Robotaxi & AI Operations: What Uber’s Bet Signals

Trips on Uber rose 22% in Q4, yet the company still warned of lower-than-expected near-term profit and a 22%–25% effective tax rate in 2026. That combination—strong demand, margin pressure, and regulatory complexity—is exactly why Uber is still spending on autonomous vehicles (AVs) and robotaxis.

Most companies get this wrong: they treat AI as a “nice to have” that gets cut when profits tighten. Uber’s latest results (and its robotaxi push) show the opposite approach—investing in AI-powered operations while simplifying the unit economics elsewhere. For businesses in Singapore thinking about AI adoption—especially in logistik dan rantaian bekalan (logistics and supply chain)—this is a practical case study, not sci-fi.

Robotaxis aren’t just about passenger rides. They’re a preview of how AI will reshape fleet utilisation, route optimisation, dispatching, forecasting demand, and cost-to-serve. The same patterns apply to delivery fleets, cold-chain distribution, service scheduling, and warehouse-to-door fulfilment.

Uber’s robotaxi plan is really a supply chain strategy

Uber’s headline is “robotaxis”, but the underlying play is a capacity and supply strategy.

According to the Reuters report carried by CNA, Uber is committing capital to vehicle partners to secure early supply and speed up deployments, while working with banks and private equity to finance most of the autonomous fleets. That’s not a moonshot mindset; it’s classic supply chain thinking:

• Secure constrained supply (early vehicle access)
• Diversify suppliers (multiple AV partners)
• Shift capex off the balance sheet (third-party fleet financing)
• Optimise utilisation (keep assets earning)

In logistics terms, Uber is trying to avoid the AV equivalent of a component shortage: if autonomous fleets arrive slower than demand, you lose market share to whoever has vehicles on the road.

Why utilisation matters more than the “self-driving” headline

Uber’s CEO Dara Khosrowshashi highlighted that vehicles operating through Uber’s platform have achieved higher utilisation and shorter pickup times than standalone robotaxi services. That’s a big statement because utilisation is where unit economics live.

A simple rule I’ve found useful across transport and fulfilment: If an asset isn’t moving (or being used), it’s costing you twice—once in fixed cost, and again in missed throughput.

Robotaxis are the extreme version of this. The vehicles are expensive; the software is expensive; compliance is expensive. So the only way it works is:

1. Keep the vehicle busy (high utilisation)
2. Minimise deadhead time (empty distance)
3. Reduce pickup times (dispatch optimisation)

That’s the same trio that matters for:

• Last-mile delivery fleets
• Field service scheduling
• Cross-dock operations
• Linehaul planning

Margin pressure is why AI operations becomes non-negotiable

Uber’s profits were dented by two very “real world” forces:

• Cheaper ride products (shared rides and lower-cost options to expand the user base)
• Higher taxes (operating across 70+ countries)

When prices soften and costs rise, the only sustainable response is operational efficiency. For logistics and supply chain teams, this is familiar:

• Customers demand faster delivery but won’t pay more
• Labour costs rise
• Fuel and compliance fluctuate
• Service-level agreements don’t budge

AI becomes useful here when it’s not treated as a chatbot project, but as an operations system.

What “AI in operations” actually means (beyond hype)

AI in logistics and supply chain is most valuable in three areas:

1. Prediction: demand forecasting, ETA prediction, risk prediction (delay, damage, churn)
2. Optimisation: route optimisation, load planning, slotting, inventory positioning
3. Automation: exception handling, document processing, dispatch decisions, customer updates

Robotaxi deployment is basically all three at once.

• Predict demand by neighbourhood and hour
• Optimise dispatching and repositioning
• Automate decisions that used to require human judgement

Robotaxis expand the market by lowering the “cost of reliability”

Uber argues robotaxis will expand the mobility market by adding supply, improving reliability and lowering prices—leading to higher trip volumes.

The key phrase is “improve reliability.” In transport and fulfilment, reliability has a hidden price. If your service is unreliable, you pay for it via:

• Buffers (extra vehicles, extra labour, extra inventory)
• Customer support overhead
• Refunds, credits, service recovery
• Lost repeat purchases

A reliable autonomous fleet, if it arrives, reduces that reliability tax.

The supply chain analogy Singapore operators will recognise

Think of robotaxis like a new kind of capacity layer—similar to adding a flexible 3PL partner or micro-fulfilment nodes. It’s not just more volume; it’s more predictable throughput.

If you’re managing delivery or dispatch in Singapore, the “robotaxi lesson” is:

Reliability is a cost centre until you engineer it into the system. AI is one of the few tools that can do that at scale.

What Singapore businesses can copy (without owning a fleet)

You don’t need robotaxis to benefit from the same AI playbook. Most Singapore SMEs and mid-market firms can get real wins by applying AI to bottlenecks that show up every week.

1) Use AI to reduce empty time and empty distance

Robotaxis are all about minimising idle vehicles and deadheading. For logistics, this becomes:

• Dynamic route planning that updates with traffic and job changes
• Smarter batching for multi-drop routes
• Predictive repositioning (move drivers/vehicles before the surge)

Practical starting point: take 4–8 weeks of historical dispatch data and calculate:

• Average idle time per vehicle/day
• Empty kilometres as % of total
• Late deliveries as % of total

If any of those are high, optimisation will pay back fast.

2) Automate exception handling before you automate everything else

Many teams jump to “full automation” and get burned. The faster win is exceptions:

• Late pickup
• Failed delivery
• Address issues
• Damaged parcel
• Customer not available

AI can triage, prioritise, and propose the next best action (reschedule, reroute, notify, refund rules). This reduces support load and improves customer experience.

3) Forecast demand in a way operations can actually use

Forecasts that don’t change staffing, inventory, or vehicle planning are just dashboards.

A useful operations forecast answers:

• How many jobs tomorrow by zone?
• What’s the peak hour window?
• How many drivers/packers do we need?
• Which SKUs should be pre-positioned?

Even a modest forecast improvement can cut overtime and reduce late orders.

4) Structure financing and partnerships like Uber is doing

Uber is leaning on banks and private equity to finance fleets. In supply chain terms: don’t fund everything yourself.

For Singapore companies, the parallel is:

• Lease vehicles vs buy
• Use on-demand warehousing in peak seasons
• Partner with integrators that offer AI route optimisation as a service
• Negotiate outcome-based contracts (e.g., SLA + cost-per-drop)

The stance here is simple: keep your balance sheet flexible while you test what works.

A quick “AI readiness” checklist for logistics and supply chain teams

If you want AI outcomes (not experiments), treat it like a deployment. Here’s a field-tested checklist.

Data and workflow basics

• Single source of truth for orders, jobs, and statuses
• Clean location data (postal code, coordinates, delivery windows)
• Clear definitions (what counts as “late” or “failed”)

Operational ownership

• Named owner from ops, not just IT
• One metric that matters (on-time %, cost-per-drop, utilisation)
• Weekly review loop (what the model changed, what humans overrode)

Risk and governance (especially relevant in 2026)

As Uber’s tax-rate warning hints, operating across jurisdictions raises compliance costs. For Singapore firms handling cross-border deliveries and data:

• Document what data is used and why
• Put humans-in-the-loop for high-impact decisions
• Track model drift during seasonal peaks (CNY, Ramadan, year-end)

AI fails quietly when governance is vague. Define accountability early.

What Uber’s CFO change signals (and why operators should care)

Uber announced CFO Prashanth Mahendra-Rajah will step down, with Balaji Krishnamurthy taking over. Leadership changes are normal, but the timing matters when a company is balancing:

• Near-term profitability targets
• Large, long-horizon AI investments
• Multi-country tax and regulatory complexity

For operations leaders, the lesson is that AI programmes live or die on finance alignment. If finance only sees costs and not throughput gains, AI gets labelled “experimental” forever.

A stronger internal approach is to present AI like a capex decision:

• Baseline current cost-to-serve
• Model savings from utilisation and exception reduction
• Run a pilot with a hard stop date
• Scale only when the unit economics work

Where this is heading in Asia (and why Singapore should pay attention)

Uber plans to facilitate robotaxi trips in up to 15 cities globally by end-2026, including Hong Kong as its first autonomous ride market in Asia.

Hong Kong is a meaningful signal for Singapore operators because it’s dense, regulated, and operationally complex—more similar to Singapore than sprawling, car-centric cities. If robotaxi operations can work reliably in that kind of environment, it strengthens the case that AI-driven dispatch, routing, and fleet orchestration will become standard across urban logistics too.

For this “AI dalam Logistik dan Rantaian Bekalan” series, I’d frame it this way:

• Robotaxis are one visible outcome of AI operations
• The transferable value is in how the system is run: prediction, optimisation, automation
• The winners will be firms that improve reliability while keeping cost-to-serve under control

Next step: treat AI as an operations upgrade, not a tech trend

Uber’s quarter shows a reality Singapore businesses already feel: demand can be strong and margins can still get squeezed. That’s when AI tools earn their keep—by tightening routing, improving forecasting, reducing exception workload, and increasing asset utilisation.

If you’re looking at AI for logistics and supply chain, don’t start by copying the robotaxi. Copy the discipline behind it: secure capacity, measure utilisation, automate exceptions, and finance growth without trapping cash.

The question worth asking your team this week: Which part of your delivery or dispatch operation is paying the biggest “reliability tax”—and what would it be worth to reduce it by 10%?

Source context: CNA/Reuters report on Uber’s results and robotaxi plans (04 Feb 2026): https://www.channelnewsasia.com/business/uber-pushes-robotaxi-plans-even-cheaper-rides-higher-taxes-dent-profit-5907091