To support its AI ambitions, the UK is considering sovereign AI compute resources—essentially, government-backed AI supercomputing power to reduce reliance on foreign cloud providers.

But is this the best use of resources, or should the UK partner with industry to build AI infrastructure more efficiently?

The UK government has set an ambitious goal to train tens of thousands of AI professionals, aiming to build a world-class AI workforce. However, AI talent is in high demand globally, and current training initiatives may not be enough to close the skills gap.

The UK government’s proposal for a National Data Library (NDL) aims to create a centralised repository of public datasets to support AI research and innovation. The goal is to make high-quality, well-structured data available to businesses, researchers, and developers, accelerating AI advancements across multiple sectors.

One of the headline proposals in the UK’s AI Opportunities Action Plan is the creation of AI Growth Zones (AIGZs) areas where planning rules will be relaxed to fast-track the development of AI data centres and associated infrastructure. The government hopes this will attract investment, support regional development, and reinforce the UK’s AI capabilities.

The UK government has positioned artificial intelligence (AI) as a central pillar of its economic and industrial strategy. Framed as both a growth engine and a national priority, AI is expected to boost productivity, open up new markets, and reshape public services for the digital age.

In Part One, we explored the different AI deployment options, cloud, on-premise, and edge computing, along with their advantages and trade-offs.

In Part One, we explore the different approaches to AI deployment, cloud, on-premise, and edge computing, highlighting the key benefits and challenges of each.

In Part 1, we explored the fundamental differences between cloud-based AI and self-hosted AI solutions, outlining the key pros and cons of each approach. Now, in Part 2, we dive deeper into the practical considerations when choosing between these options and provide insights on how to make an informed decision for your business.

Artificial Intelligence is a transformative tool for businesses, but deploying AI solutions requires careful consideration of infrastructure, cost, and long-term strategy. One of the most critical decisions is choosing between cloud-based AI and self-hosted AI. Each approach offers distinct advantages and drawbacks, depending on your business needs, technical expertise, and budget.

Artificial intelligence is no longer a futuristic concept, it is an essential tool that businesses can leverage for efficiency, insight and competitive advantage.

However, building an AI capability requires investment into infrastructure, expertise and strategy. Depending on the type of AI you want to deploy, your business will need to make different considerations.

Artificial Intelligence offers businesses powerful tools to enhance products and services, but not every AI implementation delivers real value. For SMEs, the key is to determine whether AI genuinely improves outcomes, rather than adopting it for the sake of innovation. Below is a practical checklist to help you assess whether AI is the right fit for your business.

The rise of AI in business began with the widespread adoption of ChatGPT, developed from language translation algorithms. Almost immediately, every conceivable application of Large Language Models (LLMs) was soon identified, leading to the rapid emergence of targeted AI services designed to monetise ChatGPT and other LLMs.

AI is expected to revitalise the metaverse concept by providing more efficient ways to generate meaningful content. Whilst creative industries have valid concerns about evolving interpretations of copyright, only large organisations are likely to have the resources to defend their intellectual property within the emerging legal framework.

When Facebook rebranded as Meta in June 2022, it anticipated rapid growth in virtual and augmented reality. However, the structure of the market has not evolved as Meta had envisioned.

Annual financial reports for heritage organisations suggest that digital engagement is scarcely recognised as a revenue opportunity by management boards. Primarily, revenue comes from membership fees, in-person visits, café and gift shop sales, and benefactors.

The graph shows the positions of three types of corporate AI solutions, namely Chat GPT, DALL E and Neuralangelo. These three products are contemporary examples of Large Language Models, Generative AI and 3D image reconstruction, respectively.

DeepSeek is the most publicised result of these efforts, but other forms of neural networks have achieved even greater reductions in processing. We will discuss a case where elimination of uncertainty within the numerical method results in better performance, and how that has been achieved.

In our last blog post we introduced the concept of experts, which play an important role within Large Language Models.

Experts are one method of reducing the complexity of training and using an LLM translator. In daily life human experts have much greater confidence in their conclusions than those who lack experience, they avoid endless trawling through data by using their knowledge of a subject.

During the evolution of automatic translators, it was realised that the chance of a word occurring in a sentence was highly dependent on the structure of the sentence. A sentence that contains the word ‘tomato’ is much more likely to include ‘greenhouse’ than ‘particle accelerator’.