Jan 15, 2025, 9:52 AM ETFireside Chat, Interview

Victor Riparbelli, CEO @Synthesia: OpenAI vs Anthropic vs X.ai - Who Wins and Why | E1246

Synthesia raised $100 million in a Series D round led by NEA, with participation from all existing investors.
The company's valuation is estimated at $2.1 billion, with early investor Mark Cuban holding approximately 15% of equity from his initial $1 million investment in the Seed round.
The funding provides capital to reach "escape velocity," expand the go-to-market team, and dominate the category.
Founder Victor has maintained "cockroach mode" financial discipline, having raised subsequent rounds without depleting prior capital to preserve unit economics and operational control.

In 2017, the founders pitched a vision of generative video creation but faced rejection from 80–90 investors, primarily due to a European VC industry dominated by private equity mindsets unfamiliar with early-stage technology risks.
A failed attempt to raise $8 million in a later round forced the company to rewind and raise a leaner $3.1 million, which the founder credits for forcing product-market fit and customer obsession.
Victor argues that operating with financial constraints early on prevented distraction into non-core features like deepfake detection and kept the team focused on selling the product from day one.
The founder warns against "too much money too early," noting that large teams at the idea stage can slow down the critical process of learning about the market and customers.

Synthesia targets the infinite market of text and slides, aiming to convert 5% of global text communications into video and audio, potentially creating a $50–100 billion company.
The company predicts a shift where video and audio become the default communication medium, replacing text as the primary method for consuming information, entertainment, and training.
Synthesia differentiates itself not as an "avatar company" but as a workflow platform that handles the entire video value chain, from ideation to distribution, which enterprise clients prefer over standalone model capabilities.
The company intends to build proprietary models specifically for human presentation and dialogue, viewing general-purpose video models (like Sora or Runway) as potential partners rather than direct competitors for their specific niche.
Synthesia utilizes a multi-model strategy, currently relying on OpenAI while shifting workloads to Anthropic and Gemini based on pricing, performance, and specific use cases like content moderation.

The current AI enterprise market is described as being in a "bubble" where buyers are willing to spend on innovation but often lack clear product requirements, leading to a "wall of churn" when renewals fail.
The real signal of success is renewal rates, not the signing of initial contracts or pilots.
Victor expresses skepticism regarding the hype around "AI agents" and "AI employees," arguing that these terms mislead users about the reality of software tools and that true automation of complex business workflows remains distant.
He predicts that foundation models will continue to commoditize, shifting the competitive advantage to distribution, product scaffolding, and specific workflow integration rather than raw model capability.
Victor favors investing in X (Elon Musk) over OpenAI or Anthropic for its potential asymmetric upside and integration of real-time data, though he acknowledges OpenAI's dominance in the consumer distribution layer.
He advocates for "Community Notes" and decentralized provenance tracking (like Shazam for content) over heavy-handed human moderation to solve misinformation and content authenticity issues.

AI is expected to democratize content creation to the point where the cost of producing high-fidelity video approaches zero, shifting competitive advantage to storytelling and idea quality rather than technical production skills.
Roles involving technical execution (camera operation, visual effects) will transition to human-AI collaboration, increasing individual productivity by up to 100x rather than disappearing entirely.
Discovery algorithms (exemplified by TikTok) will become more critical as content supply skyrockets, relying on interest graphs rather than social connections to surface quality content.
Victor believes the industry may be nearing the last generation to rely on text as a default communication method, with future generations consuming information primarily via interactive video and audio.

Synthesia operates out of London due to visa limitations for the founder, choosing to stay despite the US being a more established ecosystem.
Victor argues that London offers distinct advantages in talent loyalty and a lower transactional culture compared to the US, though he acknowledges the ecosystem needs more successful exits to shift employee equity mindsets.
He urges the UK government to maintain favorable tax reliefs for entrepreneurs (SEIS, EIS) and avoid over-regulation, citing Europe's regulatory density as a counter-example that stifles AI innovation.
The company avoids making decisions based on competitor actions, focusing instead on customer needs and long-term product roadmap; he notes that competitors copying their messaging eventually helped validate and accelerate the market.

Victor believes computer strategy games (e.g., World of Warcraft) foster superior decision-making skills and strategic thinking by simulating complex systems and immediate feedback loops.
He identifies his "generosity" and broad, non-specialist curiosity as key traits, allowing him to make analogies across disparate fields rather than focusing on deep expertise in one area.
He views his financial success from secondary sales as liberating, allowing for rational, long-term decision-making free from personal financial stress or the pressure of short-term exits.
He advocates for "secondaries" for founders to sell early, arguing that financial security is crucial for making unbiased strategic choices.
Victor advises entrepreneurs to never build products out of guilt regarding competitors' moves or industry hype, emphasizing that the customer is the only metric that matters.