newsfilter.io
Panel

Kong, Commit, Deepset, Strapi, LiveKit & FINOS: The OSS Paradox Free Software, Priceless Value

RAISE SummitOlivier, Russ, Alessandro, Milos, Pierre, Reza Shafi, Gabriel
Watch on YouTube →

  • LiveKit (Russ)

    • Founded in 2021 during the pandemic to address the lack of infrastructure for real-time audio/video streaming, which HTTP/HTTPS was not designed to handle.
    • Adopted an open-source-first strategy to allow developers to build streaming features without building underlying infrastructure.
    • Transitioned from a project to a company after enterprises requested paid hosting and scaling support.
    • Defines monetization boundaries by keeping the single-server open-source code entirely free while charging for the orchestration system required for horizontal scaling.
    • Offers three product tiers: single-server open source, a managed cloud service (LiveKit Cloud) for scale/latency, and "LiveKit Private Cloud" for on-prem compliance.
    • Argues that changing licenses is a defensive failure; the best strategy is to move faster than large cloud competitors who could fork the project.

  • Investment Perspective (Olivier)

    • Leverages data-driven platforms to identify early-stage open source investment opportunities, noting that most European investors avoid early-stage open source due to high valuations once revenue reaches $2-3M.
    • Views early investment as a "moral contract" ensuring the company balances community health with future monetization to avoid community betrayal.
    • Recommends selecting the appropriate license at the seed stage rather than changing it later, citing the availability of nuanced licenses (e.g., AGPL) today.
    • Believes exit value for open source companies derives from both revenue and community strength, as seen in the GitHub acquisition.

  • Deepset (Milos)

    • Initially built Deepset as a consulting firm before open-sourcing the Haystack AI orchestration framework to overcome enterprise reluctance to trust proprietary AI.
    • Adopts Joel Spolsky's "Strategy Letter 5" model: monetize the complement (services/support) rather than the commodity (open source code).
    • Contrasts this with "Open Core" models, noting that while quicker to revenue, they create a conflict of interest where proprietary features cannibalize the open source version.
    • Distinguishes between "application software" (e.g., CRM) and "technology components" (e.g., AI frameworks), arguing the latter requires a different licensing approach to ensure trust and standardization.

  • Strapi (Pierre)

    • Founded Strapi 10 years ago as students to replace traditional CMS tools; raised pre-seed, seed, and Series A rounds in 2018-2020 without revenue to prioritize community growth.
    • Adopted an "Open Core" model in 2020 with an Enterprise Edition, then launched StrapiCloud (hosting) in 2023, which drove significant revenue growth via product-led growth.
    • Identifies "deployment and scale" as the natural cleaving point for monetization, avoiding feature gatekeeping that alienates the community.
    • Notes that hosting is less alienating to the community than Open Core because the open source project remains the "best" technical choice, while hosting solves organizational pain points.

  • Kong (Reza)

    • Brings product leadership experience from MuleSoft, CoreOS, and Red Hat.
    • Distinguishes between "projects" and "products," noting that successful commercial products (e.g., OpenShift) often have open source projects with fewer stars than the product adoption implies.
    • Emphasizes that license changes (e.g., Redis, MongoDB) are often reactions to cloud providers reselling infrastructure software, whereas application software faces less risk of this model.
    • Advises that distribution, certification, licensing, and hosting are adjustable "knobs" that must be aligned with the specific product strategy.

  • Licensing and Market Dynamics

    • Cloud Provider Threat: Large cloud providers often offer "license or kill us" ultimata; changing licenses does not stop them if they have superior resources and speed.
    • Risk Segmentation: Infrastructure software (databases, models) faces higher risk of cloud provider reselling than application software (CMS, CRMs).
    • License Philosophy: Some panelists argue for transparent but non-permissive licenses for technologies where trust is paramount, while others argue permissive licenses are better for application products used as distribution channels.
    • Community Backlash: Recent license changes by companies like Elasticsearch, Redis, and MongoDB have led to community forks and accusations of "rug pulls," hosted by the Linux Foundation.

  • Investor-Founder Alignment

    • Early-stage investors (5-10% ownership) face dilution risks where new investors in later rounds may lack open source expertise or push for SaaS-style multiples that conflict with the long-term community model.
    • Successful monetization requires a clear path from free adoption to paid deployment/scale without splitting features in a way that fragments the user base.

  • Event Closing

    • Panelist Olivier noted that all four company founders are in the top 1-3% of their investment platform's tracked open source projects, though the specific 1% top performer was withheld to maintain suspense.