AI Investment Under Pressure: Uber’s Robotaxi Lesson

Uber just gave a very honest snapshot of what AI-led transformation looks like in the real world: demand is up, margins are under pressure, taxes are rising—and the company is still putting money into autonomous vehicles.

That tension is the point. Most businesses say they want AI, but the moment costs rise or profitability dips, AI becomes “next year’s project.” Uber is doing the opposite: it’s protecting affordability in the short term while funding a capital-intensive bet (robotaxis) that only pays off at scale.

For Singapore leaders following this AI Business Tools Singapore series, the headline isn’t “robotaxis are coming.” It’s this: AI adoption is a portfolio decision, not a single project. You’re balancing today’s unit economics against tomorrow’s structural advantages.

(Article source: https://www.channelnewsasia.com/business/uber-pushes-robotaxi-plans-even-cheaper-rides-higher-taxes-dent-profit-5907091)

What Uber’s robotaxi push is really saying about AI strategy

Uber’s update contains three signals that matter for any company evaluating AI tools for operations, marketing, or customer engagement.

First, demand can be strong while profits still disappoint. Uber reported 22% trip growth in Q4, driven by more people choosing shared rides and lower-cost mobility products—good for growth, tougher on margins.

Second, AI bets often require spending before they reduce costs. Uber is committing capital to vehicle partners to secure early autonomous vehicle (AV) supply and speed up deployment. That’s not “buy software and save money next month.” It’s “invest now to change the cost curve later.”

Third, scale is the advantage. Uber argues that a multi-product platform can drive better AV economics: higher utilisation, shorter pickup times, better matching. That’s a familiar AI story—models and automation don’t shine in pilot mode; they shine when integrated into workflows and demand.

The takeaway for Singapore businesses

If you’re running a SME or mid-market team in Singapore, you probably don’t have Uber’s balance sheet. But you do have the same decision structure:

• Protect today’s customer experience and prices
• Improve internal efficiency
• Keep investing in AI capabilities that compound over time

The winners don’t “go all-in” on AI everywhere. They pick a few high-leverage use cases, then scale what works.

The cost-vs-innovation tradeoff (and how to make it less painful)

Uber’s profit was hit by two very normal forces: cheaper rides (pricing pressure and affordability push) and higher taxes. Uber also warned it expects a 22% to 25% effective tax rate in 2026, partly because it operates across 70+ countries.

That’s not unique to ride-hailing. In Singapore, businesses face their own mix of cost pressure:

• Higher wages and tighter hiring
• Rising vendor costs
• Compliance and reporting demands
• Customer expectations for fast response and personalised service

Here’s my stance: if AI only works when times are good, it’s not a strategy—it’s a hobby. The right AI business tools reduce friction when the business is under stress.

Practical ways to evaluate AI tools when budgets are tight

Use a simple decision filter that mirrors what large firms do, just scaled down.

1. Time-to-value (30–90 days)

   • Can you see measurable impact within one quarter?
   • Example metrics: response time, conversion rate, invoice cycle time, hours saved.

2. Unit economics impact

   • Does it reduce cost per lead, cost per ticket, cost per report?
   • If it only improves “quality” without a measurable business metric, it’s harder to defend.

3. Adoption cost (not license cost)

   • Most AI tools fail because teams don’t use them.
   • Budget for process redesign, training, prompts/templates, QA checks.

4. Data readiness

   • If your CRM is messy or your FAQ content is outdated, your AI output will be messy too.

This is how you keep AI investment rational—without killing ambition.

Uber’s “platform advantage” is a blueprint for AI in operations

Uber’s CEO pointed to higher utilisation and shorter pickup times on Uber’s platform compared to standalone robotaxi services. That’s a platform story, but it maps cleanly to how AI operations should work inside a business.

A single AI pilot in a corner doesn’t create an advantage. Advantage shows up when AI is connected to your core systems and your daily decisions.

What “platform thinking” looks like for SMEs in Singapore

You don’t need a super-app. You need a connected toolchain:

• CRM as the source of truth (customers, deals, lifecycle stage)
• Helpdesk/WhatsApp logs for customer intent and issues
• Accounting/invoicing for cashflow visibility
• A knowledge base (SOPs, product info, policies)
• AI layer that automates and assists across these systems

When those pieces talk to each other, AI becomes practical:

• Sales gets suggested follow-ups based on deal stage and last interaction
• Support gets draft replies grounded in your own policies
• Ops gets exception alerts (late shipments, churn risk, payment delays)

Uber is betting that AVs plugged into its platform perform better than AVs floating alone. Same idea: AI works better when it’s embedded into workflows, not bolted on.

Robotaxis as a case study in “scale-first” AI (and why pilots disappoint)

Uber said it plans to facilitate robotaxi trips in up to 15 cities by end-2026, and expand to Madrid, Hong Kong, Houston and Zurich, with Hong Kong positioned as its first autonomous ride market in Asia.

That timeline matters because it shows how long these bets take even for a global company. And it explains a common AI adoption frustration in smaller firms:

“We tried AI, it didn’t work.”

Often what actually happened is:

• The use case was too broad (“automate marketing”)
• The data wasn’t prepared
• The workflow didn’t change, so the tool became optional
• Nobody owned performance metrics

A better way to run AI adoption: pick 1 workflow, then scale

If you want a reliable approach for 2026 planning, use this three-step rollout:

1. Choose one repeatable workflow with volume

   • Examples: inbound lead qualification, appointment scheduling, quotation follow-ups, invoice reminders, customer FAQs.

2. Define the “gold standard” output

   • A good AI tool needs a clear target: tone, policy constraints, escalation rules.

3. Instrument the workflow

   • Track: handle time, error rate, CSAT, conversion rate, backlog.

Uber’s robotaxi plan isn’t about proving the tech exists. It’s about getting enough real-world volume that economics improve. Your AI tools should follow the same logic: volume + integration = ROI.

The financing angle: what Uber’s AV approach teaches about risk

One of the most interesting lines in the article is that Uber is working with banks and private equity firms to finance most of the autonomous fleets, while still committing some capital to secure supply.

That’s sophisticated risk management: Uber wants exposure to upside without carrying the full capital burden.

The SME version: don’t over-buy, don’t under-invest

Singapore businesses can copy the principle even without fancy financing:

• Prefer month-to-month or annual SaaS over big upfront builds
• Start with assisted automation (human-in-the-loop) before full autonomy
• Use templates, playbooks, and guardrails before custom model training
• Keep an “AI runway” budget line for experimentation, but cap it

A good rule I’ve seen work: spend 70% on proven workflow improvements, 20% on adjacent experiments, 10% on long-shot bets.

Uber is doing a version of that—protecting core demand growth, managing near-term profit, and still funding AV scale.

A simple AI business tools checklist (Singapore-ready)

If you’re deciding what to implement this quarter, here’s a practical checklist you can run in one meeting.

1) Customer engagement tools

• AI-assisted replies for email/WhatsApp/helpdesk
• Lead routing and qualification based on intent
• Conversation summaries pushed into CRM

2) Operations automation

• Document extraction for invoices, POs, delivery orders
• SOP copilots for staff (“how do I process a return?”)
• Exception detection (late orders, unusual refunds, churn signals)

3) Management visibility

• Weekly KPI narrative (what changed, why it changed, what to do next)
• Forecast support (pipeline health, staffing needs)

4) Governance and safety

• Access control (who can generate what)
• Audit trails for regulated workflows
• Approved knowledge base sources (to reduce hallucinations)

If you can’t answer “who owns this workflow and metric?” pause before you buy.

Where this leaves Singapore businesses in 2026

Uber’s story is not a robotaxi story. It’s a leadership story about committing to AI and autonomy while the P&L is messy. Trips up 22% shows demand. Profit pressure and a 22%–25% tax rate outlook shows reality. And the continued robotaxi investment shows intent.

For the AI Business Tools Singapore series, the message is practical: your AI plan should survive a tough quarter. If the only time you invest is when margins are fat, you’ll always be late.

If you want a next step, treat your AI roadmap like Uber treats AVs: pick a handful of workflows where AI reduces cost or increases throughput, integrate it into the platform you already run on, and measure it hard. Then scale.

What would change in your business if you could reliably handle 20% more customer volume—without hiring 20% more people?