EV Sales Slump, New Launches: A Playbook for AI Adoption

A U.S. market can go from “hot” to “hard” fast. On April 1, Reuters reported from the New York Auto Show that automakers are still rolling out new EVs even after a sharp demand drop triggered by the end of the US$7,500 EV tax credit. EV sales were 9.6% of U.S. vehicle sales in 2025, but fell to 6.5% in the last three months of the year—the lowest since early 2022.

Most companies get this wrong: when demand dips, they freeze innovation and hope conditions improve. The automakers doing the opposite—shipping new models, adjusting pricing, and rebalancing EV/hybrid bets—are basically running the same playbook Singapore firms need right now with AI business tools.

This matters because the EV story isn’t really about cars. It’s about what to do when incentives disappear, customers hesitate, and your cost base doesn’t magically shrink. If you’re leading a team in Singapore, the parallels to AI adoption are obvious: the winners aren’t the ones with the loudest announcements; they’re the ones who build capabilities that survive a downturn.

Source article: https://www.channelnewsasia.com/business/automakers-unveil-new-evs-us-market-despite-sales-downturn-6031371

What the EV downturn actually signals (and why it’s not “EVs are dead”)

The clearest signal from the New York Auto Show coverage is that demand changed shape—it didn’t vanish forever.

Yes, Nissan’s Americas chair Christian Meunier was blunt: “When you look at the EV market right now, there’s no demand.” He also noted that a meaningful portion of remaining sales is being pushed by heavy incentives, meaning demand is more fragile than many projections assumed.

At the same time, rising gasoline prices are already nudging interest back. Hyundai’s CEO Jose Munoz pointed to increased EV sales in places like California driven by market conditions, not regulation. That’s a key detail: customer behavior can flip quickly when operating costs change.

Here’s the business takeaway for AI in Singapore: if your AI adoption plan only works when “everyone is excited about AI,” you don’t have a plan. You have a vibe.

The underlying pattern: incentives distort reality

When a market is propped up by incentives—tax credits, subsidies, promotional pricing—it’s easy to confuse stimulated demand with product-market fit.

In the U.S., the EV tax credit acted as a demand accelerant. Once removed, automakers had to confront what customers truly value:

• Total cost of ownership (vehicle price + charging/fuel + maintenance)
• Convenience (charging access, time, reliability)
• Trust (battery life, resale value, safety)

For Singapore businesses investing in AI business tools, the “incentive” might be internal hype, a competitor’s press release, or a C-suite mandate. When that fades, what’s left is whether AI actually improves:

• Lead conversion rates
• Service response time
• Forecast accuracy
• Cost per transaction

Why automakers keep launching new EVs in a slump

Launching new products during a downturn sounds reckless. It isn’t—if you understand the strategic math.

Automakers are shipping new EVs because:

1. Product cycles don’t pause. Waiting 18–36 months can cost an entire generation of customers.
2. The cost curve improves over time. Batteries, supply chains, manufacturing, and software platforms get cheaper and more reliable with scale.
3. Segmentation matters. A “slump” often means the early-adopter segment is saturated, not that the mainstream segment is unreachable.

At the show, Kia said it will begin selling the lower-priced EV3 in the U.S. later this year. Subaru revealed a new three-row EV (“Getaway”) seating seven. GM is back in the conversation with the Chevrolet Bolt EV starting at US$27,600.

That’s not random. It’s a shift toward accessible pricing and practical use cases.

Singapore parallel: the AI tools that win budgets in 2026 aren’t the most impressive demos. They’re the ones tied to a concrete workflow—sales follow-ups, invoice reconciliation, compliance checks, customer support triage.

A stance worth taking: “Wait and see” is the most expensive strategy

I’ve found that businesses often justify delay as “risk management.” But in fast-moving categories, delay is usually capability debt.

Automakers that stop launching EVs don’t save money long-term—they lose engineering momentum, supplier negotiating power, and customer mindshare.

For Singapore companies, “we’ll adopt AI when it stabilises” often turns into:

• Talent gap (no one internally knows how to run AI projects)
• Data mess (no clean inputs, no governance)
• Vendor lock-in panic (buying the wrong tool in a rush later)

A better move is controlled experimentation: small bets, measurable outcomes.

Three EV-to-AI lessons Singapore businesses can use this quarter

The EV market data gives a clean framework for making AI adoption decisions without getting pulled by hype.

1) Build for “natural demand,” not incentives

In the Reuters piece, Meunier said roughly half of the EV market is driven by heavy incentives—so it’s not “natural demand.” That line should be printed and taped to every strategy deck.

Natural demand in AI looks like this:

• Your sales team uses the tool even when no one is watching
• Customers notice faster, clearer replies
• Finance trusts the numbers without manual rework

How to operationalise it (simple but strict):

• Pick one workflow with clear before/after metrics
• Set a 30-day evaluation period
• Require adoption by the frontline team, not just management

If the tool needs constant pushing to be used, treat it like an EV that only sells with rebates: it’s a warning sign.

2) Compete on total cost of ownership (TCO), not “features”

EV buyers don’t just ask “How fast is it?” They ask: “How much will this cost me over 3–5 years?”

AI tools should be evaluated the same way. The subscription price is the smallest line item. The big costs are usually:

• Staff time to clean data and maintain prompts/workflows
• Integration work with CRM/accounting/helpdesk
• Risk controls (privacy, access permissions, audit trails)
• Rework when outputs aren’t reliable

A practical TCO checklist for AI business tools in Singapore:

1. What does it cost per month and per active user?
2. Who owns model outputs (and can you export your data)?
3. What’s the fallback when the tool is wrong?
4. How do you monitor quality drift over time?

If you can’t answer these, you’re not buying a tool—you’re buying ongoing uncertainty.

3) Expect a “hybrid era” in both EVs and AI

Hyundai said it has revised plans to include more hybrids, and Munoz projected EVs may rise step-by-step to around 10–15% rather than 50–60%.

That’s not failure. It’s an adoption curve that matches infrastructure, pricing, and real customer constraints.

AI adoption inside companies works the same way. Most teams won’t go “full AI.” They’ll go hybrid:

• AI drafts, humans approve
• AI triages, humans resolve
• AI forecasts, humans sanity-check

This is the steady, profitable version of AI. It’s also the version that survives compliance and customer scrutiny.

A simple “AI adoption roadmap” inspired by the automakers

If you’re reading this as part of the AI Business Tools Singapore series, here’s the roadmap I’d use if I were advising a mid-sized company making its first serious AI move.

Step 1: Pick one use case where delay is painful

Automakers feel pain immediately when they miss a product cycle. You should pick an AI use case where the pain is visible.

Good candidates:

• Customer support backlog (response time, SLA breaches)
• Sales lead follow-up lag (speed-to-lead, conversion)
• Monthly reporting (days to close, manual consolidation)

Step 2: Run a 6-week pilot with measurable gates

Define three metrics upfront:

• Speed metric (time saved, response time)
• Quality metric (error rate, customer satisfaction)
• Adoption metric (active users, workflow completion)

If you don’t hit at least two, stop or redesign. Automakers kill trims and reprice models all the time. Treat AI pilots with the same discipline.

Step 3: Standardise what works (don’t keep it as a “project”)

The biggest AI failure mode I see is leaving success stuck in a sandbox.

Operationalise it:

• Write a short SOP: inputs, outputs, approval rules
• Add permissions and audit logs
• Assign ownership (one team, one accountable lead)

That’s the equivalent of moving from prototype to production.

People also ask: “Is this a bad time to invest in EVs or AI?”

Answer: It’s a bad time to invest in vague promises. It’s a good time to invest in capabilities tied to costs and cashflow.

EVs: tax credits ended, demand dipped, but automakers are still launching because product cycles and cost curves keep moving.

AI: budgets are tighter, scrutiny is higher, but tools that reduce handling time, improve conversion, or cut rework are easier to justify than ever.

A downturn is where disciplined adopters pull ahead, because competitors hesitate.

What to do next (if you want AI to pay off in 2026)

The New York Auto Show story ends with an industry adjusting to reality: fewer subsidies, more price sensitivity, and an “evolution” toward EVs rather than an overnight takeover. That’s not doom. That’s what real markets look like.

For Singapore businesses, the smart move is to treat AI business tools the way automakers treat new platforms: invest steadily, focus on practical segments, and measure what customers (and staff) actually do—not what they say they’ll do.

If you’re building your 2026 operating plan, here’s the question worth sitting with: Which workflows will you still be running manually when your competitors have already standardised them with AI?