The Tech Sales Newsletter #101: The state of cloud infrastructure software in 2025 (part 2)
This week we continue with the deep dive in cloud infrastructure software, curtesy of the InfraRed report. Scroll to the bottom if you want to see many new logos of company worth investigating further.
The key takeaway
For tech sales: The biggest opportunity in cloud infrastructure software today is in private companies with high growth that can reward you both with strong OTE and a successful exit. AI has accelerated the timelines for companies to achieve this outcome, and after several years of most reps chasing either stability or higher OTE, we are back to evaluating the full earnings potential in front of us.
For investors: While sales reps can benefit from this wave of growth, most investors are limited to public companies. I know that some of you will not like hearing this, but it might be time to seriously evaluate also becoming a VC.
The bigger picture
Apache Iceberg has emerged as one of the leading open table formats in the data lakehouse ecosystem. This format enables organizations to store and manage large volumes of data with proper governance, transactions, and time-travel capabilities. For AI workloads, it works well for managing training data, feature stores, and batch processing pipelines, while vector databases remain the preferred choice for inference and similarity search use cases.
The side effect was that LinkedIn got overrun with aggressive and bitter spats between Databricks and Snowflake GTM members, with posts even getting C-level executives to chip in and yell at the competition.
AI workloads have other downstream effects in the data layer, with companies chasing significantly more efficient architectures in order to maintain a reasonable cost basis as data workloads increase in scope and size.
This slide is a good example that we should always take VC content with a grain of salt - Elastic already redesigned their cloud solution to be built on top of S3 buckets, while both MongoDB and Kafka have offered S3 storage integrations (as bolt-on functionality and with performance penalties).
The significant changes in how we store and access data at scale are downstream of the significant demand for scaling AI applications. As of this writing, the utilization of AI for developing, testing and deploying code has exploded significantly beyond the expectations we had even 12 months ago.
Multiple parts of the CI/CD pipeline can be taken over by agentic workflows, which also has a side effect that can drive higher quality development. One of the best practices in the industry is test-driven development, where you start with scoping what the expected outcome is, build the tests to prove it's working as expected, and then actually create the code that needs to pass the tests. Many developer teams fail to follow such a process since it adds a significant overhead of work and planning.
So, what happens if we have the people qualify and decide what they want the software to look and operate like, but then we scale agents across multiple steps of the process in a way that allows us to follow best practices and still meet aggressive deadlines?
This evolution of workflows is a logical step forward, based on expanded capabilities and opportunities to pursue. It also reveals some of the massive inefficiencies that R&D teams have been operating under for a long time. The same logic can be expanded across many other parts of the business.
The best part of this market is that there are very few companies right now offering the practical new tooling that will expand from code generation into the other parts of the CI/CD pipeline. More interestingly, we are not yet seeing innovation in the space either, i.e., what happens if we challenge core concepts and approach software development from a completely different workflow perspective.
This is a good thing btw, it means that you can be part of the next game-changing wave across parts of the industry. Our VC frens at RedPoint see a 115x market growth potential for DevOps tooling.
The investment in AI is driving downstream effects across cloud infrastructure software. One of the obvious segments with the strongest growth remains cybersecurity.
More importantly, even if an organization is doing spending cuts, it's unlikely that will hit their cybersecurity spend, unlike "muh VERTICAL SaaS" that LinkedIn enthusiasts like to associate with tech sales.
More importantly, companies continue to adopt and invest in a variety of tools, often due to a strong preference for best-in-breed software. While the long-term view of this newsletter is that platforms will own the majority of the stack (and they have already done so from a revenue perspective), at least verbally, CISOs keep insisting on best-in-breed strategies.
Ignore the logos (some obvious winners with some deeply underperforming orgs), the mix here is a good indication of where spend is—"Identify" and "Protect" parts of the typical security frameworks such as NIST.
If the existing spend is going toward proven, already deeply familiar products, the interest over the next 12-36 months would be in testing and potentially buying a variety of products that are essentially tied to AI workloads and applications.
This is where a new winner will be born.
From these, the really valuable categories are solving the AI SoC problem and Machine-to-Machine authentication. As agentic workflows start to be implemented at a large scale, they'll be targeted aggressively as a weak point. Both the authentication and afterwards lateral movement post-breach would be critical parts of managing their behavior from a security standpoint.
This, however, also opens up the potential for larger players to consolidate even more aggressively, both due to significantly larger engineering orgs, as well as access to deep funding.
This is becoming very visible in the products that existing incumbents across cloud infrastructure software are actually shipping. AI-powered features or supporting software is the majority of new R&D.
Funding for new players is also strongly correlated to AI workloads. Valuations for non-AI companies are significantly lower across the board.
More importantly, if you are seriously considering investing the next years of your life in tech sales, it's difficult to ignore the significant valuation premium that cloud infrastructure companies have and the impact that exits would have on your overall compensation from that work.
As a reminder, you get to cash out from your share in two main scenarios - company goes public or it gets bought. There is a third scenario called secondary market where another investor takes over your shares (Databricks recently had such an event as part of their Series F), but it's more rare. Realistically speaking, the company either needs to get bought or go public (in the case of Figma recently, they were forced to go IPO because the Adobe acquisition got blocked).
While most folks in GTM roles have a very narrow focus on their OTE comp, the reality today is that OTE goes hand-in-hand with equity outcomes if you are in the right company. AI has accelerated both the valuation opportunity, as well as the timing to an event (Wiz going from 0 to a massive acquisition within 4 years was partly driven by their incredible PMF as a cloud infra company and the impact it had on scaling workloads securely).
Let's close this article with a selection of the companies that RedPoint has invested in and are worth considering in this context:
