Why these markets
Freight, logistics, and the industrial systems around them are full of software and still full of manual coordination. A shipment or an order passes through parties who do not share a system of record. Someone notices the exception, writes the email, books the appointment, or holds the invoice. The work is valuable because it is physical, timed, and intolerant of a silent mistake.
That is an attractive place for an AI-native company, and a poor place for a loose claim about disruption. The firms that already run these networks will not reorganize around a demo. They will pay attention to a system that can take one defined action reliably, inside rules they recognize, and leave a record.
What AI-native means here
Conventional products add a model to software that was designed for a person to operate. The person remains the integrator. An AI-native company designs the operating loop first: observe, decide within a policy, act in a way a counterparty will accept, and record the result. People set the policy, hold the customer relationship, and take the cases the system must not close.
The themes below are a way of looking, not a closed list and not a claim that Logitech Ventures has a proprietary framework. Adjacent problems qualify when the workflow is real.
| Aspect | Software with AI features | AI-native operation |
|---|---|---|
| Unit of work | A screen that helps a person finish a task | A workflow that can finish routine work and hand off the rest |
| Where intelligence sits | Beside the existing process | Inside the decisions the process is made of |
| Data | Often a report or a chat window on the system of record | The system reads and writes the operational record |
| Exceptions | A queue for a person, with limited memory of the last case | A case with context, a policy, and a next action |
| A serious result | People adopt the tool | A loop closes, and there is a record of what was done |
The table is an editorial distinction used to explain the thesis. It is not an approved investment screen.
01
AI-Native Freight & Logistics
Freight is a chain of decisions made with incomplete information. The software most desks use can store a tender, a status, or an invoice. It still asks a person to notice when those records disagree and to carry the next step by hand.
An AI-native company in this theme treats that chain as the product. The interesting question is not whether a model can describe a shipment. It is whether the system can advance a defined slice of the work — and stop when it should.
02
Autonomous Enterprise Operations
A copilot drafts. An operation finishes a step a counterparty will accept, writes down what it did, and hands a person the cases outside its policy. The second of these is the interest here.
Autonomy, in this usage, is not a claim that people leave the workflow. It is a design in which routine actions do not each require a fresh approval, and in which accountability still has a name.
03
Intelligent Supply Chain Infrastructure
Infrastructure, here, is unglamorous on purpose: the identity of a load, the documents attached to it, the milestones parties will accept, and the permission to act on them.
A tracking page can sit on top of fragmented records and still leave every handover to the phone. The companies in this theme try to make the shared record do work.
04
Industrial AI & Connected Commerce
The common thread is a real-world constraint. Inventory, a gate, a production schedule, or a purchase order does not become simple because a model can talk about it.
Connected commerce, in this theme, means the transaction and the physical flow are one problem. Software that serves only the slide, and not the constraint, is out of scope.
Draft — not firm policy
Notes on what a relevant company looks like
The notes below are a draft for internal discussion. They are not an approved investment policy, a checklist that guarantees a meeting, or a description of a fund mandate.
- The company is aimed at a workflow in freight, logistics, transportation, supply chain, or an adjacent industrial market.
- Intelligence is part of how the operation runs, not only how the product is described.
- The founders can say which actions the system takes, which actions a person must take, and why.
- There is evidence of use in a real operation, even if the company is early. A concept with no path into live work is a weaker fit.
- Logitech Ventures could be a relevant partner. This site does not publish ownership targets, reserves, board expectations, or a required round role.
What this page will not pretend
Logitech Ventures does not publish fund names, regulatory registrations, a general check-size range, ownership targets, or performance on this site. Those facts are omitted because they have not been confirmed for publication, not because a reader should infer them. A confirmed investment in one portfolio company is not a statement of fund policy. The firm is not inviting the public to invest.
The portfolio page names one verified company: The Freight Dynamics, the brand of Freight Dynamics Inc. Logitech Ventures (Singapore) invested $5 million in the Series A of Freight Dynamics Inc. The investment was announced and closed on 20 October 2026. The capital is to optimize operations. That page does not state a valuation, name other investors, or say that Logitech Ventures led. The remaining slots are placeholders. A thesis is not a list of announcements.
Introduce a company or read the editorial essays.