A VC in Vacationland

For at least as long as I am alive…

Supply Chain Ventures is both an early- and late stage investor. AI is having a significant impact on how we evaluate potential startups as well as mentor existing portcos. The following thoughts focus on how we are going to have to manage our investment decision processes and portfolio management going forward.

Investing in all companies, not just supply chain ones, will never be the same. AI has gobbled up 55% of venture funding from January to September 2025, with 22% of the startups funded being AI based. This trend is likely to continue and expand into 2026 and beyond. It’s not a fad, it’s real. Software historically ‘ate the world’, now it’s AI’s turn. Let’s explore how AI will impact how investors like Supply Chain Ventures will have to review business plans and investing decisions going forward.

The Old Days

We have been around almost 25 years and have seen many revolutions in how supply chains can be managed–from the behind-the-firewall legacy tech, to the SAAS era in the late 1990s to the current ‘everything is AI’ world. There has been one evolving but fairly consistent constant in this process–how founders pitch their companies.

Founders develop a PPT investment pitch deck. Typical projections would say that the company would take three to five years to reach profitability. The pitch typically laid out an eighteen or so month period where founders would be coding the tech to get ready for initial pilots with prospective customers, followed by perhaps a year of pilots where the tech was improved using customer feedback, followed by a year of landing initial customers, and then entering a growth period of steadily increasing double digit sales. That is how it was supposed to work, except that it rarely did. ‘Hockey stick’ revenue projections over the first five years from zero to $20M ARR were usual in investor decks and rarely believed or happened. But that’s what mentors, online websites, accelerators and fellow founders taught fledgling founders to produce so that VCs would take them seriously.

We still see these type of pitch decks every day. We still judge them on the founders domain experience, whether the problem they are solving will create significant value for customers, will customers buy it, how are competitors reacting, can the company can become profitable, and if they have (or can acquire) the right team to make it happen to become a company with a $100M valuation–all with a five year time horizon.

Change is coming to our ability to judge and because of AI, it is coming fast.

The New Days–Early-Stage Startups

We also used to get pitch decks (pre AI) from ‘two undergrads at Stanford’ saying that their supply chain tech startup would revolutionize supply chain management, basically using software to automate various supply chain functionality in execution, planning or design. Same basic deal–it would take five years to build the tech, get customers and reach profitability. We would generally reject them (not always, or we would not be savvy early stage investors) on the basis of lack of domain knowledge. Supply chains are inherently very complex systems involving often hundreds of players to get a shipment from origin to destination. The idea that a magical piece of tech could pull all this off is dreaming (and still is). Human intervention is a critical part of making stuff happen in supply chains, although much can be automated. The concept of ‘dark supply chains’ (analogous to dark factories where humans intervene only when there are problems) is still a long way off.

We are now seeing a different set of ‘two undergrads from Stanford’ proposing a whole new way of creating supply chain tech solutions, using AI coding tools such as Anthropic’ s Claude. They claim to create sophisticated supply chain tech in eighteen DAYS instead of eighteen months (well, not really eighteen days, but months less than eighteen months). We are still quite skeptical–the founders still have minimal supply chain domain experience but the time frame reductions from AI enabled coding is, in a word, concerning.

These founders are not claiming to immediately replace legacy full-scope supply chain solutions across execution, planning, and design. They are promising to answer sophisticated questions, such as ‘how much inventory do I have across all my channels back the suppliers and are where are the opportunities to reallocate product to maximize downstream profitability?’ Legacy software providers are feverishly working to answer the same questions, but are hindered by having to live in the context of hard coded tools that cannot handle real-time data sets, may not easily respond to modification, or take too much development time. It’s the ability to craft simple code in a few days instead of months to ask relevant questions that ‘parasitically’ sit on top of existing software and data that is changing the investing game.

If these founders can significantly reduce software development time, can they also shortcut the process of acquiring domain experience? There is no lack of supply chain domain expertise in legacy software providers or in supply chain intensive industries. Attracting the expertise to even a well-funded startup may be difficult, depending on perceived risk, compensation and flexibility to assume multiple roles besides domain expertise in a startup. We are already seeing younger founders with five or so domain years of experience looking to funding AI enabled supply chain companies.

We can cite many reasons why these founders will not immediately challenge existing software suppliers or venture investors–defensibility of a simply coded platform, security, point rather than broad supply chain solutions, for a few. For investors like us, the lingering thoughts are troubling and twofold:

1. How long will it take those ‘two kids from Stanford’, given help on domain expertise, to develop a useable supply chain execution et. al. solution? Probably less than 18 months. meaning we have to judge pitch decks differently going forward. Whether they can significantly shorten the customer acquisition process, is an open question. Perhaps AI will shorten the sales cycles as well, as customers strive to outdo or keep up with AI-enabled competitors.
2. How does this impact our existing early-stage portcos? Probably a lot more than we would like, given their tools were developed over the previous eighteen month or so time frames and are now ‘legacy systems’ whether they or we like it or not.

Let’s not count out the later stage startups or legacy software players. They benefit from the same fast coding capabilities inherent in Claude and other AI coding platforms. Universal availability of fast coding tools, built at VCs expense, levels the playing field for everyone going forward. It will be a race to the finish among startups and legacy players, with domain expertise driving success as opposed to fast coding.

How we can evaluate these new startups will tackle the markets with faster, tech, new value propositions and innovative business models going forward will be a big challenge, requiring a reevaluation of how we judge the new breed early stage companies.

The New Days–Late Stage Investments

Many of these trends in early stage investing also apply to our investments in later stage companies. Our later stage portcos, such as Capstone, Baxter Planning and nShift, have been around for decades. They have well established, primarily multi-tenancy SAAS tools, to serve their clients.

The ability to quickly develop code to sit on top of these systems to answer sophisticated customer questions cuts many ways for later stage companies–one option is to develop internal IT teams to use Claude et.al. to code the solutions by aggressively copying new external innovators, another is partnering with these innovators to include their current software, or the third is ignoring the whole trend and potentially becoming an also-ran in the space.

Our position as an early stage investor will allow is to be a early-warning system for our late-stage portcos–alerting them to new startups that might challenge them with more sophisticated tool sets. We can also nag them to develop internal teams or external partnerships to advance their AI-enabled tech capabilities.

What’s the Game?

Supply Chain Ventures is far from the only investor to face these challenges. We all live in two worlds today–somewhat tied to the past by our previous investments and having to decide how best to invest in the future.

We will have to change how we do business to meet these challenges. Our existing portco founders will need to step up their AI game. often with our help. We will have to evaluate startups in a whole new light (although still following the fundamentals of investing–team, business models and tech). Finally, we will need to keep a much closer eye on the early-stage space to determine which founders can truly revolutionize supply chain decision making, even if their ‘simple AI coded tech’ is only one piece of a successful company longer term.

It’s always (mostly) a fun day in venture investing…we’ll be sharing a more detailed look into how AI will impact supply chain technology solutions in a later Blog post.