YC Infra Companies + Trends
YCombinator launched in 2005 and in the past 15 years has funded some of the most high-flying, influential startups in the world. Companies like Airbnb, Doordash, Stripe, and Instacart have been funded by YC, participated in its accelerator program, and have since gone on to revolutionize industries.
Though software infrastructure and dev tools startups are not YCombinator's most widely recognized winners (with the clear exception of Stripe), the accelerator has backed many infrastructure companies over the years and has seen considerable success in doing so. Companies like Pagerduty (incident response), Scale AI (infra for AI), Segment (application ETL), and Fivetran (managed ETL) have all come out of YC. These four companies alone represent over $12B in total value created.
YCombinator, because of its wide reach and clear knack for backing winners at an early stage, can give a meaningful view into current trends in technology and in markets, providing an extra finger to the pulse of the day. In this article, I've compiled a list of the companies YC has backed in 2021 that are labeled on the firm's website as infrastructure software. I've grouped these 11 companies into the categories that represent some of today's trends in the space.
DevOps + Deployment Management
DevOps is the operations of software development. DevOps practices and tools are used to ensure that software can be developed quickly and of high quality and be delivered to the customer with the least possible overhead for the developer.
Argonaut enables "one click deployments" allowing engineering teams to focus most on business value, rather than operational overhead. The platform automatically generates terraform, CI/CD definitions, and deployment observability configurations from simple user specifications. A nice touch of the product is broad integration with open-source tools like kubernetes, elasticsearch, and git, which prevents vendor lock-in and theoretically allows users to switch deployment platforms while maintaining the same core tech stack.
Arpio eliminates downtime by continuously replicating one's entire AWS environment to another AWS region. If for example an app is deployed in the US-West Oregon region but that region has an outage for a few hours, with your environment present in a different region (say US-East N. Virginia) you can serve traffic to the healthy region until the issue is resolved. The trick to this product is replicating complex environments completely and accurately. If the product works reliably, companies would undoubtedly save money by using it even at a high price, when the alternative is costly engineering time to do the same thing manually.
Serverless Stack is an open-source framework and company that allows developers to easily build serverless apps. Serverless is a backend application-development paradigm in which one thinks purely in terms of functionality and application logic; the hardware on which code is run is hidden from the developer entirely. With Serverless Stack, you develop apps for AWS Lambda locally while connecting to real AWS services; local changes are deployed to the local lambda immediately, making iteration cycles much faster.
Monitoring + Incident Management
Application monitoring and incident management are becoming more important than ever for a few reasons. One reason is that such monitoring protects a company's bottom line, as downtime can result in customer refunds or lost customers. Also, increasingly powerful dev tools have made writing and shipping software easier than ever, which means that SLA uptime can become a new dimension of competition. Application monitoring is naturally crucial in such competition. I was not surprised to see that a few monitoring and incident management products have come out of YC so far this year:
Rootly is an incident management platform that helps companies resolve service issues more quickly by automating administrative tasks. One example is its formulaic approach (with a manual override) to assign an owner to new issues and automatically create Jira entities or Zoom meetings to help triage. Additionally, the product provides trend insight to operators to help prevent future issues from arising.
SigNoz is an open-source alternative to Datadog. Datadog is a $34B publicy traded cloud application monitoring service, administered via a SaaS analytics platform. The reason to build such monitoring tooling as open-source is to help users avoid vendor lock-in and not define crucial monitoring configurations in a proprietary service. Building tooling and DevOps solutions around open-source can provide huge flexibility in the long run.
Dev Tools + SDKs
The Dev Tools + SDKs category is a catch all when grouping infrastructure software startups. That explains why this section has the most companies of the three categories that I've created from this year's YC batches. Anything that does not address a broad challenge with respect to buliding software (like the two challenges represented above), but is built to help enterprises develop more quickly and effectively is likely a dev tool. Here are those to have come out of YC so far in 2021:
Warrant builds APIs to help developers add access control and authorization to their applications using fewer than 20 lines of code. Each API endpoint and every sub-action within the endpoint can be accepted or rejected using dynamic auth definitions provided by a single Warrant API.
Appollo enables developers to create one app and seamlessly launch it to all major eCommerce platforms. With a single, user-friendly API and SDK, one can connect to Shopify, WooCommerce, Square, and a handful of other services.
Svix builds webhooks as a service for developers. Webhooks are messages sent from applications to a pre-configured url when certain events occur. They work similarly to SMS notifications. Svix allows developers to use the power of webhooks without having to build their own infrastructure, manage their own security concerns, and spend costly engineering time.
Dyte builds SDKs and hosts infrastructure for in-app audio and video streaming. Incumbent competitors like Twilio and Agora have already built huge businesses in the space, but Dyte is attempting to build an experience with even less friction and that requires no more than 10 lines of code.
QuickNode offers a platform for easy and robust development of blockchain applications. The service gives quick access to Ethereum, Bitcoin, BSC, and other blockchains. On each of these chains, QuickNode offers shared nodes which run the protocols and accept client requests, but on the platform customers can also spin up dedicated nodes if their workload requires it.
Neuro builds serverless compute for machine learning. The premise is that ML engineers spend a lot of time managing compute infrastructure to run and train their models; that time could be spent on the models themselves and integrating them into applications. Like the big cloud platforms, Neuro billing uses a "pay-as-you-go" model to allow customers to optimize their spend.
Most Compelling YC Infra Opportunity
The company that I’m most excited to follow over the coming years is the first one listed above: Argonaut. I may be biased by the soft spot that I have for DevOps and infrastructure management products as opposed dev tools because of the more technical nature of the products. More technical products of course carry higher technical risk, so a lot would come down to the experience of the team behind the company and that team's execution. Managing software deployments is such a broad problem that there’s little doubt in the opportunity that Argonaut has if it can nail the product faster and better than competitors.
And speaking of competitors, it’s worth mentioning that this specific problem area has received much investment of late. Companies and products like CircleCI, Harness, Atlassian Bamboo, and big cloud platform offerings have received huge investment over the past five years and offer pretty robust solutions. The space is simply developing and being redefined so quickly that the competitive landscape remains dynamic.
Argonaut’s approach is strategic in building its platform on top of open-source technologies. A lens through which one can analyze the defensibility of an infrastructure software company is whether the product is built on a specific technology, or built on a concept or way of doing things.
A perfect example is HashiCorp, which I wrote about in this article just a couple weeks ago. It would have been easy for HashiCorp to build its terraform product to only provision AWS resources or its nomad orchestration product to work only with a certain type of application container like Docker. But Hashicorp instead built terraform on the concept of idempotent resource provisioning and nomad on the concept of scalable microservice orchestration using any underlying technology, vastly increasing its TAM and giving its customers the flexibility that keeps them around for the long haul.
A few of the YC companies listed above mention that they are built to allow you to do X or Y specifically on AWS. That’s a natural way to first iterate on your technology when engineering resources are few and time is your greatest asset, but I still wouldn’t advise backing your product positioning into your corner by forcing customers into a single-vendor.
It's no surprise to see that YCombinator has already backed a dozen or so infrastructure, application monitoring, and dev tool startups so far this year. I look forward to seeing how these companies progress and how future trends in infrastructure influence YC's coming cohorts.