Welcome back to another episode of Upside at the EUVC Podcast, where Dan Bowyer, Mads Jensen of SuperSeed and Lomax Ward of Outsized Ventures gather for a holiday-home special to cut through the noise around Europe’s tech, geopolitics and AI shifts.
What begins as an innocent debate about whether DeepMind is “still a UK company” quickly spirals into a tour of sovereign AI strategy, the SpaceX mega-raise, Europe’s increasingly uncomfortable place between China and the US, defence-spending reality checks and a surprisingly uplifting set of deep-tech deals across the continent.
It is classic Upside: the takes are sharp, the geopolitics gets spiky, and the optimism… well, it arrives eventually.
What’s covered:
04:36 AI-for-Science, robotics and the new “AI scientist” era
06:50 A national-curriculum Gemini and the vision of a tutor for every child
09:39 The SpaceX 2026 IPO: what investors are actually buying
14:00 Starship, orbital compute and the trillion-dollar imagination gap
18:07 Why Europe missed the space race once again
19:43 Portugal flips the script: “Economy of the Year”
22:58 Europe between China’s export tsunami and America’s cold shoulder
32:07 Defence budgets: the hype, the delay and the reality for startups
34:25 AI Corner: bubble fears, Mistral’s comeback, Meta goes closed, China goes full-stack
✍️ Show Notes
DeepMind x UK Government: Sovereignty, Science and a New National AI Strategy
The episode opens with an almost comic question: is DeepMind even a UK company anymore? Lomax answers bluntly that, no, Alphabet owns it outright, but the group agrees that it’s not that simple. In a world where sovereignty suddenly matters again — especially in AI, defence and critical infrastructure — cap tables, ownership structures and national alignment all matter more than they did ten years ago.
Against this backdrop, DeepMind and the UK’s Department for Science, Innovation and Technology have launched a new partnership that will create a London research lab, deepen work with the AI Safety Institute and expand into AI-for-Science. Lomax puts it in context: the “AI scientist” thesis is finally real. Machine learning and autonomous robotics are now able to run experiments at a scale that no lab staffed purely by human researchers could match. Materials discovery, drug discovery, physics — all of it accelerates.
For Mads, the most transformative piece is education. The partnership includes a national-curriculum-grounded version of Gemini that could finally move us toward the holy grail: a personalised tutor for every child in the country. The team agrees that if implemented well, this is one of the most exciting and accessible applications of frontier AI so far.
SpaceX’s 2026 IPO: The World’s Most Powerful Monopoly Machine
From UK labs, the conversation rockets into orbit. SpaceX is reportedly preparing a 2026 IPO that may raise thirty billion dollars at a valuation of one-point-five trillion. Lomax breaks it down with unusual clarity: investors are effectively buying three companies at once.
Starlink is the cash engine, already at millions of users and expected to hit sixteen billion in revenue and roughly ten billion in EBITDA by 2026. The launch business remains a near-monopoly, and nothing in Europe is remotely close. And then there is Starship, which is the uncapped upside: a reusable super-heavy vehicle programme so far ahead of the pack that SpaceX may be the only organisation on earth capable of delivering deep-space missions at scale.
Mads adds a twist: the emerging thesis of orbital data centres. Space offers intense solar energy, free cooling and faster data transmission than fibre on earth. The US Department of Defense is experimenting. China is experimenting. And Musk already controls the biggest AI compute cluster in the world through X, plus the launch economics, plus the battery know-how via Tesla. It is wild, borderline speculative — and increasingly plausible.
Europe’s Space Reality Check
While SpaceX accelerates, Europe… does not. Lomax lists the painful facts: ISAR’s first launch failed, Rocket Factory Augsburg is behind, and Eutelsat’s constellation strategy is small compared to Starlink. Europe is fragmented, undercapitalised and late to a category that is structurally winner-takes-all. No one on the panel sees a short-term turnaround.
Portugal: The Comeback Kid of European Macroeconomics
There is at least some sunshine. The Economist has named Portugal the “Economy of the Year” across 36 wealthy nations, screening for inflation discipline, GDP growth, employment strength and stock-market performance. Portugal comes out on top with controlled inflation, over two percent GDP growth and a buoyant market. The irony is rich: a decade ago, the “PIGS” narrative dominated European financial commentary. Now southern Europe is outperforming the UK, Germany and the US on several metrics.
Europe Gets Punched From Both Sides: China and the US
The mood turns again as Mads lays out the squeeze Europe finds itself in. Washington’s new national security priorities increasingly deprioritise Western Europe. Trump’s team is openly exploring a fast settlement in Ukraine and warmer ties with Russia. Meanwhile China is expanding its export surplus at historic scale — over one trillion dollars in eleven months — even as tariffs bite.
China’s productivity and cost advantage is structural. It does not have the political constraint of a welfare state absorbing its economic surplus. Europe does. And Europe’s democratic structures, while admirable, make coordinated, long-term industrial strategy harder.
Elon Musk adds colour by railing against EU regulators after an X fine and praising Orban and Meloni. It all adds up to a difficult week for the continent.
Defence Spending: The Hype Meets the Procurement Cycle
Lomax provides a dose of realism. European defence budgets are not opening nearly as fast as political speeches suggest. The UK’s 2.5 percent target for 2027 is far away. The five percent target for 2035 is, in his words, largely meaningless. Startups expecting procurement windfalls in their next funding cycle may be disappointed. Defence is slow, bureaucratic and subject to multi-year political drift.
AI Corner: The Bubble Talk, the Open-Source Surprise, and China’s Silicon Awakening
AI Corner this week is unusually packed. Pension funds are trimming US exposure out of bubble concerns, but nobody is actually fleeing the market. Nvidia keeps running, institutions keep hedging, and 2026 still looks strong.
The real surprise comes from France. Mistral releases DevStral 2, an open-source coding model that performs astonishingly well on human evals while remaining lightweight and cost-efficient. Europe is officially back in the open-source game.
Meta moves in the opposite direction, shifting toward a closed model after the struggles around Llama 4. The reversal is striking: Meta, once the champion of open AI, is now leaning proprietary just as Europe leans open.
And then there is China. Following a political handshake facilitating Nvidia’s export permissions, China doubles down on building its own silicon ecosystem. Huawei continues rolling out Ascend chips. MoreThread completes a successful IPO. Full-stack autonomy is the strategy.
Deals of the Week
The headline deal is a wild one. Unconventional AI raises four hundred and seventy-five million dollars at seed, barely two months after founding. Their bet is neuromorphic computing — processors inspired by the brain that could be a thousand times more energy-efficient than GPUs. If they succeed, the entire AI compute landscape could shift.
Solve Intelligence raises forty million to reinvent IP workflows with large language models, a rare example of legal tech with real depth.
Relation Therapeutics signs a substantial deal with Novartis, underscoring London’s growing strength in AI-enabled drug discovery.
And the week closes with GoCardless effectively merging with Mollie in a share-for-share combination, creating a more muscular European payments platform.
One-Line Takeaway
Europe is being buffeted by forces far larger than its current policy machinery, yet the entrepreneurial engine keeps spinning up breakthrough science, frontier AI and globally competitive companies. The founders who thrive will be the ones building for the world as it is becoming, not the world Europe still wishes it were.
