AT&T Ventures' leader says AI has flipped seed-stage risk from 'can they build it?' to 'is it defensible?' — forcing corporate venture arms to offer real-world validation beyond just capital.

What happened: Vikram Taneja, head of AT&T Ventures, argues that AI tools have lowered the barrier to building working software so dramatically that the traditional measure of seed-stage technical risk — whether a team can execute — is becoming less relevant. Instead, the critical question has shifted to defensibility: whether a company's technology can compound and withstand competition from frontier AI labs moving into application layers.

Why it matters: As institutional venture firms rush into seed rounds to secure larger stakes earlier, and founders can prototype production-ready applications faster, distribution and go-to-market strategy now arise at seed instead of later in funding cycles. For corporate venture arms like AT&T Ventures, this reshapes competition — pure capital is no longer enough. The body suggests AT&T's advantage lies in offering proof-of-concept partnerships and real-world technical validation from a major network operator, which financial VCs structurally cannot provide.

What to watch: AT&T Ventures is actively seeking seed-stage companies with proprietary data, embedded domain expertise that general-purpose AI lacks, or highly specialized niche markets that are too targeted for frontier labs to pursue directly. The firm is running ongoing proof-of-concepts with portfolio companies in areas such as AI-RAN (artificial intelligence for radio access networks), connected infrastructure, and computer vision.