Funding The Cloud Challengers

As the Big 3 cloud providers have expanded their services over the years, each category has spawned new opportunities for challenger startups to compete with them. Venture capital dollars have followed closely behind.

🎧 Listen to this article >

As evidenced by success stories such as Snowflake’s triumph in spite of Redshift, winning against a major cloud vendor is possible. This is why although AWS, GCP and Microsoft Azure collectively offer hundreds of services, there have been billions of VC dollars backing a growing cadre of challenger startups with increasingly specialized and sophisticated products.

This is represented in our accompanying data visualization chart, with each bubble representing a standalone opportunity.

Five Trends in VC Funding

In reviewing the VC funding data between 2019-2020, five trends have stood out in particular as areas of interest to us at Greylock:

  • An (unsurprising) increase in the number of AI and ML services and funding;
  • Growth in funding into (and the creation of) data and analytics startups;
  • Increased funding into security services;
  • A maturation of startups operating in the observability sector, which we predict will lead to more funding;
  • An unexpectedly small API tooling market.

What’s driving these trends?

1/ AI and ML: The Whales of the Cloud

The promise of the potential impact of artificial intelligence and machine learning has prompted everyone, from the largest cloud providers to the tiniest challengers, to chase after the sector. In both the number of services and number of VC dollars invested, this category greatly outweighs the others tracked in our database.

Between the Big 3 Cloud providers, there are 74 AI and ML services offered. This is the highest number of services provided in a market, and is almost double the size of the second-largest category of Developer Tools, which has 44 services. In spite of the strong focus from the cloud vendors, venture capital firms have put $4 billion over 2019 and 2020 into startup challengers in the sector.

Given the surface area covered by the cloud providers in this sector, it’s particularly impressive to see challengers go up against the incumbents here.

The Big 3 possess seemingly impenetrable moats: Proprietary datasets (for example, AWS can offer pre-trained fraud detection models based on their retail fraud datasets they’ve accrued), combined with seamless interoperability and economies of scope. In AWS’ case, those three advantages allow for interoperability between services, from getting the data from S3 and all the way to model training in SageMaker.

While their sheer scale makes them impossible to replicate, almost every cloud service is nonetheless facing a VC-backed competitor. Challenger startups can go down two routes: from end-to-end or point-specific ML infrastructure (e.g. Snorkel), to applied AI /ML in enterprise-specific applications (e.g. Abnormal Security). Challengers have been able to command large market share against the cloud, such as DataRobot competing directly with household ML name AWS SageMaker.

The cloud providers have also made significant effort in services around text, speech, and image recognition, which is conducive to AI applications. Startup challengers have likewise stepped up with practical applications of such tools in a variety of formats – from improving writing with Grammarly to unlocking information with Instabase.

Investors have shown a particular excitement around these AI-enabled workflows, which are bringing long-needed efficiencies to areas with historically poor optimization. Cresta, for example, provides a real-time AI platform to large enterprises that leverages expertise to all employees, regardless of skill level.

While there are a myriad of standalone AI-leveraged SaaS opportunities that the cloud providers might not focus on, we do wonder if where we will see a battle between Big Cloud and challengers is in the end-to-end model creation and deployment infrastructure.

In these markets, the biggest advantage a cloud provider will have is easily interacting with and moving data across services in their ecosystem, and the volume of proprietary datasets available to pre-train models. That said, we still think cloud providers will have to go up against a number of highly capable and fast moving startups.

2/ Analytics on Fire

Analytics has long been a market that has captured the attention of both the big cloud providers and VC-backed challengers. But the pace at which new companies are being founded and funded in recent years is unprecedented.

Between 2019 and 2020, venture capital investors poured around $3B dollars into around 35 challenger startups valued over $500M in the analytics sector, which are competing with a similar number of services offered by the big cloud providers.

The move from storage to databases to analytics over the past years has led to outsized outcomes in modern data warehousing and data lakes. The aforementioned success story of Snowflake provides a great case study for how challengers do sometimes win against the existing cloud products, thanks to innovative product architecture, customer centricity, and speed of execution.

And the opportunity for challengers doesn’t stop there.

The modern data stack is still a work in progress, but already, data scientists and analysts have no shortage of tools thanks to the innovation of a growing number of challengers. A myriad of analytics applications are being built around the data warehouse: data quality, data governance, data privacy, and data cataloging – all categories currently present standalone solutions an enterprise can implement.

For example, data scientists and developers can use products from the emerging class of cloud-only startups like Rockset as an analytics engine to build data-driven applications as simply as possible.

There is a common theme between all of these categories: the amount of data being collected is untenable, unclean, and (as the organization grows) inconsistently tagged. At the same time, data science teams are over-burdened with requests for bespoke analysis and questions about data from business users.

Challengers in this category are ushering us into an era where the ability for employees to ingest and perform analytics on their data (and in a secure and compliant way) is available to all.

Of course, we can envision the cloud vendors doubling down on these capabilities within their own warehouses. But overall, we believe the private sector is at an inflection point in regards to spending and talent to the point where we would question whether it is in the cloud vendor’s best interest to devote the substantial resources required to catch up.

3/ Security: The Gift that Keeps on Giving

Security is an ever-present and ever-evolving problem for all industries.

So it’s no surprise that venture capital investment into startups operating in the sector grew considerably between 2019 and 2020. For that period, investment was over $3Bn into around 30 challenger startups that are each valued above $500M. They are competing with 25 services from the big cloud providers.

At the beginning of the Covid-19 pandemic, most industry reports forecasted a decrease in overall IT spend. But enterprise organizations actually spent more on security tools during that time, due to widespread shifts to virtual work and school. Moreover, this was happening alongside an acceleration of migration to the cloud by nearly all industries.

VC spending in 2020 was characterized by uncommonly large raises and valuations at the earliest stages of a company’s existence. Hundreds of millions of dollars have gone into CNSP (Cloud Native Security Platforms) that focus on CSPM (Cloud Security Posture Management) like JupiterOne, and sometimes also CWPP (Cloud Workload Protection Platforms) like Wiz, and Orca.

The increase in work from home affected endpoint platforms (such as Crowdstrike) and zero trust network access platforms (such as Twingate and Tailscale). The well-funded Cloud Security Posture Management (CSPMs) companies like Wiz, Orca, and Open Raven were the main beneficiaries during this time as regulated industries relied heavily on their services in their rush to migrate to the cloud.

This adoption of more technology brings an additional problem to solve.

As organizations modernize in response to Covid-19 and its subsequent acceleration of cloud migration, they are utilizing more diverse tech stacks.

This means they are leveraging an increasing number of platforms to service customers, which results in a proliferation of customer data across multiple locations. As a result, these organizations have increased focus on compliance and privacy, and an enhanced need for visibility and tracking of all customer data. This is why we are seeing more companies offering privacy management solutions like One Trust, and centralized access control with compliance workflows, like Privacera.

Within the security vein, the identity and authentication sector has matured to the point where we are starting to see consolidation, such as Okta’s acquisition of Auth0. We expect to see further consolidation down the road as companies within the various security submarkets go after similar customers and use cases, such as the way CSPM and CWPP functions are starting to overlap.

Complex and risky as these undertakings are for the challenger startups that attempt to take on these problems, we have seen a lot of funding go into these solutions. We expect further investor excitement for startups finding even better solutions.

4/ Observability: The Under-Served Market?

There have been some great historical outcomes in observability, from traditional approaches in AppDynamics and New Relic to modern platforms like DataDog.

Now, we’re seeing the industry evolve from the three elements of metrics, traces, and logs, to a more “jobs to be done” approach of know, triage, understand – and then fix.

With the evolution of the hyper-scale era of cloud comes the need to monitor cloud-native apps and make precise, data-driven decisions. Querying and storing one dimension of a company’s data is simple. But the cost and complexity to query and store all the metrics within all the dimensions quickly becomes not only cost prohibitive, but also slow and impossible to use. That’s where a company like Chronosphere comes in, with a hosted cloud service allowing enterprises to ascertain deep insights from their metrics.

At Greylock, we believe that the cloud monitoring space is large and growing, reflected by our current investment in Chronosphere and our earlier investment in Sumo Logic, which went public in 2020.

Recent activity like DataDog’s acquisition of Timber Technologies (the creators of Vector) and of Sqreen, or ServiceNow’s acquisition of Lightstep, has demonstrated that marquee companies see value in this market.

That said, the investor dollars spent in this market are significantly lower than others – just above $300M in the last two years.

One could argue the lack of VC investment is because of strong competition from cloud vendors. Azure and AWS both have solutions in this category, and GCP is leading in the number of observability services.

Moreover, the incumbents’ deep technological moats and the significant lift a challenger must make to get a new observability product off the ground means few startups have been willing to get into the sector. Chronosphere is an outlier because founders Martin Mao and Rob Skillington had built the M3 platform at Uber, which provided a solid jumping-off point for their startup.

However, we are bullish that opportunity exists for more challengers to get into the sector. While all of the large cloud providers have observability products available, none of them is a truly strong product that can compete with the level of complexity a DataDog can handle.

A service like AWS Cloudwatch may have the advantage of raw processing power and ease of integration with other AWS services, but once a customer needs observability for larger and more complex data environments, they will be best served by a product from a challenger.

Additionally, we are still a few years away from a strong cloud-provided multi-cloud observability product, meaning there is ample white space for challengers to go after.

5/ API: Small Footprint, Big Opportunity?

APIs have arguably enabled singular products to become platforms. Without the Google Maps API, we never would have had Uber. The Skyscanner API allows companies to find and aggregate flights. The Clearbit API allows a multitude of sales companies to pull in prospect data.

The evolution of the API ecosystem could mirror the evolution of the SaaS market, wherein the explosion of SaaS companies in turn created a market for enterprise SaaS management. Over the past two decades, we saw the creation of numerous SaaS businesses – from horizontal apps like Salesforce in CRM and Workday in HR / Financials, to vertical apps like Veeva in life sciences and Blend in financial services.

Similarly, some of the largest tech companies in the past decade are API-based businesses – from Twilio in the public markets to Stripe in private, along with many new challengers. As a consequence to this explosion of API-based businesses, large markets are being created around finding, connecting, and managing the thousands of APIs a company uses.

That’s why it’s very surprising that the market for API management and development has not grown to command more of a sizable share of the overall cloud ecosystem. This is particularly notable in the low VC dollar spend. In 2019 and 2020, financing was a mere $400M, and there were fewer than five large challengers competing with the big cloud providers.

We haven’t seen many challengers attempting to compete with the big cloud provider API Gateways (from Google’s Apigee to AWS’s API Gateway), as the large incumbents have a strong foothold and enable easy integration with all existing services. One of the few remaining gateway challengers is Kong, known for its open source gateway.

That said, we have seen challengers making a name for themselves with tools that bring visibility into a company’s API consumption. Promising startups in API management and tracking include RapidAPI, and Solo.io, while other challengers are focused squarely on API security, like Salt, and CloudVector. These solutions are still in their infancy, and we believe there is a greenspace opportunity for new, high-quality entrants.

Conclusion

Despite the Big 3 Cloud castles having massive moats, there are certain markets like AI / ML and Analytics where the pie is so large the challengers and VCs are willing to go toe-to-toe in an attempt to claim their slice.

Similarly, there are markets like Security that move at such a fast pace – and are predicated on larger economic macro-trends – that we expect VC spend to be effectively evergreen.

Lastly, in the Observability and API markets, we see signs of opportunities for ‘Act II’ companies in areas where the Big 3 cloud vendors haven’t been able to focus as strongly.

As we continue to add companies to the database, we are eager to see how the above trends play out and which other trends are revealed.

We invite participants to weigh in on our analysis and predictions, and to share any insights they may have as we work to identify more opportunities for founders.