Business

Preply has raised $150 million in a Series D funding round led by WestCap, valuing the global language learning marketplace at $1.2 billion and marking its transition into unicorn territory.

The round, for which Goldman Sachs International acted as sole placement agent, will be used to accelerate product development, expand Preply’s AI and data capabilities and fuel international growth as the company seeks to reshape education through personalised, tutor-led learning supported by artificial intelligence.

Founded as an online language tutoring marketplace, Preply now connects more than 100,000 tutors with learners in 180 countries, offering one-to-one lessons across more than 90 languages. Its model blends human instruction with an AI-powered tutoring co-pilot designed to track progress, surface insights and automate administrative tasks, allowing tutors to focus on teaching and learners to progress faster.

The funding comes amid strong growth in global demand for language learning. According to data from HolonIQ, around 1.8 billion people worldwide are currently working towards proficiency in a second language. The global direct-to-consumer language learning market is forecast to reach $227 billion by 2035, following threefold growth over the past five years.

Since its Series C round, Preply has more than tripled the number of bookable tutors on its platform and expanded its offering by adding more than 40 new languages. Over the past year, the company has continued to improve profitability and has become EBITDA positive.

Alongside WestCap, the round was supported by existing investors including the European Bank for Reconstruction and Development and Horizon Capital. WestCap’s team, which has previously backed and scaled global marketplaces such as Airbnb and StubHub, will work closely with Preply as it enters its next phase of growth.

Kirill Bigai, co-founder and chief executive of Preply, said the investment underlined the company’s belief that the future of education lies at the intersection of human connection and AI.

“We feel extremely fortunate and deeply responsible for shaping how people will learn in the future,” Bigai said. “We connect learners with the world’s best tutors, amplified by AI, bringing learning efficiency to a level that was previously unreachable. This partnership will help us continue to innovate and create opportunities for people everywhere to progress in their lives.”

Preply said it would use the fresh capital to expand its product and engineering teams, deepen its AI-driven personalisation and accelerate global expansion to reach more learners and tutors. The company is positioning itself against fully automated learning platforms by doubling down on human-led instruction, using AI as an enabler rather than a replacement.

Research published in Preply’s 2025 Efficiency Study, conducted with LeanLab Education, found that 96 per cent of learners believe real conversations with human tutors are essential to their progress, while 97 per cent said such interactions were key to building confidence. The study also found learners progressed up to three times faster on Preply compared with average benchmarks, with one in three advancing a full CEFR level within 12 weeks.

Allen Mask, a partner at WestCap and former senior executive at Airbnb, who will join Preply’s board, said the platform was setting a new benchmark for personalised education at scale.

“Learners thrive when real human instruction is supported by technology,” he said. “Preply has the market-leading product, experienced leadership and vision to shape how people communicate globally.”

The Series D brings Preply’s total funding to around $299 million to date, reinforcing investor confidence in its growth trajectory. With a growing global footprint and a focus on measurable learning outcomes, the company says it is well positioned to cement its role as a category leader in human-led, AI-enabled education.

Read more:
Preply raises $150m at $1.2bn valuation to scale human-led, AI-enhanced learning

Labour has been urged to stop workers accessing their private pensions from the age of 55 in an effort to curb early retirement and tackle rising unemployment, according to a leading think tank.

The Resolution Foundation, which has close links to senior Labour figures, said current pension and tax rules encourage wealthier workers to leave the labour market years before state pension age, worsening labour shortages and weakening the public finances.

Under existing rules, savers can draw on their private pension from age 55, 11 years before the state pension age, which is due to rise from 66 to 67 this year. Up to a quarter of a private pension, capped at £268,275, can be taken tax-free from that point.

In a new intervention, the think tank said ministers should consider limiting access to private pension wealth before state pension age, either by raising the minimum access age or by reducing the amount that can be withdrawn tax-free.

“To reduce the financial incentives for people to retire early, the Government should consider limiting access to private pension wealth before the state pension age,” the foundation said. “This could be done either by raising the age at which tax-relieved private pension wealth can be accessed or by reducing the amount that is tax-free.”

The call comes amid signs of a weakening labour market. Figures from the Office for National Statistics show the UK unemployment rate has climbed to 5.1 per cent, up from a post-pandemic low of 3.6 per cent in 2022. The Resolution Foundation said the rise has been driven largely by people under 25 and over 50 leaving or failing to enter work.

Employment rates among so-called “prime-age” workers in the UK remain comparable with high-employment European economies such as Denmark, Germany and Norway, but Britain lags behind when it comes to keeping older workers in jobs.

Data cited by the think tank shows that by age 55, around a quarter of people in Britain are not in employment. That figure rises to more than a third by age 60 and over half by 64. At the current state pension age of 66, only 30 per cent remain in work.

Among people aged 50 to 65 who are not working, 41 per cent cite sickness or disability as the main reason, while 31 per cent say they are retired. A further 12 per cent are home-makers and 6 per cent are unemployed and actively seeking work.

The minimum age for accessing private pensions is already scheduled to rise to 57 from April 2028, a change the Resolution Foundation itself recommended in a post-pandemic report in 2023. The think tank now suggests further reform may be needed.

Nye Cominetti, an economist at the Resolution Foundation, said generous tax reliefs were distorting behaviour at the top end of the income scale.

“Our pensions and tax rules currently incentivise very wealthy people to retire early,” he said. “These generous tax breaks should be restricted. By doing so, the Government can boost both employment and the public finances.”

“The UK does reasonably well on its overall employment rate compared with other rich countries,” he added, “but trails the best-performing nations when it comes to the share of people over 50 in work. The Government should offer a mix of carrots and sticks to improve their job prospects.”

The foundation noted that UK unemployment is now close to the European Union average of 6 per cent for the first time since the euro was launched in 2002, suggesting the problem is largely domestic rather than driven by global conditions.

Some countries have already gone further. Denmark, often cited as a high-employment economy, has linked its state pension age to life expectancy, meaning workers must now wait until age 70 to receive payments. The UK state pension age is scheduled to rise to 68 by 2042, fuelling speculation that future governments could adopt a similar approach.

Despite incentives to retire later, the number of pensioners still working has been rising as cost-of-living pressures bite. More than 1.5 million people over the state pension age are now in employment, according to estimates from HM Revenue & Customs, based on the latest Survey of Personal Incomes. Around 1.56 million over-65s are on payrolls, a 12 per cent increase compared with 2020–21, while 562,000 pensioners were self-employed in 2024–25.

A Treasury spokesperson said the government remained focused on retirement security, pointing to its commitment to the triple lock, which it said was worth £470 a year to recipients of the new state pension.

“We have also launched a pensions commission to look at what more is required to ensure the pensions system is strong, fair and sustainable,” the spokesperson added.

Read more:
Labour urged to raise private pension access age to curb early retirement

UK businesses are being urged to tighten controls around their use of artificial intelligence in 2026, as legal experts warn that poorly governed AI systems are exposing companies to mounting legal, financial and reputational risks.

From unclear ownership of AI-generated content to data protection breaches and misleading outputs, advisers say many organisations have adopted AI tools faster than they have put safeguards in place, leaving them vulnerable as regulation accelerates.

Copyright and ownership disputes remain unresolved

One of the most pressing risks for businesses using generative AI is uncertainty around copyright and ownership of AI-generated outputs. Legal experts warn that AI tools can unintentionally reproduce copyrighted material, creating disputes over who owns, or is liable for, the content produced.

A high-profile example is the case of Getty Images versus Stability AI, which highlighted the legal grey areas surrounding AI training data. Getty alleged that its copyrighted images had been used without permission to train an image-generation model. While Getty’s main UK copyright claim did not succeed, the court found limited trademark infringement linked to early outputs that reproduced Getty watermarks, underlining the legal uncertainty businesses still face.

Lawyers say companies should carefully review the licensing terms of any AI tools they use, implement internal review processes to check outputs for potential infringement, and clearly define ownership rights in contracts. Commercial use of AI-generated content is particularly risky where training data sources are opaque.

AI ‘hallucinations’ pose serious business risks

Accuracy remains another major concern. Studies suggest that around 20 per cent of AI-generated outputs contain significant errors, including fabricated or outdated information. When relied upon for legal, financial or operational decisions, these so-called “hallucinations” can expose businesses to misrepresentation claims, regulatory penalties and reputational damage.

In March 2024, a chatbot powered by Microsoft was reported to have given incorrect legal guidance to business owners, including advice that could have led to breaches of employment law. Legal experts warn that similar errors could result in fines of up to €7.5 million (£6.5 million) for providing misleading information to regulators.

To mitigate the risk, businesses are advised never to treat AI as a final authority. Human verification should be mandatory for high-stakes decisions, with clear disclosure when content has been AI-generated.

Weak AI governance is a growing liability

Many organisations have rolled out AI tools without establishing internal governance frameworks, a gap advisers describe as a “ticking time bomb”. Without clear policies, employees may misuse AI systems, input sensitive data, or fail to recognise harmful or biased outputs, increasing the risk of data breaches and legal claims.

Legal specialists recommend introducing company-wide AI policies that define acceptable use, establish review protocols and assign accountability for decisions informed by AI. Treating AI as a regulated business tool rather than a productivity shortcut is increasingly seen as essential.

Data protection breaches carry heavy penalties

AI systems often process vast quantities of personal data, including customer and employee information. Using this data without proper consent, transparency or anonymisation can lead to serious breaches of data protection law, resulting in fines and loss of trust.

Businesses are being urged to minimise data collection, document lawful bases for processing, maintain clear consent records and ensure transparency around how AI systems handle personal information.

Regulation is evolving faster than many businesses expect

Perhaps the biggest challenge for 2026 is the pace at which AI regulation is changing. Governments are introducing new rules that can apply across jurisdictions and, in some cases, retrospectively to systems already in use.

Legal experts warn that companies failing to monitor regulatory developments or audit their AI systems regularly risk falling foul of the law even when acting in good faith. Flexible compliance strategies and ongoing legal oversight are increasingly seen as essential as AI moves from experimentation to core business infrastructure.

The message from advisers is clear: AI remains a powerful competitive tool, but in 2026 it also represents a growing legal exposure. Businesses that fail to embed governance, oversight and compliance into their AI strategy may find the technology creates more problems than it solves.

Read more:
Legal experts warn UK firms of rising AI risks in 2026 as regulation tightens

Apple is developing a wearable “AI pin” capable of listening to conversations and responding to spoken commands, as the iPhone maker looks to regain momentum in the fast-moving artificial intelligence race.

The disc-shaped device, which would be equipped with speakers, microphones and cameras, is at an early stage of development but could eventually go on sale later this decade. According to reports, the wearable would be designed to clip onto clothing or be worn as an accessory, allowing users to interact with AI without taking out their phone.

The project was first reported by The Information, which said Apple is exploring a standalone product or a companion device for future smart glasses. Internal estimates suggest Apple could manufacture up to 20 million units when the product eventually launches, though it is not expected before 2027.

The prototype device is understood to include three microphones, pointing to a strong focus on accurately capturing voice commands, as well as a physical button. Some AI wearables are activated manually, while others continuously monitor surrounding conversations, a distinction that could prove sensitive for a company that has long positioned privacy as a core brand value.

Any always-listening device would present a challenge for Apple, whose chief executive Tim Cook has repeatedly emphasised user privacy as a competitive advantage. Apple’s existing products, including the iPhone and Apple Watch, already use microphones that activate when users say “Hey Siri”, but a dedicated listening wearable could attract far greater scrutiny.

The move would also represent an attempt to revive Apple’s faltering AI ambitions. The company has repeatedly delayed a major upgrade to Siri intended to rival conversational chatbots such as ChatGPT and Gemini. Apple recently struck a deal with Google to integrate Gemini into Siri, and is expected to overhaul the voice assistant later this year so it can handle more fluid, chatbot-style conversations, according to Bloomberg.

Apple is not alone in exploring AI-first hardware. A growing number of technology companies are searching for a post-smartphone interface powered by artificial intelligence. Meta has embedded AI into its smart glasses, while Amazon has acquired Bee, a start-up behind a listening wristband designed to capture and summarise conversations.

Elsewhere, Sir Jony Ive has joined OpenAI to develop a new generation of AI-powered devices, which could be unveiled as soon as this year, putting additional pressure on Apple to respond.

Not all experiments in the category have succeeded. Last year, Humane, a company founded by former Apple employees, halted production of its own AI pin after weak sales and heavy criticism. The $700 (£521) device was widely panned for poor performance, and the start-up behind it was later sold.

Apple declined to comment on its plans. However, the reported work underscores how intensely Big Tech is competing to define the next major consumer device — and how artificial intelligence, rather than touchscreens, may sit at the centre of that future.

Read more:
Apple explores wearable ‘AI pin’ that could listen to conversations

The builder of the superyacht Bayesian, on which British technology entrepreneur Mike Lynch died, has launched a £400 million legal claim against his widow, alleging that the tragedy caused a catastrophic collapse in its sales.

The Italian Sea Group (TISG) has filed a €456 million (£399 million) lawsuit in a Sicilian court, claiming that the yacht’s crew and the vessel’s holding company were responsible for the sinking and the resulting damage to the company’s reputation and revenues.

Lynch, the former chief executive of Autonomy, died alongside his teenage daughter Hannah and five others when the £30 million superyacht capsized during a violent storm off the coast of Sicily in August 2024. His wife, Angela Bacares Lynch, survived the incident and is the legal owner of Revtom, the Isle of Man-registered company that owned the vessel.

According to the claim TISG alleges that the crew’s “incompetence and negligence” led to the vessel capsizing, and that the company has since lost hundreds of millions of euros in prospective yacht sales after being blamed for the disaster.

The lawsuit has been filed by TISG and GC Holding Company, owned by Italian yachting entrepreneur Giovanni Costantino, in a court in Termini Imerese, near the site of the sinking. It names Revtom, the yacht’s captain James Cutfield, and crew members Timothy Eaton and Matthew Griffiths as defendants.

The claim argues that the Bayesian was “unsinkable” but that the crew failed to shut hatches, heed weather warnings or lower the vessel’s keel, causing it to capsize in high winds. It further alleges that Revtom is legally liable for the actions of those operating the yacht.

The allegations contrast sharply with findings from the UK’s Marine Accident Investigation Branch, which last year said the superyacht had “vulnerabilities” that were not known to the crew. The vessel, built in 2008 under the Perini Navi brand, featured one of the tallest masts in the world.

A source close to the Lynch family strongly rejected TISG’s claims, describing the lawsuit as “desperate, opportunistic and in bad faith”.

“The UK investigation has raised serious and unresolved questions about the yacht’s design, stability and operating characteristics,” the source said. “This action appears designed to distract from those issues, but it will not prevent proper scrutiny of how the vessel was designed, approved and built.”

TISG claims the fallout from the sinking has been financially devastating. The company said it has been unable to sell a single Perini-branded yacht since the incident, with expressions of interest from brokers and potential buyers drying up entirely. It also claims the tragedy has triggered a slump in its share price and a collapse in the value of the Perini Navi brand, which TISG acquired out of bankruptcy in 2021.

Before the disaster, the group said it had planned to sell close to €1 billion worth of yachts by 2028.

Ms Bacares Lynch declined to comment on the lawsuit. TISG did not respond to requests for comment, and the captain and crew members named in the claim could not be reached. Italian prosecutors have confirmed that members of the crew are under criminal investigation.

Lynch founded Cambridge-based software firm Autonomy, which was sold to Hewlett-Packard for £7 billion in 2011. He was later charged with fraud by US prosecutors but was acquitted in 2024. He was celebrating his legal victory with family and friends on board the Bayesian at the time of the sinking.

Separately, Lynch’s estate is facing a major civil claim in the UK from Hewlett-Packard, which is seeking £1.5 billion in damages linked to the Autonomy acquisition. TISG also sued The New York Times last year following a report that raised questions about the yacht’s design.

Read more:
Mike Lynch superyacht builder sues widow for £400m over Bayesian sinking

Calling a colleague “old” because they struggle with computer skills does not, on its own, amount to age discrimination, an employment tribunal has ruled.

The decision came in the case of Farah Janjua, 39, who brought claims against her former employer, Harvey Jones Ltd, after a younger manager told her her lack of IT skills was “because you’re old”.

Ms Janjua argued that the comment, made by a colleague in his late 20s, amounted to unlawful age discrimination. She was dismissed from her role as a sales designer following the end of her probation period and subsequently launched legal proceedings.

However, an Employment Tribunal sitting in Reading rejected her claim in full, concluding that the remark did not meet the legal threshold for age discrimination.

The tribunal heard that Ms Janjua began working at a Harvey Jones kitchen showroom in Marlow in July 2022. During one incident, a sales manager, Nawaz Salauddin, intervened while she was working on a document, showing her how to add attachments using a computer mouse.

When Ms Janjua said she did not know the function existed, Mr Salauddin replied: “Cos you’re old.”

Ms Janjua complained about the remark, arguing it was ageist given that she was 39 at the time. She also alleged a separate incident in which a regional sales manager appeared “disgusted” on learning her age.

In dismissing the claim, Judge Naomi Shastri-Hurst said the tribunal accepted that the comment had been made, but found that it was not discriminatory in law.

“We find that a lack of technical knowledge is not infrequently deemed, rightly or wrongly, to be connected to age,” the judge said. “On the balance of probabilities, we accept that this conversation took place as suggested.”

However, she added that the evidence showed Mr Salauddin would have made the same comment to anyone older than him, rather than targeting Ms Janjua specifically because she was 39.

“In light of the evidence we have as to his character and behaviour, in terms of his desire to assert his authority, we find that he would have said this to anyone older than him,” the judge said.

Ms Janjua was dismissed in December 2022 following concerns about her performance. She launched legal action the following month, bringing claims of age discrimination alongside allegations of race and sex discrimination, sexual harassment, harassment related to sex, and victimisation.

All claims were rejected by the tribunal.

“We reject the claim of age discrimination in its entirety,” Judge Shastri-Hurst concluded.

The ruling highlights the distinction tribunals draw between inappropriate or ill-judged workplace comments and conduct that meets the legal definition of discrimination under employment law.

Read more:
Calling colleagues ‘old’ over IT skills is not age discrimination, tribunal rules

A bitter trademark dispute over the Beckham name has emerged as a key factor in the growing rift between the Beckhams and their son Brooklyn Beckham and his wife Nicola Peltz, underscoring how intellectual property law can collide with family relationships when a surname becomes a commercial asset.

The row centres on control of the Beckham name, which has long been protected as a trademark across multiple commercial categories. While the dispute has played out in the public eye as a deeply personal family fallout, legal experts say it is also a textbook example of how trademark ownership can override individual autonomy in business.

Hannah Finster, an intellectual property lawyer at Marks & Clerk, said the situation highlights a fundamental principle of trademark law.

“Trademark ownership trumps personal identity in commerce,” she said. “The Beckham family has strategically protected the ‘BECKHAM’ brand since 2000 across multiple classes of goods and services. This isn’t just family drama, it’s a clear example of trademark strategy colliding with personal autonomy.”

Finster said that, in legal terms, even being born with a famous surname does not automatically confer the right to commercialise it. “Brooklyn Beckham cannot simply monetise his own name without permission if the trademark is owned elsewhere,” she said.

She pointed to historic precedents, including Chelsea FC’s registration of José Mourinho’s name, which meant the manager himself could not authorise merchandise linked to his new club. “It’s a stark illustration of how trademark rights can sit above the individual,” Finster added.

The Beckham dispute also echoes earlier high-profile cases where founders lost control of their own names after selling businesses, such as Jo Malone and Bobbi Brown. In those cases, the personal brand became a corporate asset, limiting the founders’ future freedom to trade under their own identities.

Finster noted that trademark disputes often capture public attention in a way commercial law rarely does, precisely because they blur the line between family, identity and money. She pointed to popular culture, including the latest season of Emily in Paris, which features a storyline centred on family trademark infringement.

She added that many of the world’s most commercially successful celebrity families have moved early to avoid similar conflicts. Figures such as Beyoncé and Jay-Z, along with the Kardashians, have aggressively registered trademarks and secured social media handles for their children from a young age, often as a defensive measure to prevent third-party exploitation.

“The Beckham situation shows what happens when the desire for personal autonomy clashes with a brand that has already been locked down,” Finster said. “When your name is the asset, breaking free is not as simple as doing whatever you want.”

As the family fallout continues to attract headlines, legal specialists say the episode serves as a cautionary tale for celebrity families and entrepreneurs alike: without clear agreements on ownership and future use, even the most personal of assets — a family name — can become a source of conflict rather than legacy.

Read more:
The trademark dispute at the heart of the Beckham family feud

Nigel Farage has renewed his threat to strip commercial banks of billions of pounds in interest payments made by the Bank of England, reviving a controversial proposal that has already alarmed financial markets and the banking industry.

Speaking at a Bloomberg event on the fringes of the World Economic Forum, Farage confirmed that Reform UK still intends to scrap the interest paid on reserves held by banks at the central bank, money created during the era of quantitative easing.

Quizzed on whether the proposal, first outlined in Reform UK’s 2024 manifesto, remained party policy, Farage replied bluntly: “We are going to do it.”

“Some of the banks won’t like it,” he added. “Well, I don’t like the banks very much because they debanked me, didn’t they? This will be tough for banks to accept, but the drain on public finances is just too great.”

Farage rejected claims that the move amounted to a new tax on banks, insisting instead that lenders were benefiting unfairly from central bank policy. “They are just not going to get free money anymore,” he said.

Under current arrangements, commercial banks earn interest on the reserves they hold at the Bank of England, a mechanism designed to transmit monetary policy. Critics of the system argue it has resulted in large, politically sensitive transfers from the public purse to the banking sector as interest rates have risen.

At the time of the general election, banking groups warned that scrapping the payments could have “real consequences” for financial stability, lending conditions and investor confidence in the UK.

Asked why bond investors appeared uneasy about the prospect of a Reform UK government, Farage dismissed market concerns and suggested he would take a contrarian approach.

“As a former commodities trader, if there’s a big consensual view, take the opposite trade position,” he said. “I’ve always believed in that quite strongly.”

Farage also sought to draw lessons from the market turmoil that followed the 2022 mini-Budget under former prime minister Liz Truss and chancellor Kwasi Kwarteng.

“The big lesson is they did not propose to cut spending,” he said. “For our programme to work, we absolutely have to tell people that we are going to reduce welfare spending and cut excessive government spending. If we do that, I think the markets will applaud it.”

Farage also revealed that he had recently met Andrew Bailey, describing the encounter as “a very interesting clash of cultures”.

The renewed proposal is likely to reignite debate over the independence of the Bank of England and the relationship between monetary policy and fiscal decision-making, particularly as political scrutiny of bank profits and public finances intensifies ahead of the next phase of the economic cycle.

Read more:
Nigel Farage renews pledge to tax banks by scrapping Bank of England interest payments

A Leeds-based technology start-up has raised £23 million to accelerate the commercial rollout of its light-powered computer chips, in a move that could strengthen the UK’s position in the global race to develop next-generation artificial intelligence hardware.

Optalysys will use the funding to scale its photonic computing technology and expand operations into the United States. The investment round was led by Northern Gritstone, with backing from the UK government’s National Security Strategic Investment Fund.

Optalysys develops chips that process and transmit data using light rather than electrical currents flowing through copper wires, a technology known as photonics. The approach dramatically reduces energy consumption, generates far less heat and allows data to be processed securely while remaining encrypted.

Unlike conventional semiconductor firms, Optalysys designs its chips in-house but outsources manufacturing to specialist foundries. These include facilities at University of Southampton, as well as partners in Belgium and Singapore, with a further site planned in New York state.

Nick New, founder and chief executive of Optalysys, said building fabrication plants internally would be prohibitively expensive and unrealistic given the dominance of players such as TSMC.

“Trying to compete head-on with traditional semiconductor giants would be futile,” said New. “But in photonics, where the next generation of processing will come from, the costs are far lower. That gives the UK a genuine opportunity to become a global leader in photonic computing.”

The company’s immediate commercial focus is secure computing. Because photons allow data to be processed while in motion and encrypted, sensitive information can be analysed by AI systems without being exposed, a capability New said is increasingly vital across finance, defence and critical infrastructure.

“In financial services, for example, transactions can be processed confidentially, which makes this technology suitable for mainstream banking in a way traditional AI systems are not,” he said.

Optalysys is currently pre-revenue but launched its first commercial secure-computing product last year and is beginning to ship it to customers in the coming weeks. New believes secure computing will underpin the next phase of AI and cloud infrastructure.

“Where companies like Nvidia provided the foundations for today’s AI boom, we can do the same for the next generation of AI and secure computing,” he said.

The company is also targeting AI data centres, where tens of thousands of graphics processing units generate intense heat and consume vast amounts of electricity. As AI models such as ChatGPT proliferate, the industry is increasingly constrained by power availability.

Light-based connections do not generate heat in the same way as electrical currents, significantly reducing cooling requirements and overall energy demand. New said this efficiency advantage could prove decisive as data centres become some of the world’s largest consumers of electricity.

“Any meaningful reduction in energy use at that scale saves enormous amounts of money and resources,” he said.

Optalysys currently employs just over 50 people across Leeds, Bristol and the US, and plans to hire a further 35 staff as it scales. The company is also establishing a commercial base in San Francisco to deepen its presence in the US market and tap into American investment networks.

With geopolitical concerns growing around chip supply chains and AI infrastructure, investors believe photonic computing could offer the UK a rare chance to carve out global leadership in a strategically vital technology.

Read more:
Leeds startup raises £23m to push UK to the forefront of the AI chip race

Government borrowing fell sharply in December after a surge in income tax and national insurance receipts helped narrow the monthly deficit, offering short-term relief for the Treasury but little comfort over the longer-term health of the public finances.

Figures published by the Office for National Statistics showed public sector borrowing fell to £11.6 billion in December, £7.1 billion lower than the same month a year earlier, a 38 per cent reduction.

The improvement was driven by a strong rise in revenues. Total tax receipts increased by £7.7 billion year-on-year to £94 billion, while public spending rose by a more modest £3.2 billion to £92.9 billion.

Tom Davies, senior statistician at the ONS, said the fall in borrowing reflected “receipts being up strongly on last year, whereas spending is only modestly higher”.

However, the broader fiscal picture remains stretched. Over the first nine months of the financial year, borrowing totalled £140.4 billion, only £300 million lower than the same period last year and still the third-highest April-to-December borrowing figure since records began in 1993.

Public sector debt now stands at 95.5 per cent of GDP, up from 35 per cent before the 2008 financial crisis and 0.9 percentage points higher than a year ago. Higher interest rates continue to exert pressure, with £9.1 billion spent servicing government debt in December alone.

Dennis Tatarkov, senior economist at KPMG UK, said rising debt interest costs remained a structural challenge despite the month-on-month improvement.

The figures come amid continued economic uncertainty, following years of shocks including Brexit, the pandemic, energy price volatility after Russia’s invasion of Ukraine, and the fallout from Liz Truss’s 2022 mini-Budget.

Chancellor Rachel Reeves has raised taxes by around £70 billion across her first two Budgets, but the ONS data suggests this has yet to materially reduce borrowing over the course of the year.

James Murray, chief secretary to the Treasury, said the government was “stabilising the economy, reducing borrowing and ensuring public services deliver value for money”.

Independent commentators were less convinced. Philly Ponniah, chartered wealth manager at Philly Financial, said December’s improvement should not be mistaken for a turning point.

“The UK is still running one of the largest peacetime deficits on record,” he said. “High debt limits future choices and reduces resilience when the next shock arrives.”

With inflation still above target and growth fragile, economists warn that sustained improvement will depend less on tax windfalls and more on long-term productivity and investment-led growth.

Read more:
Government borrowing falls on income tax windfall, but pressure on public finances remains