Welcome to the latest edition of the World of DaaS Roundup.
A lot has happened since our last edition.👀
Humanoid robots are cleaning San Francisco homes and the data they generate may be worth more than the service itself. OpenAI, Anthropic, and Google are quietly working through what AI governance actually looks like. And in 10 days, MoMA will open its first-ever exhibition dedicated to information design.
Different stories, same bigger question: where is data becoming more valuable, more strategic, and harder to ignore?
Here’s what caught our attention this week.👇🏻
🚀Beyond the Summit
The Summit is over. The community isn't.
What happened in Washington doesn't have to stay in Washington. World of DaaS membership is how the conversation continues: peer circles, warm introductions, private gatherings across major cities, and a guaranteed seat at next year's Summit.
The people you met in that room? They're in here too.
📰 News & Reads
Revolut confirmed a data breach after hackers used fake government requests to trick the company into handing over customer data.
The attackers impersonated law enforcement agencies to file emergency disclosure requests, a legal mechanism that lets authorities bypass normal privacy protections in urgent cases. Revolut complied. The attack didn't require breaking any technical defenses. It required faking the identity of the people authorized to request data.
AI is learning the language of financial markets, and Chicago Booth researchers just showed that the same transformer models behind ChatGPT can map Wall Street’s genome.
The Market Genome Project trained a BERT-style model on institutional investor portfolios instead of text. The result: asset embeddings that explain over 70% of cross-sectional variation in stock returns, far outperforming decades of traditional factor models. For anyone building financial data products, this signals where the edge is moving.
Anthropic CEO Dario Amodei just published a plan for how the AI industry should govern itself, and committed Anthropic to implementing the first step unilaterally.
The proposal: embedded third-party evaluators with access to AI companies' data pipelines, training processes, and safety incidents. Anthropic is inviting them in now. Sam Altman agreed. As the companies training on the most data in the world voluntarily open their doors to oversight, the governance era of AI is no longer theoretical.
SIX Swiss Exchange just launched a marketplace that gives institutional and retail users real-time access to European exchange data, and lets them feed it directly into AI and quantitative workflows.
SIX Market Signal combines real-time consolidated best bid/offer data from Swiss, Spanish, and 16 other European markets. Users are explicitly licensed to train trading models and AI agents on the data, removing one of the biggest friction points in building AI for financial markets.
🎙️ DaaS Podcast Spotlight
What happens when the person sourcing your data is no longer a person?
→ Freeman Lewin co-founded Brickroad after years negotiating data contracts with frontier AI labs. His central argument: the "messy middle" of data procurement, sorting through thousands of sources, filling out DDQs, acquiring samples, is exactly what AI agents should be doing. Not replacing the data strategist, but freeing them to do the work that actually requires judgment.
In this episode, they discuss:
Why the ICP for data buyers is shifting from "silver-haired guys at three-martini lunches" to engineers and researchers who will pay a premium to never get on a call.
How Brickroad's agents infer data set quality without explicit signals, using things like how long an app has been in the App Store or when a company was incorporated.
Why the real frontier for data sales isn't OpenAI, it's the thousands of small, well-funded specialist labs buying rights-cleared structured data for post-training and evals.
What predatory licensing from AI labs actually looks like and why some companies are signing $50M MSAs and extracting only $5,000.
Why agents meter naturally and what that means for the future of enterprise data pricing.
An Unexpected Data Read For Today.
Before a humanoid robot can clean your house, it needs to watch a human do it first. Thousands of times.
That's exactly what Mecka AI is building. The startup pays people to record themselves doing everyday tasks: vacuuming, folding laundry, making coffee, using body sensors and smartphones. That footage becomes the training data for the next generation of humanoid robots.
Sequoia just led a round valuing Mecka at $500 million.
The insight is uncomfortable: robots can't learn from text or images alone. They need physical-world data. And that data barely exists. So someone has to generate it, one recorded chore at a time.
The most valuable dataset for the future of robotics isn't in a server farm. It's in your kitchen.
Where Data Meets Art
In 10 days, MoMA opens its first exhibition ever devoted entirely to information design.
Full Disclosure: The Edge of Information Design brings together 30 works of digital and analog information design around one question: at a time when public trust in data and institutions has frayed, how do designers make invisible systems visible and who decides what gets shown?
MoMA put it plainly in their own words: "We swim in an ocean of data. Every click, weather pattern, border crossing, and blood test is tracked and fed into systems that increasingly inform our daily behaviors. Data visualization deals in trust as much as in truth."
For anyone in the data industry, it's worth the trip to New York.
September 27 — December 2026. Free.
🔥 In Case You Missed It…
M&A Funding Roundup
H1 acquired Defacto Health to expand its provider network data intelligence platform for health plans.
Moonshot Partners acquired CustomWeather to build an integrated earth, climate, and geospatial data ecosystem.
Harvey raised $550M at a $15.5B valuation led by Diffusion and Lightspeed to scale its AI data and legal intelligence platform.
FNZ raised $450M from existing institutional shareholders to expand its wealth management data and technology infrastructure.
🌱We’re Growing Our Team!
📣Calling our community, one more time 👀
Applications for our Head of Operations & Community role will be closing soon.
If you’ve been thinking about applying, this is your sign. The role spans membership, events, the Summit, and the day-to-day work behind a growing global community of CEOs, founders, investors, and data leaders.
If that sounds like you or if you know of someone great, we’d love to hear from you. Reach out to [email protected] with a little about yourself and why you think you could be part of what we’re building.🙌🏻
Till next time,
The World of DaaS Team.
