Garrett Scott, CEO of Pipedream, is building “hyperlogistics” — a system to deliver anything in under 10 minutes for less than a quarter, enabling people to treat physical objects like data: order when needed, use, then send back. He sees this as the next major leap in commerce, comparable to how streaming transformed music, and believes it will be the single largest driver of GDP growth in the coming decades.
The Vision: Hyperlogistics
Garrett defines hyperlogistics as sub-10-minute delivery at negligible cost (under ~$0.25), with equally easy bidirectional sending — you receive and return items as effortlessly as downloading and deleting files.
The core idea is that once delivery becomes nearly free and instant, consumer behavior shifts dramatically:
People stop stockpiling (no more bulk toothbrushes) and instead get higher-quality items on demand.
Clothing becomes a subscription: order many items, try them at home, send back what you don’t want, and re-rent favorites later.
Pantries disappear — you real-time chips, eggs, and groceries instead of storing them.
Renting high-quality goods (electronics, tools, specialty items) becomes far more economical than buying them outright.
The “Uber for screwdrivers” problem — where delivery costs exceed the item’s value — is exactly what hyperlogistics solves by collapsing cost and time.
From Idea to First Store
After college, Garrett took two years to find a problem worth solving for decades. He kept returning to the magic of instant fulfillment — the goosebumps from his first Postmates order in 2015, the wonder of Uber, the amazement of Amazon’s next-day delivery.
He recognized that despite incremental improvements (Amazon, Uber Eats, DoorDash), there remains a massive gap between today’s logistics and what’s possible with full autonomy.
Pipedream’s original pitch deck — an underground robot network moving totes at 100 mph to above-ground pickup kiosks — is essentially what they’re building now, but the path required a cold-start strategy.
The Technology: Otters, Pipes, and Portals
Otters: Autonomous robots carrying standard Walmart-sized totes (~99.6% of items fit, including 60 lb dog food and 48-pack water) through underground pipes at 100 mph.
Pipes: Underground conduit installed via horizontal directional drilling (no trenching), permitted through well-worn municipal processes — typically approved in 2–4 weeks, far easier than drone airspace permits.
Portals: Above-ground kiosks (like ATMs for goods) where customers drive up, grab groceries, and leave — targeting a 15-second interaction with zero phone friction.
Three lanes per site, with RFID windshield tags for first-time users so orders are ready before the car stops.
Designed as handoff points for humans, delivery drivers, drones, or sidewalk robots.
Why Underground?
Underground infrastructure has decades of established permitting pathways. Cities are comfortable with it because horizontal directional drilling causes minimal surface disruption and can run 24/7.
Garrett contrasts this with drones, which are still in a “valley of disillusionment” despite millions of deliveries already happening annually — he believes drones are underestimated by one to two orders of magnitude in impact.
Pipes are low-risk once buried: “nothing can hurt anybody.”
The First Rapid Fulfillment Center (RFC)
Pipedream’s first RFC is in a leased warehouse north of Austin, chosen for high daily traffic (30,000+ pass by) and proximity to the right demographic.
The challenge: fulfillment centers aren’t built in retail locations. They’re zoned for high-value development (hotels, etc.), and landlords don’t want to lock space for 5 years with a startup doing something “crazy.”
After ~3 weeks of negotiation, they secured the space and are now building out:
A canopy over three portals in the parking lot (using the same aluminum material as Starbucks, nearly scooped up by the Tennessee Titans stadium).
A retro neon sign and meticulously designed brand experience called Goods (not PipeDream — “horrible name for a grocery store”).
Full cold chain (freezers, refrigerators) for fresh and ready-to-eat foods.
Goods: The Anchor Tenant Strategy
Pipedream initially planned to be pure infrastructure, partnering with an existing retailer as anchor tenant. They worked with every major retailer through an “instant pickup” product deployed at Steak ‘n Shake and other locations.
They realized incumbents are too slow and not customer-obsessed enough to exploit what autonomous logistics enables. The real value is in owning the full customer feedback loop.
So Pipedream is launching its own grocery brand, Goods, as the anchor tenant — stocking 42,000 of the most commonly carried grocery SKUs (staples, fresh produce, branded goods like Oreos, Doritos, Coca-Cola, packaged meats).
Break-even is 42 orders per portal per day — a number they hit on the nose, which Garrett finds cosmically fitting (Hitchhiker’s Guide to the Galaxy reference).
Solving the Selection Problem
A 3-week guerrilla test — scraping 7-Eleven and CVS inventory, offering free 15-minute delivery, handing out hundreds of flyers — yielded exactly one order. The lesson: if you’re missing even one item a customer wants, you lose them entirely.
This mirrors Walmart’s strategy of stocking 130,000+ SKUs where some items exist solely to draw in one specific customer.
Pipedream’s solution is multi-cart: everything in the RFC is available in 5 minutes; once daily, they batch-order items from other stores (Costco, H-E-B, Whole Foods, specialty grocers) and consolidate them at the portal for pickup.
This is analogous to early Amazon’s strategy of listing millions of titles but only stocking a fraction, drop-shipping the rest.
It solves the “I have to go to H-E-B for this one thing” problem that locks customers into specific stores.
Designing the Customer Experience
Garrett and co-founder Cannon obsess over making the experience feel human and joyful despite being fully automated:
Product photos are redesigned to look gorgeous and homogeneous across brands, reducing eye strain.
The app lets you add items to cart throughout the day without checking out — when you drive near a portal, it recognizes your RFID tag and prepares your order automatically. (This was scrapped for launch because people need the psychological comfort of hitting “checkout.”)
Every detail — lighting, canopy material, neon signage, product photography — is crafted to feel like “someone really cared.”
Garrett’s philosophy: early adopters should hate the robot. If people say “I just drive up, grab my groceries, and go — it’s 10 minutes faster than anything else,” the experience has succeeded. If they say “a cool robot brings it to you,” the novelty is doing the work, not the value.
The Team: Talent Density Above All
Pipedream has stayed at ~12 people, prioritizing talent density over headcount.
For their first eight hires, the bar was “embarrassing that they joined” — each person had to be so good that existing team members looked up to them.
They lowered the bar slightly under hiring pressure and immediately felt the difference: “A minuses kill you.” They course-corrected by holding every new hire to the standard of the best person already on the team.
Cannon (co-founder) balances Garrett’s aggressive ideation with grounded execution judgment. He’s described as extraordinarily ethical — the kind of person who puts his girlfriend’s bathrobe in the dryer before her shower.
Ahmed Shubber was hired into a “wildcard” role open for 18 months. He arrived unprompted with high-quality renders of what Pipedream was doing wrong and how to fix it. His negotiation tactics (reverse-ranking contractors by review count, calling down the list, pushing prices to floor) became internal playbook material. Garrett calls him “the next Elon.”
Peyton (head of product) designed the Goods brand, the canopy, the product photo system, and the overall customer experience with obsessive attention to detail.
Company Values
“Be good partners” is the #1 company value. When a construction crew Pipedream had hired hit a gas line during the Georgia pilot (taking out power to an entire strip mall), Garrett had someone at every business within 15 minutes, collecting daily revenue figures and paying them that same night — for an outage that lasted only 3 hours.
The goal: “How do we fix this so people hope the problem happens again?”
“Give a shit” (Cannon’s phrase) is the most valued trait. People who care deeply about everything — code quality, customer experience, workspace cleanliness — catch things others miss.
Garrett values people who challenge ideas freely, even when socially uncomfortable. The best hires walk in and say, “Here’s what you’re doing wrong.”
Scaling the Network
Once a portal hits 42 orders/day break-even with strong retention and word-of-mouth (targeting a 1:2 ratio — each customer brings two more within three months), they extend the network in 5-mile loops from the initial RFC.
Early portals will be close together (unlike Chipotle’s strategy of maximizing distance between stores), because the network’s value comes from density and connected inventory.
As more portals are added, the RFC can stock more items and offer more services, creating a flywheel: more selection → more customers → more volume → more portal locations → more selection.
The long-term vision is to be the fulfillment layer (3PL) for other retailers and brands (like David Protein Bar), offering the fastest pickup and cheapest delivery by combining underground autonomous transport with multiple last-mile modalities (drones for suburbs, sidewalk robots for dense areas, portals for drivers).
What Could Go Wrong
Garrett’s honest assessment: the biggest risk is financing the business from infancy to the crawling stage — growing at the right pace without overextending.
He’s not worried about retailing itself (he and Cannon are lifelong retail obsessives who study Sam Walton, Jeff Bezos, Costco, Amazon, Ocado, and GoPuff deeply).
The biggest surprise has been how slow even “startuppy” incumbents are to adopt autonomous logistics — despite the technology already being here (self-driving cars operating, millions of drone deliveries per year). Fortune 10 partners struggle just as much as Pipedream does.
Timeline
54 days from recording: Goods opens its first portal location in Austin, with customers picking up grocery orders.
Q4 2025: Ready-to-eat meals launch.
Next 1–2 years: Network expansion in Austin through 5-mile underground loops, with additional RFCs and portals.
