Frontier Focus

DCW Frontier Focus Edition 46

By Eric Williamson
DCW Frontier Focus Edition 46

DCW FRONTIER FOCUS

Your Weekly Technology Intelligence Brief | November 22nd 2026

Artificial Intelligence, Cyber Security, Digital Infrastructure, Energy Technology & Quantum Innovation

Frontier Focus Edition 46

 

EXECUTIVE SUMMARY

  > Anthropic makes the largest stock market debut on record, raising $81.3 billion on Nasdaq; shares opened at $111.50 against an $83 offer price and closed the first day at $97.17, valuing the Claude developer at around $1.75 trillion.

  > The European Commission confirms an AI-assisted cyberattack that disrupted staff payroll, leave and medical reimbursement systems for more than ten days, as Google warns that fully autonomous AI attacks could emerge within six months.

  > Nvidia, now led by interim Chief Executive Colette Kress, reports quarterly revenue of $115.8 billion, up 103 per cent, but confirms it has stopped shipping US-licensed H200 chips to China.

  > Pacific Fusion receives New Mexico's first tritium licence for a private fusion company, while France adopts a ten-year national fusion strategy and Japan backs Helical Fusion with up to around $126 million.

  > Google books Blue Origin's New Glenn rocket for a 2029 launch of its Project Suncatcher AI computing satellites, as new US rules close the overseas remote-access route to advanced AI chips.

  > Alice & Bob doubles its 2030 target to 200 logical qubits, and an 8,000 km transatlantic quantum-safe network linking Chicago, Amsterdam and Bilbao is demonstrated at the SC26 supercomputing conference.

 

 

  🤖  ARTIFICIAL INTELLIGENCE

Anthropic Lists on Nasdaq in the Largest Stock Market Debut on Record

Anthropic, the developer of the Claude AI models, began trading on Nasdaq on Thursday 19th November under the ticker ANT. The company priced its shares at $83 on 18th November, above its planned range of $70 to $78, raising $81.3 billion from the sale of 980 million shares. Including $26 billion of private financing arranged alongside the listing, combined proceeds reached $107.3 billion, making it the largest initial public offering ever recorded.

Shares opened at $111.50, 34 per cent above the offer price, briefly valuing the company at around $2 trillion. The early enthusiasm faded through the session and the stock closed at $97.17, a gain of 17 per cent, on a day when both the Nasdaq and the S&P 500 fell by more than 2 per cent amid wider concerns about a technology bubble. At the close, Anthropic was valued at roughly $1.75 trillion including employee share options.

The listing makes Anthropic the first frontier AI developer to trade publicly, placing its finances, governance and safety commitments under the disclosure regime of a listed company for the first time. Chief Executive Dario Amodei said before trading began that the listing only strengthened the company's resolve to act responsibly. Commentators have nonetheless questioned whether the company's very large infrastructure commitments can be supported by future profits at this valuation.

 

STRATEGIC IMPLICATION

Anthropic's listing gives investors, customers and regulators a new source of audited, regularly reported information on the economics of a frontier AI developer. Organisations that rely on Claude or competing models should monitor quarterly disclosures for signals on pricing, capacity and supplier concentration. Boards should also recognise that the share price will now act as a public barometer of AI sentiment, and that volatility in this one stock may influence funding conditions across the wider AI sector.

 

AI Agents That Operate Computers Reach the Mainstream, Raising Privacy Questions

Google has added a computer use capability to its Gemini app for the Mac, allowing the assistant to work across native applications, local files and websites in the background while the user carries on working. Users see a live preview and can pause, stop or approve actions before they happen. The feature is available in English in the United States to subscribers of Google's AI Plus, Pro and Ultra plans.

Bloomberg reported that Apple is recruiting a senior engineering manager to build its own "AI agent that can use a computer", capable of browsing the web, working across applications and writing code. According to the same report, macOS 27.3 will warn users against granting full access to third-party agents. Apple has reportedly considered a launch as early as the end of next year, although the project remains at an early stage.

The risks were illustrated by a Reuters report on Meta's Muse agent, in which employees said the assistant had surfaced childcare requests, orders and bank balances that did not belong to them. Meta said staff test versions with features that are not available to the public. Muse launched in September and has passed 10 million downloads.

 

STRATEGIC IMPLICATION

Agents that can act across a user's applications create a far larger privacy and security surface than chat assistants. Organisations should set clear policies on which agents may be installed on corporate devices, what level of system access they may be granted and how their actions are logged. Data protection officers should treat agent access to personal data as processing that requires assessment under UK GDPR, particularly where agents can read messaging or financial applications.

 

Microsoft Competes on Price with Seven New In-House AI Models

At its Ignite 2026 conference, which ran from 17th to 20th November, Microsoft AI released a second wave of in-house models under its MAI brand. The lineup covers coding, cybersecurity, image editing, real-time speech and transcription, together with a routing model that decides which model should handle a given request. The releases include MAI Code 2, MAI Code 2 Flash, MAI Cyber 2 Flash, MAI Image 2.7 and MAI Realtime 1.

Microsoft's pitch centres on cost per task. The company claims MAI Code 2 delivers 96 per cent of the coding performance of Anthropic's Claude Opus 5.5 at about 5 per cent of the price, while MAI Realtime 1 is priced from $0.01 per audio minute. Microsoft also introduced MAI Licensed Weights, which allow enterprises to customise an MAI model and run it within their own Azure environment.

The release reflects a broader shift in the enterprise market from headline capability towards the economics of running AI agents continuously. Microsoft, a major investor in and distributor of OpenAI's models, is increasingly positioning its own models as alternatives to its partners' products. Separately, Fortune reported that Amazon Web Services now allows AI agents to purchase software autonomously through its marketplace.

 

STRATEGIC IMPLICATION

Vendor benchmark claims should be validated against an organisation's own workloads before procurement decisions are made. The emergence of credible lower-cost models strengthens the case for multi-model architectures that route routine tasks to cheaper models. Finance teams should also put controls in place for agent-initiated purchasing, including spending limits, approval thresholds and audit trails, before such capabilities are enabled.

 

  🔐  CYBERSECURITY

AI-Assisted Cyberattack Disrupts European Commission Staff Systems

The European Commission has confirmed that a cyberattack involving artificial intelligence tools disrupted internal systems used by its staff for more than ten days. Affected services reportedly included Sysper, the human resources platform used for payroll, leave requests and working-time records, and MyPMO, the medical expense reimbursement system. Investigators are still assessing whether personal data was taken. Politico first reported the incident on Thursday 19th November.

CERT-EU, the cybersecurity service for EU institutions, found that threat actors had used ARTEX, a freely available AI penetration testing framework, to automate campaigns against the institutions it protects, although it has not confirmed that ARTEX was used in this specific attack. No group has been publicly blamed and no ransom demand has been reported. Sysper and MyPMO resumed operations earlier this week.

The incident follows a report from Google's Threat Intelligence Group, published on 16th November, describing what it called the first fully automated vulnerability discovery and exploitation carried out by malicious actors. Google expects attacks requiring little or no human involvement within three months, and sophisticated, fully autonomous attacks within six. The FBI and Amazon warned on 13th November that attackers using ARTEX can reach privileged access in under three hours.

 

ACTION REQUIRED

Organisations should assume that attackers can now scan, exploit and move through networks at machine speed. Immediate priorities are rapid patching of internet-facing systems, phishing-resistant multi-factor authentication for all privileged accounts and monitoring capable of detecting automated reconnaissance. Incident response plans should be tested against scenarios in which the time from initial access to impact is measured in hours rather than days.

 

Expedia Confirms Third-Party Breach as Extortion Group Claims Nearly One Billion Records

Expedia Group has confirmed that attackers reached its systems through a compromised third-party provider, after a newly formed Russian extortion group calling itself n0n claimed to have stolen nearly one billion traveller-related records. Expedia, whose brands include Hotels.com, Vrbo and Travelocity, has said the group's claims are unverified and that its investigation is continuing.

The group claims to hold more than 830 million loyalty programme records from 35 programmes, 333 million consumer account records and 58 million traveller profiles, and has published a 500 megabyte sample. Researchers at Wiz and Sentra say the intrusion used a stolen vendor key from Qubole, a Google Cloud partner, which gave the attackers a trusted path into Expedia's big data platform. Unlike traditional ransomware gangs, n0n focuses on data theft and extortion rather than encrypting systems.

It is the second major supply chain incident in a week. On 16th November, malicious code was served to visitors of websites belonging to 178 organisations, including the BBC, HSBC, Zoom and IKEA, through the Optimizely website testing platform. The code was designed to swap cryptocurrency wallet addresses, skim payment cards and target older iPhones. Optimizely said attackers had abused a flaw in its account sign-up process, which it has since fixed.

 

ACTION REQUIRED

Third-party platforms with privileged access to data or customer-facing websites must be treated as part of the organisation's own attack surface. Security teams should inventory vendor keys and API credentials, rotate those with broad access and apply least-privilege principles. Websites that load third-party scripts should deploy content security policies and script integrity monitoring. Customers of affected travel brands should be alert to targeted phishing that uses genuine booking details.

 

Ukraine and FBI Arrest Ransomware Suspects Linked to $40.5 Million in Payments

Ukraine's National Police, working with the FBI, announced on 18th November the arrest of four suspected members of an international ransomware organisation linked to more than $40.5 million in ransom payments from at least 28 companies worldwide. The arrests followed a joint investigation lasting more than two years.

According to Ukrainian authorities, the group's members posed as IT support staff on telephone calls to gain access to corporate networks. Investigators linked the organisation to a wider network of more than 60 people operating across eight jurisdictions, including former members of the Conti ransomware gang. The Record reported that victims included US oil and healthcare companies.

The arrests add to recent momentum in law enforcement action. Earlier this month, the Europol-supported Operation Endgame took down 235 servers and 257 domains linked to the Gootloader, CastleRAT and StealC malware families and recovered around 6.4 million stolen credentials. StealC had, however, already rebuilt after an earlier takedown in June, a reminder that disruption often pauses criminal activity rather than ending it.

 

STRATEGIC IMPLICATION

Telephone-based social engineering, in which attackers impersonate help desk staff, remains one of the most effective routes into corporate networks. Organisations should require strong identity verification for any help desk interaction involving password resets or remote access, and train staff to verify unsolicited support calls through known channels. Law enforcement successes are welcome but should not reduce investment in prevention, given how quickly criminal infrastructure is rebuilt.

 

  ⚡  ENERGY TECHNOLOGY

Pacific Fusion Secures New Mexico's First Private Fusion Tritium Licence

The New Mexico Environment Department has approved a radioactive materials licence for Pacific Fusion, authorising the company to possess and use tritium, a radioactive form of hydrogen used as fusion fuel, at its $1 billion Research and Manufacturing Campus in Albuquerque. Announced on 20th November, it is the first such licence the state has issued to a private fusion company.

The licence authorises up to 15,000 curies of tritium, around 1.5 grams, although the company says it never intends to hold more than 10,000 curies. New Mexico regulates fusion under a framework similar to that used for hospitals and research facilities. Pacific Fusion, which is developing a pulsed magnetic fusion system, is targeting net facility gain by 2030 and expects the campus to create more than 200 permanent jobs.

Elsewhere in the private sector, TAE Technologies is upgrading its sixth-generation Norm reactor from 70 million to 100 million degrees Celsius, the threshold for deuterium-tritium fusion, with the target set for 2027. TAE plans to carry the same systems into Da Vinci, a planned 50 megawatt plant intended to demonstrate fusion energy by 2031.

 

STRATEGIC IMPLICATION

Licensing is becoming as important as physics in determining which fusion ventures reach commercial operation. The New Mexico approval shows that fusion can be regulated proportionately under existing radioactive materials frameworks rather than the full fission regime, an approach the UK has also adopted. Investors and policymakers should track regulatory readiness alongside technical milestones when assessing fusion timelines.

 

France and Japan Step Up State Backing for Fusion Energy

France has adopted a ten-year national fusion strategy, unveiled by Philippe Baptiste, Minister for Higher Education, Research and Space, during a visit to the ITER site and the CEA's Cadarache centre. The strategy centres on three facilities: ITER, the CEA's WEST tokamak and the Laser Megajoule inertial confinement facility. Its medium-term goal is to develop the technologies needed for a European pilot fusion plant by 2035, and it aims to double the number of students in plasma physics, nuclear engineering and cryogenics.

In Japan, Helical Fusion announced on 18th November that it had won a maximum subsidy of about 19.2 billion yen (around $126 million) under a programme backed by the Ministry of Economy, Trade and Industry, bringing its cumulative funding and financing commitments to about 33.2 billion yen. The money supports Helix HARUKA, a demonstration device in Toki, Gifu, which will begin testing its high-temperature superconducting coils in 2027, and design work for Helix KANATA, its first power plant.

In the UK, First Light Fusion is partnering with the Central Laser Facility at the Rutherford Appleton Laboratory to develop fast ignition physics for its FLARE approach, with the aim of reducing the laser energy needed for ignition. Separately, a French state auditor's report highlighted financing challenges facing the government's plan to build six new pressurised water reactors, a reminder that fission and fusion programmes are competing for constrained public budgets.

 

STRATEGIC IMPLICATION

State support for fusion is shifting from research grants towards industrial strategy, with national roadmaps tied to skills, supply chains and pilot plants. UK organisations involved in fusion should note that France and Japan are positioning to capture supply chain and skills advantages, and that partnerships with European programmes may offer access to facilities and funding. Workforce planning in plasma physics and cryogenics deserves early attention, as every major programme is competing for the same scarce talent.

 

Study Finds America's Largest Data Centres Use Nearly 2 Per Cent of US Electricity

A Harvard-led study published on 16th November in Nature Sustainability estimates that 403 hyperscale data centres in the United States consumed around 82 terawatt-hours of electricity between May 2024 and April 2025, equivalent to about 1.8 per cent of national electricity consumption. Across the study's range of assumptions, consumption was between 68 and 99 terawatt-hours.

Around 54 per cent of the electricity attributed to these facilities came from fossil fuels, principally natural gas at 38 per cent and coal at 15 per cent. The researchers linked the facilities' electricity use to roughly 25.7 million tonnes of carbon dioxide emissions, two to three times an earlier estimate for 2018. Four states, Virginia, Oregon, Ohio and Iowa, account for more than half of the electricity used by the facilities studied.

The findings arrive as nuclear power continues to expand elsewhere. Bangladesh's Rooppur nuclear power plant began supplying electricity to its national grid this week, and the International Atomic Energy Agency has raised its long-term projections for a sixth consecutive year, forecasting in its high case that global nuclear capacity could reach around 3.4 times its 2025 level by 2060, driven in part by demand from AI data centres.

 

STRATEGIC IMPLICATION

The study provides one of the first facility-level estimates of the energy and carbon footprint of AI-era data centres. Organisations reporting Scope 2 and Scope 3 emissions should seek location-specific emissions data from their cloud and AI providers rather than relying on corporate-level renewable energy claims. The concentration of demand in a handful of US states also highlights grid resilience risks for workloads hosted in those regions.

 

  🏗️  DIGITAL INFRASTRUCTURE

Nvidia Beats Forecasts Under Interim Leadership but Halts Licensed Chip Sales to China

Nvidia reported revenue of $115.8 billion for the third quarter of its 2027 financial year, ended 25th October, up 103 per cent year-on-year and 20 per cent on the previous quarter, and ahead of analyst expectations of around $109 billion. Data centre revenue reached $107.6 billion, and the company confirmed that its Vera Rubin NVL72 platform has begun shipping. Nvidia guided fourth-quarter revenue to about $130 billion.

The results were the first under interim Chief Executive Colette Kress, the company's long-serving Chief Financial Officer, who took charge after co-founder Jensen Huang resigned on 12th November amid reports about his personal relationships that prompted an independent board inquiry. Harvey C. Jones has been appointed the first chair of Nvidia's board, which is searching for a permanent successor and has formed a new panel to oversee national security and export-control compliance.

On the earnings call, Kress confirmed that Nvidia is no longer shipping US-licensed Hopper data centre products, including the H200, to China, and that its guidance assumes no such sales. The company sold $2.4 billion of these products to China in the quarter. Beijing has been restricting imports of the H200 as it encourages domestic alternatives, and around 600,000 of the chips were reportedly being held in Hong Kong awaiting approval.

 

STRATEGIC IMPLICATION

Nvidia's results confirm that demand for AI infrastructure remains exceptionally strong, but a leadership transition combined with the effective closure of the China market introduces new governance and geopolitical risk for the industry's dominant supplier. Organisations dependent on Nvidia hardware should monitor the succession process and any changes to supply allocation, and continue to evaluate alternative accelerator platforms as part of supply chain resilience planning.

 

Google Books Blue Origin to Launch AI Computing Satellites in 2029

Google has signed a launch services agreement with Blue Origin to carry satellites for Project Suncatcher, its research programme into AI computing in space, on a New Glenn heavy-lift rocket in the first quarter of 2029. The companies announced the deal on 19th November. Each satellite will carry Google's Tensor Processing Units and roughly the computing capacity of a data centre rack.

The satellites will be placed in a sun-synchronous orbit that provides near-continuous sunlight. Google says satellites in such orbits can generate up to eight times more solar power than panels on Earth, which forms the energy case for the project. Google plans to link two prototype satellites in 2027 ahead of the larger launch. Neither company has disclosed the number of satellites or the financial terms.

Significant engineering problems remain unresolved. Chips in space cannot be cooled by air, satellites must fly in tight formation to communicate at high bandwidth, and the economics must compete with continually improving ground-based infrastructure. Google is not alone in exploring the concept: SpaceX has outlined plans for up to one million orbital AI satellites, and a start-up has filed with the US communications regulator to launch up to 100,000.

 

STRATEGIC IMPLICATION

Orbital computing remains a long-horizon research bet rather than a near-term procurement option, but it shows how severely energy and land constraints are now shaping AI infrastructure strategy. Organisations should continue to prioritise terrestrial capacity planning, power availability and efficiency. Regulators and space agencies, including in the UK, will need to consider the spectrum, orbital debris and data jurisdiction questions raised by computing infrastructure in orbit.

 

New US Rules Close the Overseas Route to Advanced AI Chips

The US Commerce Department has issued an interim final rule that treats remote access to advanced AI chips through overseas data centres in the same way as a physical chip shipment, closing a route that Chinese AI developers have used to reach hardware they cannot buy directly. Data centre operators currently serving restricted customers will have 90 days after the rule takes effect to obtain a licence or lose access to future US hardware.

The rule also introduces licensing requirements for advanced chip shipments to Singapore, Malaysia, Indonesia, Thailand and Brazil, which Washington identifies as presenting an elevated risk of diversion, and formally cancels the Biden administration's January 2025 AI Diffusion Rule. Data centres, foreign foundries and chip exporters must carry out due diligence on the ownership of their customers.

The effect appears to have been immediate. According to The Information, Beijing-based Moonshot AI, developer of the Kimi models, cancelled a $150 million agreement to rent Nvidia Blackwell Ultra B300 capacity through a data centre in Indonesia. Moonshot, Nvidia and the host operator have not confirmed the report. Analysts expect other Chinese-funded overseas data centre projects to reconsider their orders.

 

STRATEGIC IMPLICATION

Cloud and data centre operators in South East Asia and beyond now face substantially heavier compliance obligations, including checks on the beneficial ownership of their customers. UK-based providers and their clients should review whether any compute arrangements could involve restricted end users, as export control exposure increasingly attaches to services as well as goods. Firms should expect continued divergence between US, Chinese and allied technology regimes.

 

 

  ⚛️  QUANTUM COMPUTING

Alice & Bob Doubles Its 2030 Target to 200 Logical Qubits

Paris and Boston-based Alice & Bob has published an updated roadmap that doubles its 2030 target to 200 logical qubits, to be delivered by a system called Carbon built from just 3,600 physical qubits. The company's previous roadmap, published in December 2024, targeted 100 logical qubits by 2030.

Alice & Bob uses so-called cat qubits, which suppress one type of quantum error at the hardware level. This means far fewer physical qubits are needed to build each error-corrected logical qubit: the new roadmap implies around 18 physical qubits per logical qubit for Carbon, compared with roughly 67 for its Beryllium system planned for 2028. Carbon would link nine modules together, and the company describes it as its first utility-scale system, intended to surpass classical computers on materials science simulations.

The announcement should be read with appropriate caution. Alice & Bob has revised its timing before, and intermediate milestones on its roadmap have yet to be publicly completed. Nonetheless, the reduction in hardware overhead, if achieved, would substantially lower the cost and energy demands of fault-tolerant quantum computing.

 

STRATEGIC IMPLICATION

Error-correction efficiency is becoming the decisive competitive measure in quantum hardware. Organisations tracking quantum timelines, particularly for cryptographic risk, should pay attention to logical qubit roadmaps rather than physical qubit counts. Each credible reduction in overhead brings forward the point at which commercially and cryptographically relevant machines become feasible.

 

Transatlantic Quantum-Safe Network Spanning 8,000 km Demonstrated at SC26

A consortium of research and technology organisations has demonstrated a hybrid quantum-safe network linking Chicago, Amsterdam and Bilbao across more than 8,000 kilometres at SC26, the supercomputing conference held in Chicago this week. Partners included Luxquanta, Ciena, the Dutch research network SURF, the University of the Basque Country, GÉANT, Poland's PCSS and Spain's RedIRIS.

The network combines two complementary approaches: quantum key distribution, which uses the physics of light to exchange encryption keys, and post-quantum cryptography, which relies on new mathematical algorithms designed to resist quantum attack. It protects data at three levels, from the optical fibre up to applications, and uses standardised interfaces so that key management systems from different suppliers can work together, relaying keys through intermediate nodes where needed.

The demonstration builds on earlier commercial work, including a 1.6 terabit-per-second quantum-safe encryption trial on a live US network completed by Ciena, Toshiba and Quantum Corridor in August. SC26 organisers said this year's conference featured 55 quantum computing vendors, a sign of how closely quantum technology is now tied to mainstream high-performance computing.

 

STRATEGIC IMPLICATION

The demonstration supports the view that quantum-safe protection will be layered rather than reliant on a single technology. For most organisations, migration to NIST-standardised post-quantum cryptography remains the practical priority, consistent with NCSC guidance, while quantum key distribution is likely to remain confined to high-value links. Interoperability between suppliers is essential to avoid lock-in and should be written into procurement requirements.

 

Nvidia, IBM and Fermilab Plan Platform to Bring Quantum and Supercomputing Together

Nvidia, IBM and the US Department of Energy's Fermi National Accelerator Laboratory plan to build a modular quantum computing codesign platform that places Nvidia DGX B200 systems alongside IBM quantum computers, including machines powered by IBM's 120-qubit Nighthawk processor. Announced at the start of SC26, the platform will support research into error-correction codes, decoding algorithms and hybrid quantum-classical applications, with an initial focus on high-energy and nuclear physics.

Alongside the announcement, Nvidia released TCBP-JOBS, a decoding algorithm within its CUDA-Q platform that it says delivers up to a 13,000-fold speedup for quantum error-correction research. Decoding, the classical computation that interprets error signals from a quantum processor in real time, is one of the main bottlenecks in building fault-tolerant machines, and its demands grow rapidly as systems scale.

The modular theme continued later in the week, when Infleqtion and memQ signed a memorandum of understanding to support the clustering of quantum computers. Linking several smaller processors, rather than building a single very large one, is increasingly seen across the industry as the most realistic route to machines powerful enough for commercially useful work.

 

STRATEGIC IMPLICATION

Quantum computers are likely to operate as accelerators within classical supercomputing environments rather than as standalone machines. Organisations planning quantum pilots should consider integration with existing high-performance computing and GPU infrastructure, scheduling tools and skills from the outset. The growing role of classical decoding also means that quantum progress will depend partly on continued access to advanced GPUs, linking quantum timelines to the same supply chains discussed elsewhere in this edition.

 

CONCLUSION

This week's edition, our first as a weekend publication, captures a technology landscape in which capital, capability and risk are all accelerating at once.

In artificial intelligence, Anthropic's record-breaking listing has made a frontier AI developer a publicly traded company for the first time, bringing new transparency but also exposing the sector's valuations to daily market judgement. At the same time, AI agents that can operate computers on a user's behalf are moving into mainstream products, raising privacy and governance questions that organisations need to address now.

In cybersecurity, the European Commission incident and Google's warning about autonomous attacks indicate that AI-enabled intrusion is no longer theoretical. The Expedia and Optimizely incidents underline that third-party providers remain one of the most exploited routes into organisations, while the arrests in Ukraine show that international law enforcement cooperation continues to deliver results.

In energy technology, fusion is advancing on regulatory and policy fronts as well as technical ones, with New Mexico's tritium licence and new national commitments from France and Japan. The Harvard data centre study is a timely reminder that the AI boom currently depends heavily on fossil-fuelled electricity.

In digital infrastructure, Nvidia's results show undiminished demand, but its leadership change, the end of licensed China sales and new US remote-access rules all point to an increasingly politicised hardware market. Google's orbital computing plans illustrate how far the industry is prepared to look for power.

In quantum computing, the focus is firmly on error correction and integration, from Alice & Bob's more efficient roadmap to the convergence of quantum and supercomputing showcased at SC26. Organisations should continue to prioritise post-quantum cryptography migration while these hardware advances mature.

DCW will continue to monitor these developments and provide timely intelligence as the landscape evolves.

 

DISCLAIMER: This newsletter is produced for general informational and educational purposes only. The content herein does not constitute legal, financial, regulatory or professional advice. The Digital Commonwealth Limited makes no representations or warranties, express or implied, as to the accuracy, completeness or timeliness of any information contained in this publication. Readers should seek independent professional advice before acting on any information contained herein. All opinions expressed are those of the author at the time of writing.

Copyright 2026 The Digital Commonwealth Limited. All rights reserved. This publication may not be reproduced, distributed or transmitted in any form without prior written permission from The Digital Commonwealth Limited.

ABOUT THE DIGITAL COMMONWEALTH LIMITED

The Digital Commonwealth Limited (DCW) is a technology intelligence and advisory firm committed to helping organisations navigate the complex landscape of emerging technologies. Our mission is to provide clear, actionable intelligence that empowers decision-makers across public and private sectors.

Published: 22nd November 2026  |  Authored by Eric Williamson, Director of Compliance and Risk