Livepeer (LPT) : a 1-Pager Analysis on the World’s Open Video Infrastructure

5 min readMar 5, 2024

Executive Summary

Livepeer (LPT) is a decentralized video streaming network which provides a marketplace for video transcoding and distribution.
The network aims to reduce costs to integrating video streaming by up to 50x.
LPT is the current leader in the decentralized video streaming space in usage terms, and AI generative videos capabilities are planned for Q3/Q4-2024.
With a strong positioning in a market with important network effects, we think that LPT is undervalued and will outperform the current market cycle.
FDV is estimated at $2.5B. The recommendation is to Buy.

Product, Roadmap & Competition

Architecture : Livepeer is composed of its own P2P network layer and the revenue distribution layer is managed on Arbitrum L2. The network layer is composed of Orchestrators who manage the transcoding routing and/or supply computer resources, Delegators who stake LPT to economically-secure the network and Broadcasters who pay for the service.

Current Product & Usage: The main product is Livepeer Studio — a dashboard which allows users to manage live-streams, video uploads, billing, etc. Customers are still niche-users, mainly decentralized applications, creator economy professionals or crypto-native startups. The equivalent of 23,000 days of videos have been transcoded in Q4–2023.

Future products: Livepeer released a new roadmap “Job type abstraction” to introduce generative AI capabilities in the network. Generative video based on open-source models such as Stable Diffusion represents significant usage & revenue opportunities for Livepeer as generative videos unit revenues are much larger than current transcoding unit revenues.

Competition: Livepeer faces competitors from both centralized and decentralized products on transcoding and generative videos.

Decentralized transcoding : Theta Network, Golem…
Centralized transcoding : Akamai, Cloudfare, Internet Giants (AWS, Azure, Google Cloud)…
Decentralized GPU networks : Akash, Render…
Centralized GPU networks : CoreWeave, Lambda Cloud, Internet Giants (AWS, Azure, Google Cloud)…

A key element on competition analysis is that users main preferences are not based on cost consideration but rather user experience, quality of service or SLAs. Despite much lower costs for decentralized products, the complex user experience and compliance issues (own tokens to use the service) reduce their attractiveness compared to the Giants who profit from a “lock-in” effect on their customer bases.

Economic architecture & Tokenomics Overview

Overall economic architecture: Livepeer’s current network revenue consists of :

Demand-side revenue: Broadcasters pay ETH fees to Orchestrators for transcoding services. ETH Revenues for Orchestrators depend on two variables → Transcoding capacities + LPT staked amount
Supply-side revenue: Delegators (& Orchestrators) earn LPT staking rewards and a part of ETH fees for transcoding services

Tokenomics : LPT tokens serve as:

A “Bond & Slash” model i.e. a bonding mechanism in a delegated proof of stake system, in which slashing can occurs due to protocol violation. This model create an economic incentive to secure the network against a number of attacks.
A “Stake for Acces” model i.e. LPT is required for staking to perform a work as an Orchestrator.
A coordination mechanism to aims a 50% participation rate (LPT staked to Orchestrators). When the participation rate increase, the daily issuance decreases, and inversely.

Token allocation & issuance model:

Initial distribution: 10M LPT tokens allocated to various stakeholders with notably 63.4% for network participants
Current inflation: Dynamic model where LPT issuance increases by 0.00005% when the participation rate falls below 50% in a given round (1 round = 22.4 hours). Inversely, when the participation rate exceeds 50%, inflationary issuance decreases by 0.00005%.

Market & Financial Analysis

Market & Opportunity: With 1.8 trillion hours of contents created per year, video is the main media on the internet today. It represents more than 80% of internet bandwidth and its share will continue to increase in the next years. Currently, Livepeer is transcoding around 150 million hours of video per year or 0.009% of the total per year. On the GPU networks segment, usage is more complex to quantify due to various products and market segments. However, estimations on unit revenues are $3/minute on generative AI videos workload and $0.003/minute on transcoding.

Valuation: Livepeer was trading at roughly 700x Price to Sales (FDV before AI product announcement/annualized Revenues ≤=> $250M/$350K). By transposing this valuation with AI videos workload prices, we can imagine a positive but realistic scenario where annualized Revenues increase by 10x with a constant Price to Sales, thus a FDV around $2.5B. Moreover, Livepeer is currently the lowest valued network based on GPU resources available with approx. $530k/GPU. As GPU resources are considered as a commodity, a convergence of price/GPU is expected.

Price History & Performance: LPT has experienced an increase of approx. 110% in YTD (as of 24/02/24) while the Total Crypto Market Cap increased of approx 48%. However, LPT is still at approx -72% from its previous Market Cap ATH ($1.6B end of 2021). Therefore, the estimation of $2.5B seems coherent with the previous ATH and the expected Total Crypto Market Cap in 2024/2025.

Investment Rationale & Conclusion

Growth Potential: Video streaming services and AI generative videos are experiencing an increasing demand. Decentralized Physical Infrastructures (DePIN) space is still undervalued and will grow along Crypto space growth.

Opportunities:

Strong market position with “Positive network effect” and “Winner takes all” tendencies in Decentralized Infrastructure space
Decentralized Trans-coding & AI capabilities still represent a very small share in the market, thus lot of room for growth is available
Lowest valued decentralized network based on GPU resources

Challenges :

Complexity of the go-to-market against Centralized Giants with strong user base & lock-in effects.
Dependency on Open-Source models such as Stable Diffusion vs In-House models.
GPU Providers arbitraging strategies to optimize revenues by switching between different networks where the economic incentives are the most lucrative i.e the ones with most inflationary models or better price actions.

Sources : Livepeer, Messari, TradingView