DX Today AI Daily Brief - Monday, March 30, 2026

DX Today AI Daily Brief - Monday, March 30, 2026

Mistral AI secures eight hundred thirty million dollars in debt financing to build a massive Nvidia-powered data center near Paris as Europe pushes for A.I. compute sovereignty. Anthropic faces fallout after a data leak reveals its most powerful model yet, Claude Mythos, which the company warns poses unprecedented cybersecurity risks. Meta six-folds its El Paso Texas data center investment to ten billion dollars, targeting one gigawatt of capacity by twenty twenty-eight. Bluesky debuts Attie, an A.I. assistant that lets users build custom social media feeds with plain-language prompts. Apple hires veteran Google VP Lilian Rincon to lead A.I. product marketing for Apple Intelligence and Siri. Crusoe announces a nine hundred megawatt A.I. factory campus in Abilene Texas to support Microsoft infrastructure. A new pro-A.I. political group plans to spend over one hundred million dollars in the twenty twenty-six midterms. The European Commission confirms a cyberattack by ShinyHunters claiming three hundred fifty gigabytes of stolen data. University of Michigan researchers unveil Prima, an A.I. that diagnoses brain MRIs in seconds with ninety-seven point five percent accuracy. A CFO survey reveals A.I. driven layoffs could be nine times higher than last year. Yann LeCun's AMI Labs raises a record one point zero three billion dollar seed round to build world models. And the EU Council agrees to streamline A.I. regulation under its Omnibus Seven package.

DX Today - Your daily AI news briefing, delivered every morning.

Send us fan mail

#AINews #ArtificialIntelligence #AIInfrastructure #Cybersecurity #DXToday

Transcript

SPEAKER_03 0:00

It's Monday, March 30, 2026. You're listening to the DX Today AI Daily Brief. Today, Mistral secures$830 million to build a massive AI data center near Paris. Anthropic's most powerful model yet is revealed in an accidental data leak, and Meta sixfolds its investment in an El Paso data center to$10 billion. Let's get into it.

SPEAKER_00 0:27

French AI startup Mistral has locked in$830 million in debt financing from a consortium of seven global banks, including BNP Par, HSBC, and Credit Agricole. The funds will go toward building a 44 megawatt data center near Paris, powered by 13,800 NVIDIA GB, 300 GPUs. Operations are expected to begin in the second quarter of this year. Mistral says the facility will handle both model training and inference workloads. The company is also pursuing a broader plan to reach 200 megawatts of compute capacity across Europe by the end of 2027, part of a growing push for European AI sovereignty.

SPEAKER_05 1:12

From chips to leaks, Anthropic is dealing with the fallout from an accidental data leak that revealed the existence of its most powerful AI model to date, called Claude Mythos. Security researchers discovered an unsecured content management system containing nearly 3,000 internal files, including a draft blog post describing the model in detail. Anthropic says, Mythos represents a step change in capabilities, dramatically outperforming Claude Opus 4.6 on coding, academic reasoning, and cybersecurity benchmarks. But here's the concern. The company describes Mythos as currently far ahead of any other AI model in cyber capabilities, warning it could enable large-scale cyber attacks that outpace defenders. The model is being trialed by early access customers, but a general release remains uncertain.

SPEAKER_02 2:13

Now to massive infrastructure. Meta is dramatically scaling up its AI ambitions in Texas. The company announced it's increasing its investment in an El Paso data center to$10 billion, up from an original commitment of$1.5 billion when construction began last October. The facility aims to bring one gigawatt of capacity online by 2028. Meta says the expansion will create 300 permanent jobs and require more than 4,000 construction workers at peak. The company is also contracting more than 5,000 megawatts of clean energy projects in Texas to power the build-out. This is Meta's third data center site in the state, reflecting the broader hyperscaler race to secure compute capacity as AI workloads continue to surge. Shifting to social media.

SPEAKER_01 3:03

Blueski is making its AI play. The company unveiled ADI, a new app that lets users build custom social media feeds using plain language prompts, no coding required. Want a feed showing electronic music from people in your network? Just type it in. Addy runs on Anthropics Claude under the hood and was presented at the Atmosphere Conference by BlueSky's chief innovation officer Jay Graber and CTO Paul Frazy. At launch, Addy generates feeds viewable within BlueSky or any app built on the AT protocol. The longer-term vision is even more ambitious, letting users vibecode their own social apps. Addy is currently an invite-only closed beta with a public wait list open for signups.

SPEAKER_03 3:48

Over to Cupertino now.

SPEAKER_04 3:50

Apple is bringing in outside firepower for its AI push. The company has hired Lillian Rincone, a veteran Google vice president, to lead product marketing for Apple Intelligence and Siri. Rincon spent nine years at Google, most recently running the Global Product Organization for Google Shopping and Google Assistant. She'll report directly to Greg Joswiak, Apple's senior vice president of worldwide marketing. The hire comes as Apple prepares a long-delayed overhaul of Siri, its voice assistant that has struggled to keep pace with rivals powered by large language models. Rincon announced her departure from Google on LinkedIn last week, saying she was incredibly excited about what comes next.

SPEAKER_03 4:36

Staying with big tech.

SPEAKER_00 4:38

Microsoft is scaling its AI infrastructure to industrial proportions. Crusoe announced a new 900-megawatt AI factory campus in Abilene, Texas, built specifically to support Microsoft's compute needs. The expansion includes two new buildings, each delivering 336 megawatts of critical IT load, plus a dedicated on-site power plant for grid resilience. With this addition, Crusoe's total projected capacity at the Abilene site reaches 2.1 gigawatts. The campus features closed-loop liquid cooling systems designed for next generation GPU architectures. Land clearing is already underway, with the first building expected to come online by mid-2027. The project is expected to create thousands of construction jobs and hundreds of permanent positions.

SPEAKER_05 5:29

Now to Washington. AI is becoming a major force in American politics. A new group called Innovation Council Action is preparing to spend more than$100 million in the 2026 midterms to back candidates who support AI deregulation. The organization is led by Taylor Budowicz, a former Trump White House Deputy Chief of Staff, and is championed by tech investor and White House AAI advisor David Sachs. The group has developed a scorecard ranking lawmakers on their alignment with the administration's AI agenda. Combined with other pro-industry groups, total AI election spending now exceeds$300 million, marking a dramatic escalation in the tech sector's efforts to shape policy from the ballot box.

SPEAKER_02 6:33

The hacking group ShinyHunters has claimed responsibility, alleging the theft of more than 350 gigabytes of data from the Europa.eu platform. The stolen data reportedly includes emails and attachments, a full single sign-on user directory, DKIM signing keys, and internal admin URLs. The Commission says internal systems were not directly impacted and the website remained operational, but the full scope of the breach is still under review. This marks the second confirmed data breach at the Commission this year, following a separate intrusion reported by CERT EU in February. From security to healthcare.

SPEAKER_04 8:40

Executives believe AI is delivering larger productivity gains than the data actually shows, suggesting a delay between adoption and realized revenue. As one co-author put it, it's not the doomsday scenario the headlines might suggest.

SPEAKER_03 8:57

Two more stories to go.

SPEAKER_00 8:58

Yanlikun's new venture AMI Labs has raised$1.03 billion in what's believed to be the largest European seed round ever. The company, co-founded after Lacun departed Meta, is valued at$3.5 billion and is focused on building world models using joint embedding predictive architecture, or JeepA. Unlike large language models that predict text token by token, Jeepa learns abstract representations of how the world works. The round was backed by Bezos expeditions, NVIDIA, Toyota, and Samsung, among others. AMI Labs plans to build teams across Paris, New York, Montreal, and Singapore.

SPEAKER_03 9:37

And finally, regulation in Europe.

SPEAKER_05 9:40

The European Union Council has agreed on its negotiating position to streamline AI regulations as part of the Omnibus 7 legislative package. The move, announced on March 13th, is part of the EU's broader simplification agenda aimed at reducing regulatory burden while maintaining oversight. The Council's position will now be used in negotiations with the European Parliament. The effort comes as Europe tries to balance its role as a global leader in AI regulation, with growing pressure from the industry to reduce compliance costs and keep pace with the rapid development happening in the United States and China. It's a delicate balancing act that will define Europe's AI future.

SPEAKER_03 10:31

That's your briefing for Monday, March 30th, 2026. For DX Today, stay curious.