ElevenLabs lists BlackRock, Jamie Foxx and Eva Longoria as new investors

ElevenLabs expanded its $500 million Series D with new backers including BlackRock, NVIDIA, Salesforce, and Deutsche Telekom, while also surpassing $500 million in ARR after ending last year near $350 million. Management said Q1 2026 added $100 million in net new ARR to roughly $450 million and the company also closed a $100 million tender. The startup has rapidly scaled valuation from $6.6 billion last September to $11 billion in February and continues adding enterprise contracts across voice AI.

Analysis

This is less a pure startup financing story than a signal that voice is moving from “nice-to-have demo layer” to budgeted infrastructure inside customer-service, payments, and enterprise workflow stacks. The second-order winner is the picks-and-shovels layer: hyperscale GPU supply, inference orchestration, and enterprise software vendors that can embed voice without building models from scratch. That makes the strategic participation by large cap tech and software investors more important than the celebrity names — it validates that voice is becoming a distribution and retention feature, not just a model benchmark.

For NVDA, the implication is not just training demand but a broader expansion in inference intensity as voice becomes a default interface for real-time customer interactions. Low-latency, always-on audio is computationally expensive relative to text, so even modest penetration into call centers or in-car assistants can meaningfully lift token-equivalent consumption and keep enterprise AI workloads on a high-growth path. For CRM, the risk is mixed: near term, embedded voice agents can improve seat stickiness and ARPU, but over time they can also compress differentiation if customers route around traditional UI layers and buy orchestration directly from model-native vendors.

The contrarian risk is that valuation and ARR growth are now high enough that any slowdown in enterprise conversion will be punished quickly. Voice products are notoriously sensitive to latency, accent robustness, and security incidents; one material failure in a regulated use case could create a multi-quarter sales reset. The other overhang is channel conflict: if retail access through a venture product becomes real, it may improve brand reach but also raise governance and liquidity questions around secondary supply and future pricing discipline.

Bottom line, this is bullish for the AI infrastructure complex and selectively for enterprise software that can monetize voice workflows, but less clearly positive for incumbents exposed to AI feature commoditization. The cleanest expression is to own the beneficiaries of rising inference demand while being cautious on software names where voice becomes a substitute rather than an enhancement.

Key Decisions for Investors

• Long NVDA on a 3-6 month horizon: voice adoption is an incremental inference tailwind, with upside if enterprise call-center deployment accelerates; use dips around broader AI rotation for entry, target risk/reward ~2:1.
• Pair trade: long NVDA / short a basket of lower-quality AI software names that trade on model wrappers rather than infrastructure economics; thesis is that voice shifts spend toward compute and away from thinly differentiated application layers.
• Tactical long CRM into any pullback over the next 1-2 quarters: voice can expand workflow attachment and services revenue, but size modestly because some value accrues to native voice providers rather than incumbent SaaS platforms.
• Buy downside protection on CRM 3-6 months out if the stock re-rates on AI optimism: the market may be overestimating how much voice preserves SaaS pricing power as interfaces become commoditized.
• Watch KLAR for 1-2 quarters as a beneficiary of better conversational automation in consumer support, but avoid chasing until there is evidence that voice reduces service costs without increasing fraud or compliance losses.