HeyGen launches Avatar V for identity consistency

Hot take: this looks like a meaningful step forward for avatar video because it targets the real failure mode, not just prettier lip sync.

• –The core pitch is stronger than “more realistic”: it focuses on keeping the same person stable across shots, looks, and long-form output.
• –The 15-second capture requirement is a practical wedge if the quality holds up outside controlled demos.
• –Separating motion/performance from appearance is the most interesting product distinction here; that could make avatar generation much more useful for teams producing variants at scale.
• –The main risk is the usual one for avatar products: the demo may be impressive, but consistency and trust under real production conditions will decide whether this is a true model upgrade or just a marketing reset.