Today’s theme is consolidation without consequence. The data industry spent late August writing checks to buy governance, infrastructure, and engineering talent at scale — while the same week’s legal and policy stories show how little actual accountability travels with that money. My take: until “data governance” means enforceable rights for the people whose data it is, these acquisitions are risk-management theater dressed up as strategy.
The buying spree
Nasuni’s purchase of DryvIQ is its second deal of 2026, following April’s Resilio buy, and it folds content classification into a platform serving DryvIQ’s 1,100-plus customers. That’s a vendor betting governance sells better as a bundled feature than as a standalone product — which tells you something about how commoditized “we’ll classify your data” has become. Meanwhile Deloitte grabbing Wavicle’s Databricks practice (terms undisclosed, naturally) is the consulting giants doing the same thing one layer up: buy the engineers who build the pipelines, then sell the governance story to enterprise clients. And at the infrastructure layer, KKR and Singtel closed their S$6.6 billion (US$5.1 billion) buyout of the remaining 82% of ST Telemedia Global Data Centres — the physical plumbing all this governed data has to live on. Three deals, three layers of the stack, one direction: consolidate first, ask about rights later.
The rights nobody gets
That’s what makes EFF’s review of five state location-data bans so pointed: Connecticut, Maryland, New Jersey, Oregon, and Virginia all banned selling precise geolocation this year, and not one gives consumers a private right to sue. Enforcement is left entirely to attorneys general who have better things to do. Same posture shows up in EFF’s fight against “market dilution” copyright theory, where courts in the US, UK, Germany, and EU are diverging on whether AI merely competing with human artists should count against fair use — a fight about who controls training data with almost no individual recourse built in either way. And California’s unanimous passage of AB 1709’s ban on addictive recommendation algorithms for under-16 users, which EFF wants Newsom to veto, is the same pattern in policy form: sweeping restriction, no clear mechanism for the people affected to enforce it themselves.
Training data’s shaky economics
Underneath all this, the actual training-data business looks wobblier than the M&A headlines suggest. Moonshot AI is offering a 30% cut to Microsoft, Amazon, and Google just to host Kimi K3 — a sign compute leverage still runs one direction. Appen is touting an AI-data rebound while its own shares slid 7.7% to 1.14, which is the market openly disagreeing with the press release. And Bolt’s up-to-$27 million pay-to-play bridge, which strips equity from investors who don’t re-up, shows how punitive capital has gotten for anyone outside the frontier labs. Add Vin Vashishta’s warning that SME-written eval answer keys are a skills-gap red flag and HUMAIN and MinIO’s terms-free, customer-free “data fabric” announcement, and you get an industry better at announcing infrastructure than proving it works. Separately, Apple’s new forensic evidence against a former engineer and OpenAI is a reminder that even corporate IP protection is now an argument about what an AI agent did, not just what a file contained.
Watch tomorrow for whether Newsom signs or vetoes AB 1709 — it’ll be the first real signal of how much appetite state government has left for regulating algorithms versus just banning data sales nobody can enforce.
Stories covered
- Moonshot Seeks 30% Cut From Microsoft, Amazon, Google to Host Kimi K3 (Deals & Funding)
- Nasuni Buys DryvIQ, Its Second 2026 Deal, to Bundle Data Governance (Deals & Funding)
- Appen (ASX:APX) Touts AI-Data Rebound While Its Own Shares Slide 7.7% (AI Training Data)
- HUMAIN and MinIO Announce AI Data Fabric Deal — No Terms, No Customers Named (Data Markets)
- Deloitte Buys Wavicle’s Databricks Shop; Terms Undisclosed (Deals & Funding)
- Bolt’s Breslow Raises Up to $27M Pay-to-Play Bridge to Avoid Collapse (Deals & Funding)
- Five States Ban Location Data Sales, But None Let Consumers Sue (Licensing & Legal)
- KKR-Singtel’s S$6.6 Bil STT GDC Buyout Closes — Deal Terms Check Out, Timeline Barely (Deals & Funding)
- EFF Fights ‘Market Dilution’ Theory as AI Copyright Rulings Diverge Worldwide (Licensing & Legal)
- Vin Vashishta: SME-Written Eval Answer Keys Are a Red Flag, Not a Fix (AI Training Data)
- Apple’s MacBook Evidence Against OpenAI Hinges on an AI Agent, Not Just a File (Licensing & Legal)
- California Passes AB 1709 Youth Social Media Ban; EFF Presses Newsom to Veto (Licensing & Legal)
Rhea Rundown is an AI-assisted column persona of The Data Commenter; every column passes the newsroom quality gate before publication. Nothing here is investment advice.
