2020 Investment Focus

Logos are a sampling of Crosslink portfolio companies

At the end of 2019, my wife Nadia and I were lucky enough to be blessed with an awesome gift heading into the holidays, our first baby boy! So… I took some time off, and in between eat, sleep, and other extracurriculars, I was able to carve out time to think about where I would like to focus in 2020.

In my journey thus far in the world of venture capital over the past 6 years, I have learned a few valuable lessons, one of which is the importance of focus. And I believe focus is becoming increasingly important in today’s environment given the volume of startups being forming across the globe and incredible infrastructure now in place to launch and scale up companies. For clarity, being focused does NOT mean the same thing as being overly thesis-driven or closed minded. I believe that the truly best ideas come from outstanding individuals who have a unique vision into the future, and the passion and fortitude to commit to the grueling task of making their vision a reality. For that reason, I aim to put myself in a strong position to meet these unique individuals so they can articulate to me why their thesis is so compelling and must be a future reality.

In order to put myself in a position of strength to meet these extraordinary founders, I must be knowledgeable enough to have a view on where the most interesting areas in the market exist. These are where the likely unicorn founders should exist after all. For that reason, I thought I’d share where I’m spending time this year and hopefully I am lucky enough to cross paths with one of the needles in the haystack!

First Principles

Before diving into my specific areas of focus, I think it’s important to consider some first principles I believe are crucial in identifying attractive early stage investment opportunities:

1. Invest in extraordinary founders with a unique insight and clear vision for how they will define a massive category
2. Align with the founders on why timing is right to either disrupt an existing massive market or create a new market that enables a $1B+ company to be built — keep a prepared mind and an open mind to where these opportunities might exist
3. Have a thesis for how the company can build long-term defensibility (network effects, data moats, IP, distribution advantage, etc.)
4. Identify a credible path to a business model that will become highly profitable over time
5. Take early stage risk: Welcome GTM, product/technology, and competition risk, NOT founder or market timing risk

Sector Focus #1: Enterprise Software

There are a few different ways to organize the enterprise software stack today, but this is how I tend to think about it in the simplest form:

Infrastructure Overview

Starting at the bottom of the stack, infrastructure is becoming dominated by cloud infrastructure, which is dominated by the big three: AWS, Microsoft Azure, and Google Cloud. Given the pace of innovation from the cloud giants, it is becoming much more difficult for startups to compete and win share at the lowest levels of infrastructure (storage, compute, networking). That doesn’t mean that there are no interesting opportunities in infrastructure, it just means investors need to be pickier and always conscious of how the cloud giants will respond. I believe the most compelling pockets of the market for early stage infrastructure startups are in fast growing areas such as data security, edge computing / IoT, hybrid / multi-cloud, microservices architecture, and serverless computing.

Infrastructure Highlights:

• Cloud-native apps will continue to grow share, but AWS, Azure, and GCP are so dominant and continuously innovative that they leave little room for startups to compete in the lower level infrastructure categories (storage, compute, networking)
• Proper Cloud Security and Data Security / Privacy has fallen behind as larger enterprises have fully embraced cloud, SaaS, and APIs — still good opportunities for startups in these categories
• Edge computing is still an emerging category with interesting use cases in industrial, smart cities, IoT, etc. but adoption has been slow due to the nature of end customers
• Multi-Cloud support is important to enterprises so they don’t suffer from vendor lock-in — independent infrastructure startups should highlight this as an advantage where cloud vendors cannot compete
• Serverless growth will continue, as abstraction is the name of the game — there may be opportunities for serverless tooling startups
• Microservices adoption has grown rapidly and these environments are difficult to manage — need more tooling in security, visibility, and control

Developer Tools Overview

At the Dev Tools layer, open source software adoption has begun to dominate across nearly every industry all the way up through the Fortune 500, as developers are empowered to choose their own toolsets to enable faster software innovation. Another lever companies are using to increase the speed of innovation is inserting automation across the software development lifecycle (SDLC). Companies are deploying both open source technologies and proprietary solutions to increase automation throughout the build, test, and deployment process, enabling developers to focus on higher value work.

Developer Tools Highlights:

• The Software Development Lifecycle (SDLC) is the assembly line for building applications, and dev tools serve to increase the quality and efficiency of building products on that assembly line
• Continuous Integration (CI) and automated testing are now mainstream concepts — Jenkins OSS and CircleCI are leading the way here
• Continuous Delivery (CD) adoption is quickly gaining steam where emerging leaders such as Armory (Crosslink portfolio co) are capitalizing
• HashiCorp and GitLab have grown rapidly with big ambitions to abstract infrastructure and simplify DevOps tasks across the SDLC
• Open Source Software (OSS) is a great model for driving developer adoption given both GTM and R&D benefits, but it’s a tricky business model to get right for commercial vendors

Applications Overview

The software application landscape over the past few decades has been dominated by behemoths Microsoft, Oracle, and SAP. And they are still the three largest software vendors by revenue. That said, Salesforce is catching up, alongside others from the SaaS era such as ServiceNow, WorkDay, and Zendesk. However, most of the traditional horizontal opportunities have already been gobbled up and reinvented again with a SaaS / cloud delivery model.

The next generation of great enterprise software companies will be those that can win big with either 1) Vertical market focus or 2) Product-led growth. The vertical software market has already begun to explode with great outcomes in Veeva, Guidewire, and ServiceMax* as well as up-and-comers Procore, ServiceTitan, and Weave*. By product-led growth, I am referring to those companies that throw out the traditional enterprise sales approach and instead find inexpensive ways to deliver great products into the hands of end users, which then proliferate and expand across the entire organization. We have already seen some phenomenal companies develop with this strategy, such as Slack, Dropbox, Atlassian, Twilio, Stripe, and Zoom, and I believe it’s going to be a key ingredient of the dominant software vendors of the future.

Applications Highlights:

Vertical Software:

• Multiple $billion+ industries that are way behind and would benefit from software innovation to help businesses grow faster, increase automation, and improve profitability
• Industries: Construction, Energy, Food/Ag, Financial Services, Healthcare, Home Services, Manufacturing, Real Estate, Retail, Supply Chain
• Particularly focused on businesses that can offer unique datasets / insights and workflow automations that show clear ROI to key stakeholders

Product-Led Growth:

• Disruptive GTM model to the status quo: Products with fundamentally better end user experiences offered at freemium or low cost per-user licenses that then proliferate throughout the enterprise
• Due to the maturity of cloud-based software and APIs, self-service installation is now easy AND employees are now being empowered to make their own purchase decisions (decentralizing IT)
• Forrester (2019): 68% of B2B buyers prefer to research online on their own and 60% prefer not to interact with a sales rep
• Best in breed products win → dollars go to the players that can stay in tune with the user needs and out-innovate
• Vendor lock-in disintegrates → Need to find other methods of defensibility such as network or data advantages

Bringing it all together, this is how I view the most attractive opportunities in enterprise software this year:

Sector Focus #2: AI

Some believe that AI is just the next iteration within the existing enterprise software stack. In many ways, I agree with that statement and the AI stack looks quite similar as what was discussed above. However, AI applications require a different set up infrastructure — GPUs, TPUs, etc. — and an entirely new method of programming — data pipelining, model development, model training, model deployment, etc. There are also fundamentally a new set of applications enabled by the combination of 1) Breakthroughs in AI research, 2) Scale-out cloud computing capabilities, 3) Surge in digital datasets available for model training. Existing infrastructure continues to be pushed to the limits to power these applications, though much of the innovation at the lowest levels will be owned by the well capitalized players. The application layer (in both vertical markets and horizontal functions) is an interesting place for startups to capitalize on, as is the new class of developer / data scientist tools required to develop these complex and specialized applications.

The AI stack as defined by Carnegie Mellon University:

AI Infrastructure Highlights:

• The rise in AI applications owes a huge thanks to advanced chipsets developed by NVIDIA, AMD, Intel, and Google (TPUs) and vast computing resources offered by cloud giants AWS, GCP, and Azure
• There are interesting startups building more specialized AI chips such as Cerebras and there is likely a large growth opportunity here, but those businesses are capital intensive endeavors to build from the ground floor
• AI automation applied to security and infrastructure operations are interesting areas due to the low supply of skilled security and engineering professionals — startups that can enable businesses to deploy technical resources to higher value tasks will see budget allocation

AI Developer Tools Highlights:

• Data scientists and data engineers are under-gunned with tools to do their jobs effectively, and there are large bottlenecks that exists between the massive amount of data coming into the enterprise (structured and unstructured) and the AI models, end user insights, and applications
• There are emerging leaders in this category such as Dataiku, Domino Data Lab, and DataRobot that address the need to help connect datasets, clean data, build models, deploy models, and optimize models, but this is still an evolving category with great opportunities for startups
• Automated Machine Learning (AutoML) is an interesting emerging technology in that it enables developers who are NOT experts in the field to develop practical ML applications — there are likely compelling startup ideas here targeting enterprise use cases or industry verticals
• Enterprise budgets for AI dev tools feel primed for rapid expansion given an eagerness to offer cutting edge capabilities enabled by AI applications

AI Applications Takeaways:

• AI unlocks the ability to run highly intelligent models on massive structured and unstructured datasets in arenas such as text (NLP, machine translation), sensor data (machine learning, deep learning), or images (deep learning, computer vision)
• Vertical AI applications offer great opportunities to leverage industry datasets that were only recently made available (thanks to SaaS apps, APIs, cloud) or were previously too onerous to develop insightful information with rules-based systems or legacy infrastructure.
• Horizontal AI applications offer an opportunity to enable automation in key enterprise functions such as sales, customer support, HR, engineering, security, and finance/operations, particularly where operations are bogged down by manual data entry into traditional software apps
• Differentiation and defensibility for companies in this category is established by accessing valuable proprietary datasets to train and reinforce models that become difficult to replicate — the challenge for startups are finding unique ways to access this proprietary data
• While larger tech platforms like Google or Amazon have an enormous wealth of data advantages, startups must use their advantage of being able to focus on markets and use cases where these players do not appreciate or are far astray from their core business model

Sector Focus #3: Hardware

Hardware has long been viewed as a challenged category to the venture capital industry. And while there was a brief rise in investor sentiment for the category a few years ago, the hype has since died down again due to numerous flameouts from high profile consumer hardware companies. I still believe that there are reasons to be excited about certain pockets of the hardware market due to the rapid decrease in component costs, the rise in small device / edge computing capabilities, and increased functionality delivered from advances in software and AI innovation. This combination of factors creates a unique window for startups to reimagine new experiences for both consumers and enterprises, with opportunities for some potential step function changes that may not be possible in purely a software context.

Connected Hardware Overview

Connected Hardware Highlights:

• Most interesting ideas are companies that are either re-inventing age old products or creating brand new experiences by adding value-added software or services on top of commodity hardware (i.e. Peloton, Ring, Nest, FitBit for consumer; Square, Toast, Samsara, KeepTruckin, Particle for enterprise)
• Very difficult to build massive enterprise value and maintain a competitive advantage solely on the hardware innovation alone, as that is typically replicable; rare exceptions exist where there is a deep technology innovation with core IP (i.e. Molekule)
• The most meaningful long-term value accrues through subscription software or services, particularly when these create brand new experiences and make past hardware-only approaches irrelevant

Robotics Overview

Robotics Highlights:

• Robotics is becoming a more interesting category because of a combination of two significant developments: 1) Dramatic decrease in industrial robot costs, and 2) Rapid innovation in relevant AI fields such as computer vision and deep learning
• Both developments are important but the AI innovation component is most important, as it unlocks new use cases for robots and greatly decreases the development time and integration costs — robots can automatically learn and perform tasks without herculean coding efforts
• Examples of winners thus far include Kiva Systems and 6 River Systems, both which are in the warehouse robotics space
• New winners will likely emerge over the coming years in other industries such as agriculture, construction, manufacturing, and retail given their large TAMs and heavy loads of lower skill, time consuming tasks

Closing Thoughts

Bringing it all together, I believe 2020 is going to be another excellent year to build a startup and I’m particularly keen on agents of change within enterprise software, AI, and hardware. At Crosslink, we have been investing in all three of these categories (and many others) for years, and I’m looking forward to 2020 bringing us more exciting opportunities to learn from the world’s best entrepreneurs. Cheers!

Logos are a sampling of Crosslink portfolio companies