
VC-backed startups commit more fraud, and researchers think they know why
THE SO WHAT
If fraud risk is structurally higher in VC-backed companies, boards and LPs need to treat governance as a portfolio-level control, not a company-by-company afterthought. Founders should assume diligence will shift from growth-at-all-costs to verification of metrics, customer quality, and internal controls.
READ THE SOURCE
MORE FROM THE WIRE
Startups & VentureThe Week’s 10 Biggest Funding Rounds: Safe Superintelligence And Commonwealth Fusion Lead With Billion-Dollar Deals
A reported $5 billion round for Safe Superintelligence and $1 billion for Commonwealth Fusion shows capital clustering at the extremes—frontier AI and long-horizon energy. For founders in the middle, expect higher bar on narratives that don’t either reduce AI infra costs or unlock new energy and compute regimes.
Phoebe Gates' shopping app Phia is in damage control mode and offering refunds after a payment scandal
A shopping app paying refunds after earning commissions on sales it didn’t drive is a warning shot on opaque attribution and affiliate logic. If you’re building commerce or adtech, assume regulators and users will demand line-of-sight from click to commission—document your attribution rules and be ready to surface them.
Startups & VentureEveryone else is shrinking satellites. K2 raised $500m to build giant ones, and put AI in orbit.
K2’s $500 million Series D at a $6.8 billion valuation to build large, AI-heavy satellites is a contrarian bet that compute and sensing will move into orbit, not just to the edge. If your product depends on geospatial data or global connectivity, assume new orbital platforms will become both a data source and a compute tier to integrate with.
Startups & VentureThe unsexy layer AI agents actually need: DataBahn raises $40m to sell it
DataBahn’s $40 million Series B for an agentic data control plane is a bet that the bottleneck isn’t models, it’s clean, governed, real-time data access. If you’re piloting agents, budget more for data plumbing and policy orchestration than for yet another model integration.