← RESEARCH
Pilot intent

Autonomous logistics.

Robotic warehousing, drones, autonomous last-mile.

RESEARCH FRAME

Thesis

Commerce infrastructure eventually has to coordinate the movement of goods, not just pages and payments.

Why now

Fulfillment automation, delivery expectations, and merchant margin pressure are converging.

Unresolved work

Physical-world reliability, physical-world operations, and compliance make this a slower frontier.

Graduation signal

Logistics graduates through pilots with measurable operational improvement.

enes · 27 April 2026 · 2 min read

The problem

Physical fulfillment is the slowest layer of commerce. Software shipped instantly; goods still travel by truck. The gap between the agentic operations layer (which can take a buyer from intent to payment in under a minute) and the physical chain (which takes days) is now the dominant friction in online shopping. Same-day in major cities is a band-aid. The structural answer is autonomy in the chain itself.

Our angle

The chain has three layers that need to autonomize on different timescales:

  • Warehousing — robotic picking, packing, and SKU routing. Capital-intensive but mature; the question is integration with merchant operations.
  • Last-mile — the most expensive step per unit. Autonomous vehicle pilots in dense Turkish cities have a credible regulatory path; rural last-mile is harder.
  • Aerial — drone delivery for high-margin, time-sensitive verticals where the regulatory and physical constraints are tractable. Not a 2026 surface; explicitly a longer horizon.

The bet is that the same agentic platform that runs the operations layer can, in time, dispatch and supervise the physical chain end-to-end. Mercury (the logistics agent) is already routing across human-driven networks. The substrate is in place; what changes is what each agent can address on the other end of an API call.

What we're exploring

Pilot design for controlled logistics experiments. Regulatory paths in Türkiye for drones and autonomous delivery — what's the realistic 18-month and 36-month horizon. Cost models that make this viable for SMB merchants, not just enterprise. The right unit economics for cross-merchant pooling (a fastart.co network of 10,000 stores can pool fulfillment in a way no single merchant can).

Status

Pilot intent. Research and pilot design in progress; nothing customer-facing under FaStart's brand yet.

An invitation

If you've worked on warehouse robotics, autonomous vehicle pilots, or commerce logistics at scale — and you'd rather see the agentic operations layer extend into the physical world than watch it stop at the API boundary — we'd like to talk. research@fastart.tech.

← RESEARCH