Automattic has added write capabilities to WordPress.com’s MCP integration, giving AI agents like Claude and ChatGPT the ability to create posts, build pages, manage comments, and restructure content, all through natural conversation, with human approval at every step. For most of the past six months, connecting an AI agent to your WordPress.com site has meant […]

OpenAI, Anthropic, Alibaba and other AI model giants should pay a “levy” for selling their services in Europe with the money funnelled into Europe’s cultural sector, according to the boss of Europe’s leading AI model startup.  

Arthur Mensch, the CEO and co-founder of Mistral, says such a levy would be a win-win for both the AI model providers, giving them legal certainty, and also creators, whose data the AI models are trained on, which would benefit from the levy.  

AI model companies, which train their models on vast amounts of text, audio and video data, have been hit by complaints and legal challenges from creators and copyright owners.  

In an op-ed in the Financial Times, Mensch writes that European AI model developers were at a disadvantage, operating under a “fragmented legal environment”, compared to AI developers in the US and China, who were “developing their models under permissive or non-existent copyright rules”.

He says that the current copyright rules for European creators were also not working.

Mensch says a “new approach” is needed, with his solution being a revenue-based levy, "reflecting their use of content publicly avaiable online", which would be levied on AI model companies which operate in Europe.

“Proceeds would flow into a central European fund dedicated to investing in new content creation and supporting Europe’s cultural sectors”, he says.

According to one report, Mehsch is calling for a contribution between one and five per cent of the revenues of AI model providers in Europe. This was not confirmed by Mistral.  

Mensh writes: "At Mistral, we are proposing a revenue-based levy that would be applied to all commercial providers placing AI models on the market or putting them into service in Europe, reflecting their use of content publicly available online.  

"Crucially, this levy would apply equally to providers based abroad, creating a level playing field within the European market and ensuring that foreign AI companies also contribute when they operate here.   

"The proceeds would flow into a central European fund dedicated to investing in new content creation, and supporting Europe’s cultural sectors.  In return, AI developers would gain what they urgently need: legal certainty. 

 "The mechanism would shield AI providers from liability for training on materials accessible online. Importantly, it would not replace licensing agreements or the freedom to contract.   

"On the contrary, licensing opportunities should continue to develop and expand for usage beyond training. The fund would complement, not crowd out, direct relationships between creators and AI companies.”  

Under the EU’s current rules, AI companies can use copyrighted materials for text and data mining, including AI training, unless a creator has “reserved their rights".    The EU is looking into a permanent solution to protect copyright from use by AI.

Big news this week as the European Commission presented its proposal for EU Inc., a new single set of corporate rules that will serve as the cornerstone and starting point for the EU's 28th regime.

As a journalist, well, I’m pleased and frustrated all in equal parts. Earlier this week i called it a win because I think it would be a mistake for anyone to dismiss the challenge it took to get here. The announcement represents a major milestone for EU–INC, a policy movement backed by over 22,000 signatories, including Europe's leading founders, investors, and the broader startup community.  European bureaucracy moves at a glacial pace, and the journey here has been long and complex. 

There are numerous wins — For example, entrepreneurs, founders, and companies will be able to found an EU Inc. company within 48 hours, for less than €100, with no minimum share capital requirement. There will be EU-wide employee stock option plans and increased digitisation.

BUT there are a number of areas that sound like ideas rather than actioned items (dare I say the gap between aim and execution?), and worse, fundamental elements which stray from the original idea.  Instead of establishing a truly unified legal framework, the proposal defers interpretation to national courts—risking 27 divergent outcomes—relies on national registries rather than creating real ecosystem standardisation, and ultimately introduces further fragmentation at the very layer that should be harmonised. 

Further, the call on Member States to only consider establishing specialised judicial chambers or courts to handle disputes related to EU Inc. company law is notably vague in both scope and implementation.

I’m hopeful for better, but let’s watch this space.

But that aside, this week, we tracked more than 85 tech funding deals worth over €4 billion and over 5 exits, M&A transactions, rumours, and related news stories across Europe.

Alongside the week’s top funding rounds, we’ve highlighted key industry developments, as well as notable trends in European venture activity, investor moves and emerging sectors shaping the current funding landscape.

If email is more your thing, you can always subscribe to our newsletter and receive a more robust version of this round-up delivered to your inbox.

Either way, let's get you up to speed.

? Notable and big funding rounds

??  Upvest raises $125M to strengthen its API-based investment platform

?? WorkFlex secures €37M to automate cross-border workforce compliance

?. Sequoia-backed Edra raises $30M Series A to turn enterprise data into self-improving AI agents

??‍?? Noteworthy acquisitions and mergers

?? Nscale snaps up major US data centre site, American Intelligence & Power Corporation, inks AI compute deal with Microsoft

?? Amazon acquires Zurich-based Rivr, developer of stair-climbing robots for doorstep delivery

?? TMA acquires Amsterdam’s BrainsFirst to combine psychometrics with neuroscience-based talent insights

? Interesting moves from investors

? Partech’s €300M Impact Fund targets Europe’s next generation of industrial and climate tech leaders

?.Montis VC reaches €50M first close to back energy and industrial tech startups

?. Albion Venture Capital Trusts close £90M top-up offer as demand for UK innovation investments grows

? New €70 million GVC Gaesco fund targets InfraTech startups focused on energy, industry and digital infrastructure

?️ In other (important) news

?  UK government pledges £1BN quantum computing investment

??. EU Inc. marks major win for startups as Commission unveils 28th regime proposal

? Zopa reports third consecutive year of profit, says new current account topping expectations

?? Alpine Eagle scales Sentinel production with new Munich facility and European expansion

<!--td {border: 1px solid #cccccc;}br {mso-data-placement:same-cell;}--egora and Tandem Health CEOs reject Anthropic and OpenAI threat

? Recommended reads and listens

?. Meet Rachel: the AI agent that phoned 3,000 pubs to price a pint

?  Europe’s filmtech ecosystem is growing fast — here are the startups to watch

??. eYou raises €300,000 to build a European social media platform focused on trust

??. From diversified funding to frontier innovation: the Spanish tech ecosystem

? European tech startups to watch 

??. CiaoDott raises €1.5M pre-seed to bring vertical voice AI to Italy’s medical sector

??  Cleavr raises €1M to develop an AI solution for accounts receivable

?? Rhonexum secures $1M to scale cryogenic electronics for quantum computing

?? First Concepts raises $1M to develop AI-native OS for creative work

?? Noru raises €560,000 to develop an agentic compliance platform

The UK challenger bank is rolling out Starling Assistant to personal account holders today, billing it as the UK’s first agentic AI financial assistant. It can set up savings goals, organise bill payments, and even quiz you on your own spending, all from a voice or text prompt. Starling Bank has been building methodically towards […]

Over Paddy's weekend 2026, a friendly Northern Irish "woman" called Rachel rang more than 3,000 pubs across all 32 counties to find out the price of a pint of Guinness. Over 1,000 gave a price.

 Only a handful seemed to realise Rachel was an AI voice agent. 

A data gap 14 years in the making

Ireland’s Central Statistics Office stopped tracking pint prices 14 years ago, leaving no comprehensive dataset since.

In response, Matt Cortland created the “Guinndex,” using AI to rapidly collect and analyse pint prices across Ireland, where costs have become increasingly inconsistent.

The result is what he claims is the most complete index of pint prices to date—an effort to bring transparency and potentially help normalise the price of a Guinness.

From pub owner to AI engineer

Matt Cortland is an American AI engineer based in London who previously lived in Ireland for several years. A former US-Ireland Alliance Scholar (George Mitchell Scholarship), he holds a Master’s degree in Creative Digital Media from TU Dublin (formerly DIT). Earlier in his career, he founded and operated a global pub and entertainment company spanning Ireland, the UK, and the US, including a chain of IoT-enabled “wizard bars” where he created working magic wands.  He has since transitioned into AI engineering and private consulting, developing his own AI projects while building AI products and tooling for companies, several of which are based in Dublin.

"I'm a former pub and bar owner, so I know what it's like to be on the other end of customer pricing calls," said Cortland.

"But I also know what it's like to be on the consumer end and paying a kidney for a pint. I apologise to everyone I tortured over Paddy's weekend. Rachel just wanted a wee drink."

Designing a believable pub caller

Cortland named the AI agent “Rachel” and trained her to be friendly, direct, and—if questioned—transparent. She explains she’s “putting together a wee price comparison list” and confirms she is an AI if asked.

The calls were made through ElevenLabs’ conversational AI platform and Twilio for telephony, using an old Irish SIM.  The phone numbers come from the Google Maps API, publicly available numbers for each venue.  Cortland indexed over 5,200 pubs across all 32 counties using Google's Places API. 

He explained; 

“Over Paddy's weekend, Rachel called the 3,000+ of those that had phone numbers listed. 2,052 answered the phone, and over 1,000 gave a verified price, which I then extracted from the call transcripts using Claude AI. 

The whole thing cost about €200 to run plus a lot of my time.”

Getting the voice right proved critical, with much of the work focused on refining Rachel’s accent, tone, and personality. Cortland tested dozens of options before landing on a Northern Irish accent, inspired by Rachel Duffy from The Traitors, which he felt sounded the most natural in conversation.

“Duffy was the first female traitor to win the show, and she was just so good,” he said.

“She played an absolute blinder. That’s what I wanted — someone warm, someone you’d believe. A Northern Irish accent that makes ‘we were lookin’ to come in for a wee drink’ sound completely natural.”

The script itself went through multiple iterations. Early versions had Rachel confirming the price back—“Grand, so that’s six seventy, is that right?”—but this extended calls and gave people time to grow suspicious. “The final version keeps it simple: ask the question, say ‘thanks very much,’ hang up. The transcript captures everything,”

Cortland says the biggest challenge wasn’t technical infrastructure, but making the agent feel human—particularly in an Irish context.

“Funnily enough, for the Irish market, it was training her to have banter, which she really struggled with,” he said. “Should have made her American.”

How did the staff react? 

Few people Rachel spoke to realised they were talking to an AI agent. The calls produced dozens of memorable exchanges. When questioned, Rachel told them truthfully that she was putting together a wee price comparison list. Most people accepted that and moved on.

• At Malzard's Pub in Kilkenny, the bartender laughed and offered to buy the round: "They're normally 6.20, but if you can't afford one, we'll buy you one. We'll look after you."
• At Doogies in Northern Ireland, the bartender opened with: "Twenty-five pound. But if you're coming in for a wee drink, I'll give it to you for a fiver."
• At McIntyre's Bar in Donegal, the bartender launched a full interrogation: "Five eighty. What time is it? How many are coming? Where are you coming from? What part of the country are you from? Who's this I'm speaking to?"
• At Drumlane Bar in Cavan, the person who answered was in the kitchen: "I'm in the kitchen. I don't work in the bar at all." He went behind the bar to check. Five sixty.
• At Beaufort Bar in Kerry, the bartender played coy: "It's far too cheap. You'll have to come in to find out the price." Rachel persisted. Five sixty.
• At Buddy's Bar in Tipperary, Rachel asked the price. The bartender asked her name. Rachel said she was just putting together a wee price comparison list. "Fuck off." Fair.
• At The Plough in Curraglass, Cork, the bartender refused to give the price no matter how many times Rachel asked: "You'd have to call in and I'll tell you." Rachel said she couldn't do that. "Ah, well done. You'll never know, though, will you?"
• At The Linen House in Lisburn, Rachel got trapped in a Premier Inn phone system. Two AI systems talked past each other, neither able to help the other. Rachel said "Oh, dear" four times. The virtual receptionist kept apologising. Nobody got a pint price.

Pat Hayes, owner of The Arch Bar in Thurles, Tipperary, was one of the thousands of people who picked up Rachel's call over the weekend. When he later found out he'd been chatting to an AI, he took it in good spirits.

"It was a good laugh. I had no idea it wasn't a real person," said Hayes. "But look, knowing the price of a pint is important. 

People want to know what they're paying before they walk in the door. If someone's putting together an index of every pub in the country, fair play to them. It's good for the customer and it keeps us all honest."

AI and the pub

Beyond the novelty of calling pubs, the project also highlights where AI is beginning to have real-world impact.

The Guinndex arrives at an interesting moment. Earlier this month, Anthropic published research showing that bartenders, cooks, and dishwashers are among the 30 percent of occupations with zero exposure to AI automation. Computer programmers, by contrast, are the most exposed at 74.5 per cent. "AI isn't coming for the person behind the bar," said Cortland.

 "As a former bar owner, I know this.  AI can't (yet) pour a pint, it can't read a room, it can't tell when someone's had enough. But it can call pubs on a weekend and tell you where to find a decent pint for under a fiver.

The physical work is safe. The information layer on top of it is where AI lives and where we all should be aware."

 How much is a pint of Guinness these days, anyway?

The national average price of a pint of Guinness is €5.95. The most common price is €5.50. But where you drink matters enormously.

Dublin is the dearest county by a wide margin, averaging €6.75 a pint. The cheapest pints are in the west and midlands, with Laois at just €5.38. The gap between Dublin and the cheapest county is €1.37 per pint. The cheapest pint in the entire index is €3.00 at Glynn's Bar in Dunmore, Galway, although he may have just been taking the mickey. The most expensive is €10 at The Auld Dubliner, Dublin, which, incredibly, seems to be accurate.

Despite the rising cost of a night out, the Guinndex unearthed 12 places across Ireland where you can still get a pint for a fiver or less, including one in Dublin.

Dublin doesn't fare well on any measure. Of the 46 pubs in Ireland with a perfect 5.0 Google rating, not a single one is in Dublin. They're in places like Augher (Tyrone), Kilmakilloge (Kerry), and Rathdowney (Laois).

The CSO tracked pint prices from 2001 to 2011, then stopped. In the 14 years since, the price of a stout has jumped from €3.93 to €5.95 (+48 per cent). The Guinndex fills the gap.

Not a stunt, but a dataset: the mission to map — and lower — the price of a pint

Cortland asserts that Guinndex is definitely a data project rather than an effort to test the vulnerability of humans in identifying AI:

“I want to see if we can collectively drive down the cost of a pint across Ireland. Rachel unearthed the first 1,000+ pint prices, but there are still over 4,000 to find.

Those will need to come from people who log the cost using the "Contribute" button at guinndex.ai, which instantly updates the index and prices on the site.”

He believes that if enough people get involved, he can map the country and stabilise the price.  “Or at the very least, unearth some really good gems and places where you can still get a pint for a fiver or less (spoiler, there are 19).”

Cortland admits he’s a “real AI and data nerd” and thought this would be both a practically useful and really fun project. 

“Honestly, the main goal of this is to help make a pint of Guinness affordable again. If enough people contribute prices, we can map the whole country and start to bring down the price to something reasonable.  

I will say that this is part of something bigger I'm working on, but that's for another day. For now, go contribute to your local.”

Search the Guinndex and help it grow

The project is now evolving into a crowdsourced platform, encouraging people across Ireland to contribute local pint prices.  The full dataset is live at guinndex.ai. Search by county, town, or pub name to find the price of a pint near you.

The site includes an interactive map, county-by-county breakdowns, and the ability to compare prices across the country.

Anyone can hit the “Contribute” button on the site to submit the price of their last pint, flag a correction if a price is wrong, or share photos of pints and pubs for the Guinndex social media channels. Pub owners who want to update their listing can also message to make an amendment with the “Contribute” button on the site.

Cortland concludes:

" If you're sitting there with a pint right now, tell us what you paid."

Starling Bank is rolling out what it says is the “UK’s first agentic AI financial assistant”, as it looks to leverage the new technology to help improve day-to-day banking.

The UK challenger bank, which has nearly five million customers, said that Starling Assistant can help its customers manage day-to-day finances, share personalised financial insights and give general banking guidance.

The assistant responds to voice and natural language prompts before carrying out banking tasks on the customer’s behalf, from setting up personalised saving goals to organising bill payments.  

Examples of what it can do include if a customer is planning a holiday, they could say: "I need to save £500 for a trip to Paris in July. How much do I need to save monthly and can you set up automatic transfers to a dedicated space?"

Or if a customer wants to perfect their payday routine, they could say: "Set me up with dedicated spaces for my groceries, bills, travel and eating out" and then specify how much to transfer to each space on pay day.    

Harriet Rees, Starling’s group chief information officer, said: “It’s time to embrace a new era of banking, one that’s powered by agentic AI. 

"At Starling, we want to encourage our customers to trust that AI can help them with money management and we’re excited to be pioneering the use of this cutting-edge technology to help people be good with money.”   

The assistant is built on Starling’s proprietary tech platform using Google Gemini and Google Cloud technologies.  

Previously, Starling has launched several generative AI tools, including Spending Intelligence, which lets customers ask natural language questions about their spending habits, while Scam Intelligence helps detect online marketplace scams.  

Many fintechs and neobanks are using or experimenting with generative AI tech. Across Europe, Klarna uses it for customer service purposes while Bunq launched its AI assistant in 2024. Danish challenger Lunar says its GenAI-powered voice assistant will handle around 75 per cent of customer calls over time.  

Meanwhile, Revolut is exploring a push into the AI agent space, aiming to use the technology to automate everything from customer service to sales.

Fourteen months after the Digital Operational Resilience Act became enforceable, Europe’s financial institutions are running out of room to improvise. The regulation, which took effect on January 17, 2025, was supposed to mark the beginning of a new era in digital risk management across the EU. Instead, it has exposed just how far most firms […]

The Rijen-based startup, which retrofits existing sprayers with nozzle-by-nozzle PWM control, will use the capital to commercialise its LeapEye camera system and scale LeapBox internationally from Europe to Canada. The idea behind BBLeap is disarmingly simple: most agricultural sprayers treat an entire field as a single unit, applying the same dose of pesticide, herbicide, or […]

The San Francisco B2B sales platform, which recently approached $200M in ARR and appointed a new CEO, absorbs the revenue intelligence startup’s signal-layer technology to deepen its enterprise push. Apollo.io has acquired Pocus, a revenue intelligence startup that helps sales teams identify and prioritise the accounts most likely to buy based on behavioural and CRM […]

Montis VC has launched a new venture capital fund, securing €50 million at first close. The fund is backed by the European Investment Fund through the REPowerEU programme, the Polish Development Fund, and a group of family offices and private investors from across Central and Eastern Europe.

The fund will focus on early-stage startups at the pre-seed and seed stages, targeting around 20–25 companies developing technologies across energy, industrial transformation, and artificial intelligence. Its strategy reflects broader shifts in the European economy, where demand is growing for solutions that improve productivity, support industrial innovation, and contribute to the energy transition.

Montis VC aims to invest in scalable companies with the potential to compete globally, particularly in areas where AI can accelerate development and enhance real-world impact.

We invest in founders who want to build the future of the European economy through technology and energy efficiency, particularly in areas such as the energy and industrial transition, where AI can drive innovation and scale,

said Michał Gawęda, Partner at Montis VC.

The fund builds on the track record of the team behind Montis Capital, which has invested in companies operating at the intersection of technology and industry since 2019. With the new fund, Montis VC plans to increase both the size of its investments and its level of involvement, with initial tickets ranging from €0.5 million to €2 million and additional capital reserved for follow-on rounds.

Led by an international team with experience in venture investing, industry, and entrepreneurship, Montis VC is supported by a network of venture partners who will contribute to deal sourcing and help portfolio companies expand into global markets.

The fund will take an active approach in working with founders, supporting strategy, international expansion, and preparation for future funding rounds, as it continues its fundraising process.

The AI search company launches a suite of health data connectors, linking Apple Health, wearables, and electronic health records, making it the second major AI platform to integrate with Apple Health after OpenAI. Consumer health AI has become the year’s fastest-moving product category, and on Thursday Perplexity entered the race properly. The company launched Perplexity […]

Every product team has a roadmap. Every marketing team has a funnel. But ask most SaaS and ecommerce leaders which single component has the greatest direct impact on their revenue, and you will hear a surprising amount of hesitation. The answer, increasingly, is the one piece of infrastructure that still gets treated as an afterthought: […]

The Warsaw-based fund, backed by the European Investment Fund and the Polish Development Fund, plans 20–25 pre-seed and seed investments in companies working on energy transition, industrial automation, and AI. The team behind Warsaw-based Montis Capital has raised €50 million at first close for a new fund, Montis VC, targeting European startups at the intersection […]

Ten days after founder Jay Graber stepped aside as CEO, the decentralised social platform has disclosed a $100 million Series B led by Bain Capital Crypto, a round that closed last April but was never announced. The timing tells its own story. There is a quiet irony in the fact that the person who built […]

Here is one way the AI data economy works in practice in 2026: a DoorDash courier straps on a body camera, washes at least five dishes, holds each one up to the lens for a few seconds, and earns a few dollars. That footage, mundane, specific, reproducible at scale, is exactly what AI and robotics […]

The second version of Microsoft’s in-house image model lands at #3 on Arena.ai’s leaderboard, behind only Google and OpenAI, and begins rolling out across Copilot and Bing Image Creator today. A year ago, Microsoft was generating images for Bing and Copilot almost entirely with OpenAI’s models. On Thursday, the company’s in-house team announced MAI-Image-2, a […]