🧠 At RAISE Summit in Paris, the AI debate moved beyond whether the technology works to whether the world can build enough power, chips, data centers, and networks to support it. Industry leaders unveiled massive compute investments, tangible gains in enterprise productivity, and a new race to make AI faster and more efficient. But amid the optimism, stark warnings emerged about Europe’s infrastructure deficit and autonomous agents already outrunning corporate security controls. Read the recap —>

Chris O'Brien + Helen O'Reilly-Durand

Headlines

🗞️ From the department of weird and wacky came a story from Sifted involving Yann LeCun that took multiple twists and turns in the course of just 24 hours. And I'm still not sure anyone really knows what happened. First, Sifted appeared to have a big scoop: LeCun has a big ole new fund called Extelligence Invest! But then...nope? Sifted reported a few hours later that it was shut down and had separated from LeCun over concerns that he had other obligations to other funds. During the day, the firm's website went offline, as if the firm had indeed gone dark. Later in the evening, it was back, with only 2 of the 4 partners who had been on site earlier in the day still there: Tony F Chan and Amy Hu. Even later, their names had also been scrubbed. A capture from the Internet Archive in December 2025 showed 5 partners, but LeCun was not listed. The site also claimed the fund is registered with the SEC and even linked to the SEC site, but the SEC site had no record of its registration. By late evening, those links had been removed, and the address now lists the HQ as: 92 avenue des Champs-ÉlyséesParis, Île-de-France 75008. Which is a WeWork. | Sifted

🗞️ France’s Cheap Energy Becomes the New AI Battleground
France’s low-cost nuclear electricity could be its secret weapon in the AI race. But Mistral CEO Arthur Mensch is warning that US tech giants may plug in first. With politicians now debating preferential access to power for European companies, electricity is fast becoming the next front in France’s AI sovereignty battle. | Politico

🗞️ France Backs Global Push to Protect Children in the AI Age
France, Spain, and Kenya have launched a new global coalition to protect children in the age of AI. Bringing together 18 countries and UN agencies including UNICEF and UNESCO, the initiative will focus on building AI systems that are safe and respect children’s rights.| France Diplomatie

🗞️ Paris Peace Forum Launches AI Cyber Threat Hub
The Paris Peace Forum has launched INTAiC, a new global platform tackling the malicious use of AI in cyberspace. Bringing together AI and cybersecurity experts from the public and private sectors, research, and civil society, the initiative aims to build a shared evidence base on how AI is reshaping cyber threats.| Paris Peace Forum

🗞️ Strasbourg Consortium Joins Europe’s AI Gigafactory Race
A consortium of European tech and industrial players is bidding to build two AI Gigafactory campuses near Strasbourg. Led by ÆTHER and including 2CRSi, SiPearl, Axelera AI and Outscale, the project aims to build a more sovereign European AI infrastructure spanning chips, servers, cloud and energy.| L’Usine Digitale

🗞️ French Robotics Firm Passes Key Test for Future Combat Unit
French robotics firm Vigilant Solution has completed a new round of trials for Pendragon, France’s program to build its first robotic combat unit. The tests validated autonomous multi-robot navigation and the integration of ground robots and tethered drones, as France targets a first experimental unit by summer 2027. | The Defence Blog

🗞️ Mistral Moves Into Robotics
Mistral AI has launched Robostral Navigate, its first robotics model, as the Paris AI champion expands into factories, warehouses and industrial automation. The system enables robots to navigate using a single camera, without lidar, advanced sensors or multiple-camera setups.| Reuters

🗞️ Yann LeCun Bets $1bn That LLMs Have Hit a Wall
Yann LeCun is betting that larger chatbots won’t achieve human-level intelligence. Speaking at the RAISE Summit in Paris, the former Meta AI chief argued that a four-year-old learns more about the physical world through vision than the largest LLM can absorb from all the text online. His new startup, AMI, is building “world models” designed to learn by watching. | TWN

🗞️ Cerebras Bets Billions on European AI Compute
US chip startup Cerebras is planning a multibillion-dollar expansion across France and the Nordics, targeting 200MW of European AI compute capacity by the end of 2027. The Nvidia challenger says demand is surging for high-performance AI infrastructure located closer to European users. | Euronews/AFP

🗞️ France’s Nvidia Antitrust Probe Nears Its End
France’s competition watchdog says its investigation into Nvidia over alleged anti-competitive practices is nearing its end. The next step is to decide whether there are sufficient grounds to proceed with formal objections - or close the case.| Reuters

🗞️ ZML Takes Aim at AI Hardware Lock-In
Paris startup ZML has launched LLMD, a free AI inference server designed to run models at high speed across different chips and hardware architectures. The goal is to give companies more flexibility over their AI infrastructure - and potentially cut compute costs and energy use. | TechCrunch

🗣️Responsible AI🗣️

A look at the big stories this past week, from Responsible & Frugal AI expert James Martin, of BetterTech.

With AI data centers’ resource demands growing far faster than expected, are authorities moving faster than usual to curb them? 

No sooner had we discovered Microsoft's emissions jumped 25% in 2025 - meaning all three hyperscalers’ emissions have risen by over 40% over the past five years - than it emerged ten gas generators have been fast-tracked at the Stargate Abilene facility, potentially leading to what WIRED called “catastrophic” pollution levels. Just part of an increasingly unsustainable trend, affirmed AP: “The AI boom has set in motion the biggest ever construction boom of natural gas-fired power plants.”

Not to mention the WSJ’s 'shocker': AI uses *way* more water than big tech admits.

How are authorities reacting? 19 of 50 US states have currently enacted or are considering data center moratoria, according to this tracker. New York State became the first, on July 14, to fully implement a moratorium, Reuters reported.

New Jersey Governor Mikie Sherrillo has signed into law a “first in the nation” plan to ensure new data center costs do not impact ratepayers; that operators report their impacts; that community benefit agreements are established; and that job creation is supported. In Europe, in May, Denmark imposed a 3-month pause on data center development after hyperscalers flocked to this clean-energy, cool-climate country.

Back in the boardroom, 29% of CEOs “report difficulty understanding and controlling operating costs as [AI] systems scale”, as KPMG diplomatically puts it (i.e., no one told them why their AI bills are soaring, x10 in some cases). “Organizations need clear rules for who owns AI-related costs,” concludes the bean counter. As obvious as this may seem, only 25% of US workers are aware their company even *has* an AI policy or rules re. Gallup…

Fortunately, some are working to counter big AI’s apparent inevitability: Current IA, a “public interest AI” non-profit first announced at Paris’ AI Action Summit, has revealed new responsibility-focused services. They include Alpha Chat, a transparent, sovereign, and non-sycophantic LLM, and the AI Gap Map, which shows what open-source AI can and can’t do today. Highlights on the roadmap include a mobile app, on-device AI and “affordable sovereign inference.”

“If AI is this transformative tech, it has to be open, democratic and available for everyone to invent with… [and yet] the current system is closed and proprietary”, CEO Aya Bdier told the UN’s AI for Good conference. 

Time for change?

🧠 RAISE Summit 2026: Goodbye AI Demos and Vibes, Hello Buildouts and Megawatts

Despite the general hand-wringing about the ROI of AI or runaway token budgets, nobody gathered at the 3rd edition of Raise Summit in Paris last week seemed to be asking whether AI works.

The 9,000 people who packed into the Carrousel du Louvre on July 8 and 9 had moved on to a grubbier, more practical question: can anyone actually build enough power, chips, data centers, and networks to keep up with demand?

The event placed a heavy focus on infrastructure, both as an opportunity and a barrier to realizing AI's potential. For this crowd of insiders, the clear consensus was that AI had moved from demos and vibes to megawatts and buildouts.

President Emmanuel Macron opened the event with a video address pitching France as an AI leader and pressing the case that Europe needs its own AI infrastructure in an age of geopolitical division. From there, the agenda could have almost doubled as a construction roadmap: energy, chips, data centers, networking, storage, inference, security, and only then the applications sitting on top.

"AI is the steel of our generation," said Grant Lee, CEO of AI design startup Gamma. "The question is who turns this general-purpose technology into purpose-built technology."

👋🏻 If you’re enjoying The French Tech Journal, support the project by forwarding it to friends and sharing it on your social networks. You can also comment on this post. And if you have ideas for stories, tips, or just want to harass us, shoot us an email: chris@frenchtechjournal.com / helen@frenchtechjournal.com 👋🏻‌