm, alongside a manufacturing collaboration with Kaynes Semicon to build a test-and-assembly facility supporting Mixx's optical IC packaging roadmap — positioned as the first time an international deep-tech company has anchored both design and manufacturing layers of India's semiconductor build-out.
Salience Labs
The Next Platform ran a feature on Salience Labs' silicon photonics optical switch on Jul 22 — trade-press analysis rather than a company announcement, but the only item genuinely inside the window.
Meet HPQC at SEMICON Taiwan
Charlie and Matt will be at SEMICON Taiwan next week. If you’re attending and would like to connect, get in touch with Charlie to arrange a chat.
Matthew Gould Oliver Campbell
Portfolio Manager of the ICM HPQC Fund Investment Principal
MAS Licensed Representative MAS Licensed Representative
ICM Global Funds Pte Ltd ICM Global Funds
Important Note:
The information in this article should not be considered an offer or solicitation to deal in the ICM HPQC Fund (Registration number T22VC0112B SF003) or ICM HPQC Fund 2 (Registration number T22VC0112B SF007) (the “Sub-funds”). The information is provided on a general basis for informational purposes only and is not to be relied upon as investment, legal, tax, or other advice. It does not take into account the investment objectives, financial situation, or particular needs of any specific investor. Investors should seek relevant professional advice before making any investment decision. The information presented has been obtained from sources believed to be reliable, but no representation or warranty is given or may be implied that it is accurate or complete. The Investment Manager reserves the right to amend the information contained herein at any time, without notice. Investments in the Sub-funds are subject to investment risks, including the possible loss of the principal amount invested. All forms of investments carry risks, including the risk of losing all of the invested amount. Investors should read the prospectus before deciding whether to acquire the units in the Sub-funds. The value of investments and the income derived therefrom may fall or rise. Past performance is not indicative of future performance. This document is intended solely for institutional investors and accredited investors as defined under the Securities and Futures Act 2001 of Singapore. The whole or any part of this work may not be reproduced, copied or transmitted or any of its contents disclosed to third parties without ICM Global Fund’s express written consent. This advertisement or publication has not been reviewed by the Monetary Authority of Singapore.
ICM HPQC Fund and ICM HPQC Fund 2 are registered Sub-Funds of the ICMGF VCC (the VCC), a variable capital company incorporated in the Republic of Singapore. The assets and liabilities of ICM HPQC Fund and ICM HPQC Fund 2 are segregated from other Sub-Funds of the VCC, in accordance with Section 29 of the VCC Act.
Get to know us better:
Comments