On 2 October 2026, inside the wood-panelled grandeur of Mansion House, the Lady Mayor’s Defence & Resilience Showcase opened with an unusual ask. Everyone who walked through the doors was given a pot of virtual currency and invited to spend it on the technology companies they believed showed the greatest promise. It was a deliberately interactive way of making a point: in the market for security and resilience, attention and capital flow towards whatever feels most credible in the moment. In a room full of defence hardware, cyber capabilities, and established industrial names, a counter-disinformation company called Refute could easily have been dismissed as an outlier. It was not selling a platform, a vehicle, or a weapon. It was selling the ability to defend the information environment. By the end of the event, Refute had been named the best dual-use company of the showcase and had earned a place in Sky News’ Defence Insider coverage. Dual-use is the right word. Refute sits somewhere between civilian communications and national security, between the boardroom and the front line. The showcase itself was the first in its series to focus specifically on defence and national resilience, and it brought together technology companies, investors, government customers, and established defence organisations at a moment of acute pressure. War in Europe, daily cyber threats, and rapidly changing methods of conflict were not abstract context. They were the reason the room had gathered. That a company focused on information manipulation could emerge as the standout tells you something important about how the idea of defence is evolving. It is no longer solely about protecting physical territory. It is also about protecting the information environment and the trust that underpins it.
Less than two weeks later, the UK Prime Minister announced the creation of the National Centre for Information Defence, a body with a deliberately practical remit: to detect, attribute, and disrupt hostile-state information attacks. The accompanying warning was stark. Artificial intelligence, the Prime Minister said, would only increase the scale of disinformation campaigns. It was not a hypothetical concern. Russian agencies have already used bots, fake websites, falsified newspaper articles, and forged branding from 28 British organisations to spread false narratives and sow division. Days after that, the Government announced that CEOs and industry representatives from defence and critical national infrastructure would receive closed-door briefings on the threats posed by Russia. Disinformation appeared on the agenda alongside cyber attacks and sabotage, described as one of the tactics used to disrupt organisations, take activity offline, and weaken trust. There is a crucial distinction between a cyber attack and a disinformation campaign, even though they are often grouped together. A cyber attack compromises the systems an organisation relies on. A disinformation campaign compromises the confidence people place in the information distributed by those systems. It can make a genuine warning look questionable, manufacture the appearance of a crisis, or imitate a trusted source. And this is happening at a moment when trust in online content is already dangerously low. A recent Ofcom study on news consumption and distribution found that 56% of UK adults are less trustful of an AI-written news story than one written by a person. In this environment, the question is not whether an organisation will face an attempt to manipulate information, but whether it will see it in time and know what to do.
Communications teams may see the public narrative. Security teams may spot suspicious behaviour. Government may hold intelligence about the threat behind it. But leaders still have to decide whether and how to respond. Earlier this year, we argued that counter-disinformation was beginning to look like cybersecurity did a decade ago, and the announcements since then have backed that claim. Cybersecurity became a board-level concern when companies understood that an attack could stop the business. Information defence is following a similar path because trust affects how an organisation communicates, operates, and makes decisions under pressure. Refute’s platform is built for exactly this problem. It surfaces early signs of online manipulation, traces narratives as they develop, and provides intelligence and recommended responses while a campaign is still emerging. The aim is to help governments and commercial organisations act at the speed of the attackers, rather than after the false narrative has already become the crisis itself. At Mansion House, trust was represented by virtual currency and allocated with a click. In the real world, it takes years to earn, minutes to weaken, and far longer to rebuild. That is why protecting it has become a job for government, business, and technology working together. The rest of this month’s news from our portfolio and partners, across Intelligence, Human, and Planet, is best read through that lens: how do we make the systems we depend on faster, stronger, and more trustworthy.
Across the Intelligence portfolio, the theme was making complex systems faster, safer, and more transparent. Altana partnered with Unusual Machines to improve supplier verification, traceability, and compliance across drone-component manufacturing, an increasingly urgent issue as the US works to secure its drone supply chains. iPronics raised $125 million for AI networking in a Series B co-led by Maverick Silicon and Light Street Capital, with NVIDIA joining the round and Amadeus participating; the company also began a two-year collaboration with the Barcelona Supercomputing Center to integrate its programmable optical switching platform into GPU and high-performance computing infrastructure. Photonic unveiled Project VANGUARD, a proposed C$500 million semiconductor facility in British Columbia aimed at quantum, AI, and defence technologies, and expanded its work with Microsoft on quantum-resource estimates so developers can understand the hardware needed for useful applications. Pimloc extended its privacy platform beyond video, audio, and images to include PDFs, emails, scanned records, and Microsoft Office files, and was selected by Surrey Police and Sussex Police under a joint multi-year contract to redact sensitive video and audio. In quantum, planqc delivered the first MAQCS hardware to the Leibniz Supercomputing Centre, an early step towards installing a 1,000-qubit system, received a Made in Europe Award, and had its LOGIQC proposal selected in Germany’s national quantum competition. PolyAI introduced Wren, an agent that analyses customer conversations, identifies weak points, and tests proposed improvements before deployment, essentially an AI that helps other AI perform better. Riverlane expanded its Deltakit compile and visualisation packages, demonstrated a closed, real-time error-correction loop with Qblox, opened a US headquarters in Maryland’s Discovery District, and signed an agreement with Infleqtion to combine its Deltaflow error-correction stack with neutral-atom hardware and Superstaq software. Sitehop expanded into the US with an office in Arlington, Virginia, focused on defence, national security, and critical infrastructure. V7 announced an integration with Affinity for investment teams and published new customer stories showing AI at work in investment-document workflows and OpenAI’s visual-data operations. XMOS brought voice AI to television by joining RedSquid TV’s agentic television ecosystem, adding edge voice processing to a platform designed for natural spoken interaction. It is a wide-ranging list, but the through-line is simple: the infrastructure of the future has to be both powerful and legible, capable of processing enormous amounts of information without obscuring the signals that matter.
On the Human side, the news was about earlier intervention, better manufacturing, and faster decisions. iPremom raised €15 million in a seed round led by Amadeus to support clinical validation and commercial development of its MaiRa blood-test platform, which aims to predict pregnancy risks in the first trimester. That is a powerful example of how a diagnostic tool can change the timeline of care, giving clinicians and families information earlier and, ideally, better options. Ori Biotech signed a ten-year manufacturing agreement worth up to $120 million to move a commercial cell-therapy programme onto its IRO manufacturing platform. Cell therapies have enormous potential, but they have historically been difficult to scale; Ori’s platform is designed to turn them into something that can be manufactured reliably, which is exactly what the sector needs. Nuclera appointed Carl Raimond as Chief Executive, while co-founder Michael Chen moved into the role of Chief Scientific and Business Officer, a leadership change that signals a shift toward commercial execution. Oxford Nanopore moved into the operating theatre, joining an NHS brain-tumour pilot that is exploring whether rapid sequencing can classify tumours during surgery and support faster treatment decisions. The company also joined RISE, a global rare-disease network, expanding access to genomic testing for families with unresolved genetic conditions. Behind all of these updates is a simple conviction: the sooner we know what is happening inside a body, the sooner we can act. Whether that means identifying a pregnancy complication, classifying a tumour in the middle of surgery, or giving a family a name for a rare disease, the technology is moving from the lab bench to the bedside, and from the research paper to the operating theatre.
On the Planet side, materials and space were the focus. Paragraf joined Korea’s K-Graphene Commercialisation Council as its only international member, supporting efforts to move graphene technologies from research into industry. The company also received three awards recognising its scale-up progress, international growth, and innovation, while CEO Simon Thomas was named Business Leader. Perhaps more importantly, Paragraf moved beyond graphene, becoming the lead industrial partner in the £6.4 million NEED2D programme, which is developing ultra-low-energy electronics using next-generation two-dimensional materials. In space, OKAPI published new technical analysis modelling what happens when satellites break apart and how that affects the orbital environment, an increasingly urgent question as constellations grow and debris becomes a real threat. Behind these portfolio companies are people, and this month two colleagues added their voices to important conversations. Amelia Armour, a Partner, was quoted in The Observer on AI security after speaking on a panel at the UK Private Capital Summit about keeping data in Europe and the semiconductor market. Anne Glover, co-founder and CEO, joined the Founders Forum Group VC Forum to discuss how to unlock more pension fund capital for innovation and growth. If there is a common thread running through all of this, it is resilience: resilient supply chains, resilient networks, resilient health systems, and resilient organisations. Trust takes years to earn, minutes to weaken, and far longer to rebuild, but with the right technology, the right attention, and the right partnerships, it can be protected. That was the lesson of Mansion House, and it is the theme running through this entire issue of Capitalise. If you enjoyed reading it, why not subscribe below to receive it directly to your inbox?