Argyll Data Development has signed a landmark deal with AI infrastructure company SambaNova to build the UK’s first sovereign AI cloud in Scotland, powered entirely by renewable energy, marking a major step towards sustainable data sovereignty and low-carbon artificial intelligence.

Who Is Behind The Project?

Two companies from opposite sides of the Atlantic are joining forces to redefine how AI infrastructure is built, powered and controlled. Argyll Data Development is a Scottish developer specialising in renewable powered digital infrastructure. Formed in 2023, the company’s goal is to establish a new model for data centres that combine energy independence with AI capability. Its flagship venture, the Killellan AI Growth Zone, will transform a 184-acre industrial site on Scotland’s Cowal Peninsula into a green digital campus that hosts both renewable energy generation and high-performance computing.

The other partner, California-based SambaNova Systems, founded in 2017 by former Sun, Oracle and Stanford engineers, designs specialised processors and software platforms for running advanced AI models efficiently. Its technology is already being used by governments, research institutions and enterprises to train and run large language models, with a growing focus on sovereign AI, meaning infrastructures where data stays under national control rather than being processed by global cloud giants.

First Fully Renewable Powered AI Inference Cloud

The new partnership will see Argyll and SambaNova create the UK’s first fully renewable powered AI inference cloud, where AI models are hosted and operated rather than trained. The facility will be built at Killellan Farm near Dunoon on Scotland’s Cowal Peninsula, forming the centrepiece of Argyll’s 184-acre Killellan AI Growth Zone. It will deploy SambaNova’s SN40L systems, a new air-cooled design that uses roughly one tenth of the power of conventional GPU systems, allowing high-density computing without energy-hungry liquid cooling.

Argyll will build and manage the data centre infrastructure and on-site renewable energy network, while SambaNova will supply and operate the AI platform. According to both companies, the project will provide UK enterprises with a secure and sustainable environment to develop and deploy AI systems, all within British borders.

First Phase

The first phase of the Killellan development will deliver between 100 and 600 megawatts of capacity, with plans to scale to more than 2 gigawatts once complete. It will run on a private-wire renewable network using on-site wind, wave and solar power, combined with vanadium flow battery storage for long-duration energy supply. This design will allow the facility to operate independently from the national grid in “island mode”, while still being engineered for future grid integration.

Why It’s Different From Other Data Centres

The Killellan AI Growth Zone stands apart from most data centres for reasons of sovereignty, sustainability and circularity. For example:

1. Its sovereign design. Data sovereignty has become a growing issue for both public and private sector organisations. It refers to keeping sensitive data and AI workloads within the same legal jurisdiction in which they originate. Argyll’s platform will ensure data processed at Killellan remains entirely within UK regulatory and security frameworks.

2. Its renewable-first approach. Instead of relying on grid power supplemented by renewable energy certificates, Argyll intends to generate all its electricity on-site using wind, wave and solar resources from the Cowal Peninsula. Vanadium flow batteries will store excess power, offering more stability than traditional lithium-ion systems.

3. The closed-loop design. Waste heat from the data halls, a by-product of high-intensity computing, will be captured and reused to support vertical farming, aquaculture and local district heating. The company says this will help the site operate as a “circular” digital ecosystem, recycling both energy and heat to minimise waste.

According to Peter Griffiths, executive chairman at Argyll, the project shows that “sustainability and scale can go hand in hand.” He said the goal is not only to make AI greener but also “competitive, compliant and cost-effective.”

Impact On Argyll And SambaNova

For Argyll, the project defines its core mission to create net zero infrastructure that advances UK energy and AI strategies simultaneously. The Killellan site is the company’s first major step in building large-scale digital capacity powered by renewable energy.

For SambaNova, it marks another milestone in a series of global sovereign AI projects. The firm has already supported similar renewable powered AI infrastructures in Australia and Germany. Rodrigo Liang, co-founder and CEO of SambaNova, described Argyll as “a blueprint for scaling AI responsibly”, adding that its systems are “enabling large-model inference with maximum performance per watt, while helping enterprises and governments maintain full control over their data and energy footprint.”

Economic And Regional Benefits

Argyll expects the project to attract up to £15 billion in total investment and create more than 2,000 construction jobs a year, along with 1,200 long-term operational roles. The company forecasts that it will contribute roughly £734 million annually to Scotland’s Gross Value Added once fully operational.

Located near Dunoon, the development also offers a powerful example of regional regeneration. The Cowal Peninsula has a long industrial history but limited modern investment. By repurposing the former quarry site into a hub for green digital infrastructure, Argyll hopes to revitalise the area and support new skills in engineering, energy and technology.

Also, the integration of local education partners, including Dunoon Grammar School and the University of Strathclyde, will aim to build a pipeline of digital and energy sector talent. The company says this collaboration will support both academic research and workforce development tied to the UK’s AI and net zero ambitions.

Users

For UK businesses, especially those handling regulated or confidential information, a sovereign AI cloud could solve two persistent problems, i.e., data security and compliance. For example, many enterprises currently rely on overseas cloud providers, raising questions about data handling, jurisdiction and privacy. Argyll’s system will give companies a domestic alternative. With its combination of renewable energy and energy-efficient hardware, it promises not only a smaller carbon footprint but also predictable energy costs insulated from volatile wholesale markets.

Industries such as finance, healthcare, logistics and energy could also benefit. For example, banks running fraud detection models or hospitals processing medical imaging data could use the cloud to keep sensitive workloads inside the UK while meeting environmental commitments.

Other Global Initiatives

Argyll’s model actually reflects a growing international trend towards sovereign, renewable AI infrastructure. For example, in Australia, SambaNova has partnered with SouthernCrossAI to develop SCX, the country’s first ASIC-based sovereign AI cloud powered entirely by renewables. In Germany, Infercom is preparing to launch an AI inference platform built on SambaNova technology, designed for GDPR-compliant, energy-efficient operation across the EU.

Elsewhere, hyperscale providers such as Microsoft, Google and Amazon have begun to invest heavily in renewable energy contracts for their European and US data centres. However, most still rely on external power purchase agreements rather than self-contained renewable generation, and few operate under fully sovereign data frameworks. Argyll’s combination of on-site renewables, energy storage and UK-only data jurisdiction makes it a distinctive model within this global landscape.

Sustainability And The Future Of AI Infrastructure

AI computing has become one of the fastest growing sources of data centre energy consumption. Recent studies have estimated that global AI workloads could consume as much power annually as a small country by the end of the decade. Projects like Killellan are, therefore, being closely watched as test cases for whether large-scale AI operations can be powered sustainably.

If successful, the site could demonstrate how renewables and advanced computing can coexist without compromising either capacity or carbon goals. Its closed-loop design, where heat and power are continually reused, offers a new benchmark for future AI and cloud campuses.

Challenges And Criticisms

Despite its ambition, the project faces several considerable challenges. For example, the technical complexity of generating hundreds of megawatts of renewable power on-site, combined with long-duration battery storage and the demands of AI cooling, will require significant capital investment and coordination.

There are also practical questions about whether it can truly operate entirely on renewable energy throughout the year. Variability in wind and solar output may still require grid imports unless the storage capacity is large enough to cover prolonged gaps. Independent monitoring will be important to verify the site’s energy sourcing and net zero claims.

Environmental groups are also likely to scrutinise its local impact. For example, while the site’s heat reuse and clean energy credentials are strong, data centre developments of this scale can still affect local ecosystems, landscapes and transport routes. Ensuring transparent consultation and equitable community benefits will be vital if the project is to maintain public support.

For the wider industry, Argyll’s venture highlights the pressure facing data centre developers worldwide to decarbonise operations and localise control of AI infrastructure. The success or failure of the Killellan AI Growth Zone could influence how other countries, and indeed major cloud providers, design the next generation of green, sovereign data campuses.

What Does This Mean For Your Organisation?

If the Killellan AI Growth Zone delivers on its promises, it could mark a turning point for both the data centre industry and the UK’s digital economy. By proving that high-performance AI computing can run on home-grown renewable energy, Argyll and SambaNova are attempting to demonstrate that energy security, data sovereignty and sustainability can all be achieved together rather than traded off against one another. This approach directly aligns with the UK’s ambitions to develop a competitive but responsible AI sector that also contributes to national net zero targets.

For UK businesses, access to a secure and fully sovereign AI cloud powered by renewable energy could give organisations in regulated sectors a compliant and lower-carbon alternative to global hyperscalers. It may also make advanced AI services more cost predictable by stabilising energy costs and reducing exposure to international data rules. For enterprises building or deploying AI, from financial firms to healthcare providers, that combination of energy independence and regulatory assurance could become a key differentiator in the years ahead.

For the Scottish economy, the project means that thousands of construction and long-term jobs are expected, along with skills partnerships and secondary industries such as vertical farming and district heating. However, local engagement and environmental transparency will determine whether those benefits are shared fairly and whether the project sets a genuine precedent for sustainable regional development.

For the wider data centre industry, Killellan is also likely to be a test case for a new model of infrastructure, one that links renewables, storage and high-density computing in a single closed system. If it succeeds, it could influence how sovereign AI facilities are built across Europe and beyond. If it struggles to meet its energy and performance goals, it will still serve as a valuable lesson on the limits of scaling AI sustainably. Either way, the project has already shown that the future of AI infrastructure will depend not only on processing power, but on how responsibly that power is generated, managed and shared.

Making your own agent is easier than ever with 365 CoPilot and this video gives you examples of what to use it for, as well as how to create the agent in the fist place.

[Note – To Watch This Video without glitches/interruptions, It may be best to download it first]

Did you know you can attach files from OneDrive or SharePoint directly to your Outlook emails? This feature saves time and streamlines your workflow.

To attach a file:

– Compose a new email in Outlook.

– Click on the “Attach File” button.

– Select “Browse Cloud Locations”.

– Choose OneDrive or SharePoint and select the file.

This feature is real and current in Outlook, making it easy to share files stored in your cloud storage without having to download and re-upload them. Give it a try!

Nvidia has become the first company in history to reach a $5 trillion market valuation, as investors bet that demand for its artificial intelligence chips will continue at record levels across global industries.

Who Is Nvidi? Why Does It Matter?

Nvidia began in the 1990s as a designer of graphics processing units (GPUs) for gaming computers. Those chips were originally built to render visuals quickly, but their architecture also made them ideal for performing many calculations at once, which is a capability that later proved critical for artificial intelligence (AI).

Over the past decade, Nvidia has moved far beyond gaming. For example, its GPUs now underpin most of the world’s AI infrastructure, powering data centres that train and run large language models such as ChatGPT. In fact, analysts estimate that Nvidia now controls more than 80 per cent of the global AI chip market.

The Advantage of CUDA

The company also supplies networking technology, entire server systems, and software platforms like CUDA that help developers build AI applications specifically for Nvidia hardware. CUDA, introduced in 2006, is one of Nvidia’s biggest competitive advantages. For example, once a business or research organisation builds its AI workloads on CUDA, it becomes difficult to switch to a rival chipmaker without rewriting large amounts of code (a high barrier to exit). This has created an ecosystem that ties thousands of AI start-ups, cloud providers, and universities to Nvidia’s hardware roadmap.

What Just Happened To Nvidia’s Valuation?

On 29 October, Nvidia’s share price rose more than 5 per cent in a single day to more than $212, lifting its total market capitalisation above $5 trillion. The company had reached $1 trillion only in June 2023 and $4 trillion just three months ago, marking an extraordinary rate of growth – even by technology sector standards!

The immediate catalyst was a string of announcements that reinforced investor confidence in Nvidia’s long-term dominance. For example, chief executive Jensen Huang told analysts that the company expects about $500 billion in AI chip orders over the next year. He also confirmed that Nvidia is building seven new AI supercomputers for the US government, covering areas such as national security, energy research, and scientific computing. Each of those projects will require thousands of Nvidia GPUs, underscoring the company’s position at the centre of the global AI race.

Investor optimism was also fuelled by geopolitics. For example, US President Donald Trump said he plans to discuss Nvidia’s new Blackwell chips with Chinese President Xi Jinping, raising expectations that Nvidia’s access to the Chinese market will continue. China is the company’s single largest overseas market, despite earlier US restrictions on the export of its most advanced AI chips. Nvidia has since agreed to pay 15 per cent of certain China-related revenues to the US government under a licensing arrangement introduced to manage those controls.

How Nvidia’s Chips Became Essential Infrastructure

Nvidia’s growth has been driven by the simple reality that modern AI systems consume huge amounts of computing power. Every new generation of models requires exponentially more data and processing capacity than the last. As a result, global cloud providers such as Microsoft, Amazon, and Google are spending tens of billions of dollars each quarter building new AI data centres, and almost all of them rely on Nvidia’s GPUs.

Jensen Huang’s long-term strategy has been to sell complete systems rather than individual chips. Nvidia now provides full server racks and networking systems optimised for AI workloads, creating an all-in-one platform that large customers can install and scale rapidly. The company’s H100 and newer Blackwell GPUs are currently considered the industry standard for training and running advanced AI models, including those used in robotics, autonomous vehicles, and scientific research.

Nvidia has also been expanding into telecommunications. Earlier this week, it announced a $1 billion investment in Nokia to help develop AI-native 5G Advanced and 6G networks using Nvidia’s computing platforms. The two companies said their goal is to bring artificial intelligence “to every base station” so that future mobile networks can process data and run AI models directly at the edge, reducing latency and improving security.

The Wider Tech Market

Nvidia’s $5 trillion valuation means it now sits ahead of both Apple and Microsoft, which have each passed $4 trillion. Its rise has also helped lift the broader US stock market to record highs, with AI-related firms accounting for around 80 per cent of gains in major indices this year.

The scale of spending linked to Nvidia is immense. For example, Microsoft reported capital expenditure of more than $35 billion in its last quarter, largely on AI infrastructure, and OpenAI recently confirmed that Nvidia will provide 10 gigawatts of computing power to support its future models. Oracle, Amazon, and Meta have all signed multibillion-dollar supply deals.

Self-Reinforcing Loop

This cycle creates what analysts call a self-reinforcing loop. For example, the tech firms buy Nvidia systems to build AI products, those products then demonstrate rapid adoption and revenue potential, which pushes investor optimism even higher, thereby allowing Nvidia to invest more heavily in its next generation of hardware. That in turn becomes the new industry standard.

It seems that competitors are now struggling to catch up with Nvidia. AMD, Intel, and several start-ups are developing rival chips, while Google and Amazon are designing in-house AI accelerators to reduce their reliance on Nvidia. Governments have also entered the race, i.e., China is backing domestic chipmakers to reduce dependence on US technology, and the European Union is investing heavily in semiconductor manufacturing capacity to avoid supply chain vulnerability.

The Impact On Governments, Businesses, And The Global Economy

Nvidia’s technology is now viewed as part of its country’s national infrastructure. For example, the seven US supercomputers being built with Nvidia hardware are intended to strengthen capabilities in defence, climate modelling, and scientific innovation. Access to leading-edge compute power has become a matter of strategic importance, with governments treating it as they once treated oil or rare earth metals.

For telecommunications providers, Nvidia’s partnership with Nokia signals a broader shift toward AI-driven networks that can manage themselves, predict faults, and run advanced analytics at the edge. Industry analysts at Omdia estimate that the market for AI-assisted radio access networks could exceed $200 billion by 2030.

For enterprise customers, the issue is essentially access and cost. For example, Nvidia’s GPUs remain scarce and expensive, and the waiting time for high-end systems can stretch into months. It is that scarcity that gives Nvidia immense pricing power and influence over who can deploy large-scale AI models. Businesses looking to integrate AI into their operations often find themselves competing with global tech giants for limited GPU supply, which can delay projects and inflate costs.

The company’s scale also affects financial markets. At $5 trillion, Nvidia’s value now exceeds the combined stock exchanges of every country in the world except the United States, China, and Japan. Its shares are held across pension funds and index trackers, meaning even small fluctuations in its price can move major global indices.

Growing Concerns About An AI Bubble

However, such rapid growth has led to mounting warnings about a potential AI-driven market bubble. The Bank of England, the International Monetary Fund, and several investment banks have all cautioned that valuations could fall sharply if the expected returns from AI adoption do not arrive quickly enough.

Analysts also point to what some describe as “financial engineering” in the AI sector, i.e., where companies invest in one another to sustain rising valuations. Nvidia has said it plans to invest up to $100 billion in OpenAI over the coming years, with both companies committing to deploy vast amounts of Nvidia hardware to power OpenAI’s future systems. Critics say such arrangements blur the line between commercial demand and strategic co-investment.

Tech Revolution Rather Than Speculative Excess?

It is also worth noting here that some market analysts argue that Nvidia’s growth reflects a genuine technological revolution rather than just speculative excess. Firms such as Ark Invest have suggested that AI remains at an early stage of development and that valuations could still have room to grow, even if a short-term correction occurs. That said, others, including analysts at AJ Bell, have pointed out that Nvidia’s valuation is almost beyond comprehension and likely to intensify debate over an AI bubble, although investors so far appear undeterred.

Trade Policy

Trade policy remains another risk worth mentioning here. For example, Nvidia’s share price briefly dipped in April when markets were shaken by renewed US-China tensions. Although President Trump has since reversed previous restrictions on advanced chip exports, the company’s reliance on Chinese demand makes it vulnerable to future policy changes. Beijing has been promoting local chipmakers and has already ordered state-linked companies to limit purchases of certain Nvidia models designed for the Chinese market.

For now, Nvidia’s impressive momentum looks set to continue. Its share price has risen more than 50 per cent since January, and its influence now extends across sectors from telecommunications to healthcare. Whether that momentum proves sustainable depends not just on how fast AI technology evolves, but on how far global investors are willing to believe in the story of an endless AI boom.

What Does This Mean For Your Business?

Nvidia’s extraordinary valuation is essentially a testament to the transformative potential of AI and a reminder of how concentrated that power has become. The company’s GPUs have effectively become the global standard for AI development, thereby embedding Nvidia deeply into the digital and economic infrastructure of almost every major industry. Yet its dominance also raises questions about long-term sustainability, supply resilience, and how much value creation is tied to speculation rather than fundamentals.

For investors and policymakers, the immediate concern is whether Nvidia’s growth reflects a permanent technological shift or a cycle of exuberance similar to previous tech booms. Central banks have already warned that such concentrated value could expose markets to sudden corrections if AI returns fall short of expectations. Nvidia’s role as both a supplier and investor across the AI ecosystem reinforces its strategic position, but also magnifies systemic risk if demand slows or policy barriers tighten.

For UK businesses, the implications are pretty significant. For example, access to Nvidia’s computing power underpins much of the AI capability now being adopted in sectors such as finance, logistics, healthcare, and manufacturing. As competition for GPUs remains intense, smaller firms may find themselves priced out or forced into cloud-based AI services that depend heavily on US infrastructure. This could deepen reliance on overseas providers unless the UK accelerates investment in domestic compute resources and training. The opportunity for innovation is vast, but so too is the risk of falling behind if access remains limited to global players.

At the same time, the broader technology ecosystem continues to adapt around Nvidia’s dominance. Competitors, governments, and research institutions are all pushing to develop alternative chips and software frameworks to reduce dependency. Whether those efforts succeed will determine how balanced the AI hardware market becomes over the next decade. For now, Nvidia’s scale gives it enormous influence over pricing, research priorities, and even the pace at which AI advances reach commercial use.

What happens next is likely to depend on whether real-world applications catch up with investor enthusiasm. If AI continues to deliver tangible productivity gains across sectors, Nvidia’s valuation may yet prove justified. If expectations cool, the company’s rise could be remembered as the peak of an era when optimism about artificial intelligence reshaped not only technology, but the structure of the global economy itself.

The Prompting Company has raised $6.5 million to help businesses get mentioned in AI-generated answers from tools like ChatGPT, Gemini, and Claude, signalling a major shift in how people now discover products online.

Who Is The Prompting Company?

The Prompting Company is a young, Y Combinator-backed startup (Y Combinator is a Silicon Valley startup accelerator) that wants to redefine online marketing for the age of artificial intelligence. Founded just four months ago by Kevin Chandra, Michelle Marcelline, and Albert Purnama, the company specialises in what it calls Generative Engine Optimisation, or GEO. The idea is that as people increasingly ask AI tools for advice instead of searching Google, brands must learn how to make their products visible to these systems.

The three founders, all originally from Indonesia, previously built Typedream, an AI-assisted website builder later acquired by Beehiiv, and Cotter, a passwordless authentication service bought by Stytch. Their latest venture reflects what many see as a fundamental turning point in digital discovery, i.e. AI assistants are becoming the new gateway to information, and by extension, to products and services.

Client List

The company’s early client list already includes Rippling (an HR and payroll software platform), Rho (a corporate banking and spend management platform), Motion (an AI-powered productivity and scheduling tool), Fondo (a tax automation platform for startups), Kernel (a data and machine learning infrastructure company), Traceloop (a developer observability platform), and Vapi (an AI voice agent platform), along with one unnamed Fortune 10 business.

What The Prompting Company Does

The startup’s service is actually built around a relatively simple process. For example, first, it identifies the kinds of questions AI systems are being asked in a particular market. Rather than focusing on short search keywords like “best business bank account”, GEO looks for longer, more contextual prompts such as “I’ve just set up a small company, what’s the best business account with no monthly fees?”

Once those queries are identified, The Prompting Company creates structured, machine-readable content that directly answers them. These AI-optimised pages strip away human-facing clutter like pop-ups, menus, and marketing slogans. Instead, they present clean, factual information written in a format that large language models can easily interpret and reference. The company then automatically routes AI crawlers to these pages instead of the brand’s normal website.

In short, it is search engine optimisation for AI rather than for humans. Its goal is to make brands “the product cited by ChatGPT”, as its own website puts it. The service operates on a subscription model, starting from $99 per month for basic tracking of 25 prompts and rising to enterprise plans with custom integrations and support.

The Trend

The Prompting Company’s entire business seems to rest on a simple observation that people are no longer starting their product searches on Google. They are asking AI assistants instead.

For example, Adobe’s 2025 Digital Economy Report found that US traffic from generative AI tools surged 4,700 per cent in a single year, with 38 per cent of consumers saying they had already used AI for shopping. Of those, 73 per cent said AI had now become their main tool for product research. The same report showed that visitors coming via AI assistants stayed 32 per cent longer on sites, viewed more pages, and were 27 per cent less likely to leave immediately. In other words, more shoppers are turning to generative AI tools to find deals, research products, and make buying decisions.

These changes suggest that AI assistants are now beginning to perform the filtering role that search engines once did. Instead of scrolling through links, users receive an instant shortlist of relevant products, often only two or three names. Being one of those names, therefore, has obvious commercial value.

Why Investors Are Paying Attention

That value explains why investors have been so quick to back The Prompting Company. The $6.5 million seed round, led by Peak XV Partners and Base10 with participation from Y Combinator and others, reflects growing belief that the next phase of digital advertising will take place inside AI assistants.

For investors, the logic is pretty straightforward. Whoever shapes how AI tools make product recommendations will control the top of the sales funnel for entire industries. Traditional search and pay-per-click advertising rely on visible results and bids for keywords. In AI-driven discovery, there may be no visible results page at all. An assistant could simply say, “You should try Rho for business banking,” and the conversation ends there.

That urgency among brands is reflected in the company’s own analysis, which suggests that much of the recent growth in website traffic is now coming from AI bots rather than human visitors. The founders say that developers are already using AI tools to ask for product recommendations inside their workflows, and they believe that ordinary consumers are beginning to do the same.

What Will The Funding Be Used For?

The startup says it will use the $6.5 million to scale its platform, develop AI-facing website templates for customers, and expand partnerships with major AI providers. It is also collaborating with Nvidia on “next-generation AI search”, though the details of that project have not been disclosed.

The company currently claims to host around half a million AI-optimised pages and to be driving double-digit millions of monthly visits for clients. Its customers span fintech, developer tools, and enterprise software, but the founders say the model applies to any sector where customers ask detailed, conversational questions.

The Lead In A New Sector

The funding gives The Prompting Company a clear lead in what could soon be a major new marketing sector. For example, by positioning itself as the first dedicated GEO platform, it is creating a new type of infrastructure for online visibility. The company argues that the fastest-growing “users” of the internet today are AI agents, not humans, and that brands need to design websites for those agents first.

It also aims to make GEO repeatable and data-driven, similar to how SEO matured into an industry over the past two decades. The difference is that in AI discovery, results are generated dynamically rather than ranked on a static page, meaning brands will need constant updates to stay visible.

Competitors

The rise of GEO is highly likely to unsettle traditional SEO agencies and digital advertisers. The overlap between Google search results and AI recommendations is shrinking, with some analyses suggesting it has dropped from around 70 per cent to below 20 per cent. That means a brand ranking first on Google might not even appear in an AI assistant’s answer.

Agencies built around keyword bidding and link optimisation now face the challenge of learning how to influence AI-generated answers, which rely on context and relevance rather than metadata and backlinks. This transition could change how marketing budgets are allocated, with more money flowing towards GEO-style services.

AI Companies

For companies like OpenAI, Google, Anthropic, and Meta, this trend could be an opportunity as well as a risk. For example, on the one hand, AI-driven shopping and product discovery could open new sources of revenue, especially as assistants move beyond recommending items to actually completing purchases. OpenAI’s recent integration with Stripe, for example, already allows ChatGPT to handle some transactions directly.

On the other hand, questions around bias and commercial influence are inevitable. For example, if AI assistants begin recommending brands that have paid for optimisation or have supplied AI-friendly data, users may expect clear disclosure of those relationships. Transparency will become crucial as assistants start to resemble personal shoppers or product curators.

There are also technical implications to consider here. GEO depends on AI models being able to browse the open web and ingest structured content. ChatGPT, Gemini, and Perplexity can already do this, but others, such as Anthropic’s Claude, have been more limited. This could lead to a divided ecosystem with some assistants open to optimisation, while others keep recommendations strictly in-house.

Businesses And Advertisers

For businesses, the important message is that appearing in AI-generated answers may soon matter as much as appearing on page one of Google once did. The Prompting Company claims its system allows even small or new brands to compete by creating high-quality, context-aware content that AI tools are more likely to cite.

Early signs suggest that AI-driven traffic, while smaller in volume than search, may be higher in quality. For example, Adobe’s data shows that visitors arriving from AI recommendations tend to stay longer and are more focused on buying decisions. They also use AI most often for complex or big-ticket purchases, where research matters more than impulse.

For advertisers, however, it also poses some new questions, such as how do you measure success when a chatbot’s conversation, not a click, triggers a purchase? How do you influence visibility in an algorithm that changes with every prompt? Also, how do you maintain brand trust when recommendations are made by machines rather than people?

Challenges And Criticisms

As with any fast-moving technology, the rise of generative engine optimisation (GEO) raises a number of ethical and practical questions for both businesses and consumers.

The first challenge is transparency. For example, if brands start paying to be mentioned by AI, users must be able to tell whether a recommendation is organic or commercially influenced. Regulators could extend existing advertising disclosure rules to AI assistants, just as they have done with influencer marketing.

Bias is also a key issue to consider. AI systems are only as balanced as the data they are trained on, and introducing commercial optimisation risks amplifying existing inequalities. Studies of AI in retail have already raised concerns about how these systems collect and use customer data, and whether they treat all consumers fairly. Experts have warned that businesses must prioritise transparency, bias testing, and responsible data use if AI-driven commerce is to gain public trust.

Another challenge is attribution. For example, AI traffic still converts at lower rates than traditional search or social referrals, though the gap is narrowing. Marketers can’t yet prove, with precision, that being mentioned in an AI answer directly leads to a sale. Until that attribution problem is solved, investment in GEO may remain experimental for many firms.

Finally, there are issues around the subject of dependence. For example, if AI assistants become the main interface for product discovery, the brands that are mentioned will dominate attention, and those that are not may struggle to be seen at all. For now, The Prompting Company is just positioning itself as the bridge between those two realities, betting that businesses will soon have to market to AI agents as actively as they do to people.

What Does This Mean For Your Business?

If GEO takes hold in the way its backers expect, the structure of online discovery could change faster than many realise. For UK businesses in particular, this means rethinking how visibility is achieved and measured. Instead of fighting for Google rankings or paying for search ads, companies may soon need to consider whether their products can be understood, cited, and recommended by AI systems that are shaping what customers see first. That shift could favour agile firms that adopt AI-ready content strategies early, while leaving slower competitors struggling to appear in the new recommendation landscape.

For advertising and marketing industries, GEO could become both a challenge and an opportunity. For example, traditional SEO agencies may need to retrain their focus on machine-readable design, structured data, and conversational context, while media buyers could face a future where there are no clear ad slots to purchase. Instead, visibility might depend on maintaining technical partnerships, feeding accurate data to AI systems, and monitoring how generative models respond to brand information in real time.

AI companies also face growing scrutiny as these practices expand. If assistants begin to behave like digital sales representatives, they will need to explain how and why specific products are recommended. Regulators and consumer watchdogs will expect transparency around paid optimisation, and users will demand the ability to distinguish between genuine relevance and commercial influence. Maintaining public trust will require clear standards, and the companies that set them will likely shape the rules for everyone else.

For investors and innovators, the appeal is pretty obvious. GEO creates a new layer of infrastructure in the digital economy, one that could define how brands reach audiences as AI assistants replace search boxes. However, the broader outcome will depend on how responsibly the model is used. If transparency and fairness are built into the system from the start, AI-powered product discovery could simplify choices for consumers and open new routes to market for smaller firms. If not, it risks becoming another opaque advertising channel that benefits only those able to pay for visibility.

For now, The Prompting Company has positioned itself at the centre of that debate. Its technology reflects a future in which algorithms act as gatekeepers to consumer attention, and its early funding shows how much confidence investors have in that vision. Whether this transforms online marketing or simply adds another layer to it will depend on how quickly businesses, regulators, and AI developers adapt to a world where products must be marketed not only to people but to the machines that speak to them.

Australian regulators have taken Microsoft to court, alleging the company misled around 2.7 million Microsoft 365 users by implying they had to accept a higher-priced AI-powered plan or cancel altogether, while failing to reveal a cheaper alternative that was still available.

What Happened and Why?

The case focuses on Microsoft’s handling of its consumer subscription services, i.e., Microsoft 365 Personal and Family plans, used by millions of households for applications such as Word, Excel, PowerPoint, Outlook and OneDrive. These plans are sold on monthly or annual auto-renewing subscriptions, making them a cornerstone of many users’ digital routines.

Back in October 2024, Microsoft decided to integrate Copilot (its generative AI assistant) into Microsoft 365 Personal and Family subscriptions in Australia. The rollout later expanded worldwide in January 2025. Microsoft described Copilot as a major innovation, offering “AI-powered features” that would “help users unlock their potential”.

However, this integration also triggered quite a sharp price rise. For example, according to the Australian Competition and Consumer Commission (ACCC), the annual Microsoft 365 Personal plan increased from AUD 109 to AUD 159, a rise of 45 per cent, while the Family plan rose from AUD 139 to AUD 179, a 29 per cent increase. Monthly fees also went up.

Only Two Choices Implied

The ACCC says Microsoft notified subscribers through two emails and a blog post, telling them their next renewal would include Copilot and the higher price. The messages reportedly told users that unless they cancelled before renewal, the higher charge would apply automatically.

For example, one such email stated: “Unless you cancel two days before your renewal date, we’ll charge AUD 159.00 including taxes every year. Cancel any time to stop future charges or change how you pay by managing your subscription.”

The regulator now alleges these communications implied users had only two choices, i.e., pay more for Copilot, or cancel their subscription entirely. What the company failed to mention, according to the ACCC, was that there was a third option available, which was switching to what Microsoft called the “Classic” plan.

Classic Plan Is Third Option

Although the Classic plan allowed customers to retain all the features of their existing Microsoft 365 subscription, without Copilot, at the old price, it seems that it was not mentioned in Microsoft’s emails or blog post. Instead, the ACCC says the Classic option only appeared if a customer began the cancellation process, navigating through several screens before the option was revealed.

Given how integral Microsoft 365 has become to home users, e.g., providing essential software and cloud storage, the ACCC argues this created unfair pressure. ACCC chair Gina Cass-Gottlieb said: “The Microsoft Office apps included in 365 subscriptions are essential in many people’s lives, and given there are limited substitutes to the bundled package, cancelling the subscription is a decision many would not make lightly.”

Proceedings Filed Against Microsoft

On 27 October 2025, the ACCC filed proceedings in Australia’s Federal Court against both Microsoft Corporation in the United States and its Australian subsidiary, Microsoft Pty Ltd. The regulator alleges that Microsoft engaged in misleading or deceptive conduct, and made false or misleading representations, in breach of sections 18 and 29 of the Australian Consumer Law.

Specifically, the ACCC says Microsoft falsely represented that users had to accept Copilot to maintain access to their subscription (“Copilot Necessity Representation”), that they had to pay higher prices to continue using Microsoft 365 (“Price Necessity Representation”), and that they only had the two options of accepting the higher price or cancelling (“Options Representation”).

The ACCC claims these representations were false and misleading because the Classic plan was available at the old price, without Copilot, and that by omitting mention of that plan, the regulator says Microsoft denied customers the chance to make an informed decision.

Cass-Gottlieb stated: “We will allege in court that Microsoft deliberately omitted reference to the Classic plans in its communications and concealed their existence until after subscribers initiated the cancellation process to increase the number of consumers on more expensive Copilot-integrated plans.” She added: “We believe many Microsoft 365 customers would have opted for the Classic plan had they been aware of all the available options.”

The regulator argues that this omission caused consumers financial harm. Many subscribers, believing they had no alternative, allowed their subscriptions to renew automatically at the higher Copilot rate, and those users, the ACCC says, effectively paid more for something they might not have chosen.

Why It Matters

The case is significant because it highlights how software subscription models are evolving with the introduction of AI features. For example, Microsoft’s integration of Copilot, and the resulting price increases, demonstrates how companies are bundling AI capabilities into established services, but the ACCC argues that this bundling must be transparent and optional.

For consumers, the issue is both financial and procedural. A 45 per cent increase represents a notable cost rise for households relying on Microsoft 365. More importantly, the regulator argues that burying the cheaper Classic plan behind the cancellation flow deprived users of informed consent, which is a key principle in consumer law.

For Microsoft, the allegations really go beyond pricing. For example, the case also touches on interface design and user experience. Regulators are increasingly focused on so-called “dark patterns”, i.e., design choices that nudge users into particular decisions. The ACCC says Microsoft’s renewal flow was structured to steer users towards the more expensive plan by hiding the cheaper one.

For competitors, the case could shape how AI features are rolled out across subscription products. Companies like Google, Apple and Adobe are all integrating AI into consumer and productivity tools. If the court rules that Microsoft’s conduct was misleading, others may need to rethink how they communicate AI upgrades and pricing options.

Cass-Gottlieb said the regulator’s goal is broader than this single case: “All businesses need to provide accurate information about their services and prices. Failure to do so risks breaching the Australian Consumer Law.”

What Happens Next?

The (Australian) Federal Court will now review the ACCC’s evidence, including Microsoft’s October 2024 blog post and the two key emails sent to subscribers. The regulator is seeking penalties, injunctions, declarations, consumer redress and costs.

If the court finds against Microsoft, penalties could be substantial. For example, under Australian law, the maximum fine for each breach is the greater of AUD 50 million, three times the value of any benefits obtained, or 30 per cent of the company’s adjusted turnover during the breach period. The ACCC has signalled that it will seek a significant penalty, citing the number of affected consumers and the scale of the alleged conduct.

Microsoft – Reviewing The Claims

Microsoft has said it is reviewing the claims, adding that “consumer trust and transparency are top priorities” and that it intends to work constructively with the ACCC. The company has not yet filed a detailed defence.

Wider Context

The case comes at a time when regulators worldwide are scrutinising how big technology companies integrate AI into their products. Microsoft has made Copilot central to its software strategy, embedding it into Windows, Office and Bing. The integration has been marketed as a major advance, but it has also raised questions about whether AI is being used to justify higher subscription fees.

It’s worth noting here that, earlier in 2025, Microsoft faced separate antitrust scrutiny in Europe, where it agreed to unbundle Teams from Microsoft 365 after competition regulators raised concerns about unfair bundling. The Australian case is different in that it focuses on consumer fairness rather than competition, but both point to a growing willingness among regulators to challenge how Microsoft structures its product offerings.

The proceedings also coincide with a wider policy debate about consumer protection in digital markets. Regulators in the UK, EU and Australia have been warning companies against design choices that obscure cheaper or less data-intensive options. The ACCC’s case against Microsoft is one of the first major tests of these principles in the context of AI subscription pricing.

Challenges and Criticisms

Microsoft’s defence is expected to centre on whether its communications were genuinely misleading. For example, the company may argue that the price rise and integration were communicated transparently and that the Classic plan was a courtesy option, not an advertised tier.

Critics, however, say the case exposes how complex modern subscription models have become. Consumer advocates argue that when essential software like Microsoft 365 becomes tied to expensive AI add-ons, users may have little real choice, particularly if they are steered away from cheaper options through interface design.

The ACCC alleges Microsoft deliberately hid the Classic option to increase uptake of Copilot, describing the concealment as an intentional strategy. It says consumers’ dependence on Microsoft’s software made them more vulnerable to such tactics.

Meanwhile, business users and consumers have voiced frustration online. For example, some told Australian media they were surprised to find higher charges on renewal and were unaware that an alternative existed. Others have reportedly raised concerns that global technology companies may be using AI upgrades as a pretext for universal price increases.

The case is now being closely watched by regulators and consumer organisations worldwide, who see it as an early test of how AI-linked price changes will be governed under consumer law. For Microsoft, the outcome could determine how it promotes future Copilot features, and how transparent it will need to be with millions of subscribers when the next upgrade arrives.

What This Means For Your Business?

If the ACCC’s case succeeds, it could redefine how companies communicate subscription changes and AI integrations worldwide. The issues at stake extend far beyond Microsoft’s customer base in Australia. For example, transparency in pricing, honest representation of product features, and fair presentation of choices are all central to maintaining consumer trust in a digital economy that increasingly runs on subscriptions rather than ownership. The Court’s decision will, therefore, be closely analysed by consumer regulators, legal teams and software firms around the world.

For Microsoft, the financial penalties may be less significant than the reputational and operational consequences. The company’s strategy of embedding Copilot into every tier of its software ecosystem depends on users accepting AI features as a normal, even necessary, part of productivity software. If regulators conclude that the rollout was handled in a way that misled users, Microsoft may need to re-evaluate how it introduces future AI upgrades and how clearly it differentiates between optional and bundled products. Other technology firms will also be watching closely, given that most are following a similar path of building premium AI layers into existing subscriptions.

For UK businesses, the case highlights how global developments in consumer law can have local implications. The Competition and Markets Authority has already warned UK companies about interface design that conceals key information or discourages users from exercising choice. If Microsoft is found to have breached consumer law in Australia, it may prompt British regulators to take a closer look at how AI-driven services are marketed and priced in the UK. It could also encourage businesses that depend on Microsoft 365 to examine their own contracts and renewal processes more carefully, particularly where subscription changes are linked to new technologies or price adjustments.

The broader lesson is that as AI becomes more integrated into software, the boundary between innovation and obligation must remain clear. Consumers need to know when they are paying extra for AI functionality and when they can reasonably decline it. For regulators, the challenge will be to ensure that product evolution does not erode transparency or consumer control. For the tech industry, the message is that trust will be built not only through advanced technology, but through openness about what that technology costs, how it is delivered, and the real choices available to those who use it.