A week at NADA is usually enough to kill a few assumptions and sharpen a few others.

This year’s overwhelming takeaway wasn’t about the death of dealerships, EVs or the US market — it was about defensibility.

Everyone is talking about AI. Far fewer are talking about what actually survives it.

That theme wasn’t just visible in private conversations or investor breakfasts — it was front and centre on the show floor. One of the most telling signals came from live operator commentary during NADA's Car Dealership Guy Podcast broadcasts: if a tool doesn’t clearly drive ROI, it’s on the chopping block.

That mindset framed much of what followed.

Data is the only moat that really matters.

One of the most consistent themes — echoed repeatedly at the Auto Tech Investment Breakfast and across conversations with operators, founders and investors — was that data is becoming the only truly defensible asset in automotive tech.

Software itself is being rapidly commoditised. No-code tools, “vibe coding”, and agentic AI are collapsing the cost and time required to build products. Even at the top end of the market, investors are openly questioning whether pure software businesses remain defensible over a five-year horizon.

What stood out this year was how explicit operators have become about this. As discussed live on the final day of NADA, technology is now judged on three things only:

• Does it drive measurable return?
• Does it improve process discipline?
• Does it scale across the group?

If it doesn’t do all three, it doesn’t survive budget season.

The counterargument matters here. Some believe AI agents will make proprietary data less valuable by synthesising public sources or generating convincing synthetic alternatives. The rebuttal from operators was clear: generic data produces generic recommendations.

What drives conversion is contextual data — trade cycles, local market behaviour, customer interaction history — that can’t be scraped or simulated. The businesses winning today aren’t just collecting data. They’re closing feedback loops that competitors can’t replicate.

You don’t need all the answers today — but you do need to be thinking about them now. Five-year moats matter again.

After-sales isn’t broken — but it remains under-optimised

Despite innovation elsewhere, the US after-sales still feels fragmented and misaligned with changing consumer behaviour:

• A majority of consumers continue to use independent garages
• Pricing transparency remains inconsistent
• Online booking demand is rising, but human interaction still matters

Dealers and technology providers are clearly focusing here. Marketplace-style models are emerging to sit in the middle — connecting drivers, workshops and OEMs with real-time availability, reporting and routing demand back to franchised dealers.

The opportunity is obvious. The friction is too:

• Dealers remain cautious around open pricing
• Fixed ops directors understandably protect capacity and throughput
• Scale on both sides of the ecosystem is non-negotiable

The strategic insight that stuck with me: dealers aren’t actually more expensive — but perception becomes reality. Any technology that reconnects consumers with franchised dealers while preserving operational control is quietly powerful.

Inspection, trust and workflow — not just self-service AI

Vehicle inspection is emerging as one of the most credible, high-ROI AI use cases in automotive — but not for the reasons many assume.

Self-inspection platforms are no longer just about cost reduction. Their real value lies in consistency, auditability and trust:

• More accurate damage and condition assessment
• Reduced inspection friction across logistics, remarketing and insurance
• Cleaner data flowing through downstream workflows

Captive insurers alone spend hundreds of millions annually on inspections, storage and transport — making this an obvious area for disruption. What stood out at NADA was how much more operationally mature these tools now feel, with deeper AI models and tighter integration into dealer and insurer workflows.

The broader shift is experiential rather than technical: the customer’s driveway is becoming the first point of engagement. Trust is increasingly built before a vehicle ever reaches a forecourt.

EVs, battery data and why the UK is ahead

This is one area where the UK clearly has more lived experience than the US.

A more mature EV parc, faster adoption driven by net-zero policy, and earlier exposure to used EV dynamics mean that UK operators are further along the learning curve. That gap was evident in conversations with U.S. investors and strategics — particularly around confidence, residual values and remarketing risk.

Battery health data is central to this.

Independent battery reporting — generated quickly and outside OEM ecosystems — is becoming a proprietary data asset in its own right. Platforms that demonstrate higher conversion rates do so by reducing uncertainty, not by lowering prices.

There was also notable U.S. interest in the growth of Chinese EV brands in the UK — seen as a live test market for how new entrants scale, price and position outside domestic protection.

What U.S. operators should do with this: partner now with UK-based battery data providers before the market forces the issue. Residual risk on EVs will compound faster than most groups are pricing for. The UK’s mistakes — and solutions — are a free look ahead. Use it.

Battery data may be one of the most underappreciated moats I saw all week.

The metal is the smallest part of the deal.

This came up repeatedly, and it’s worth stating plainly: the car itself is becoming the least interesting part of the transaction.

Lifetime customer value, finance, insurance, energy services, data and engagement increasingly define dealership economics. The most forward-thinking operators now regard dealerships as energy and mobility hubs, not merely retail sites.

The common traits of winners were consistent:

• Deep CRM and process discipline
• Video everywhere — increasingly AI-personalised
• Technology that embeds operationally, not cosmetically

Build, buy or partner — but don’t be lazy

Technology strategy can’t be treated like a standard asset class.

• Build delivers IP, margin and exit value — but costs time and carries blind-spot risk
• Buy accelerates capability, but integration risk is real
• Partner can be powerful, but diligence is often weaker than M&A — and shouldn’t be

A recurring warning was clear: the most significant risk in building is failing to understand what the dealer actually needs. Feedback loops are non-negotiable.

M&A, liquidity and sentiment

Despite the doom-laden AI narrative, deal activity feels anything but stalled.

While 2025 closed solidly, the more interesting signal from NADA was forward-looking: 2026 is shaping up to be an active year, based on anecdotal pipelines shared by strategics, advisors and investors across the week.

The bar has clearly shifted:

• Early stage still rewards growth
• Later stage demands growth and unit economics
• Moats are scrutinised harder than ever

Cash remains king. Headline valuations mean little without understanding structure, earn-outs and downside protection. Preparation still wins.

What this means in practice

For boards and leadership teams, the question in 2026 isn’t “Do we need an AI strategy?” — it’s:

• Which data assets, workflows and customer relationships would still matter if AI became free tomorrow?
• Any technology investment that doesn’t strengthen those three is tactical, not strategic — and should be treated accordingly.

The real message from NADA

AI isn’t killing automotive retail — it’s stress-testing it.

The winners will:

• Control proprietary data
• Embed operationally
• Move fast, test and scale
• Build businesses that are AI-proof, not AI-dependent

Sentiment is green for now. Cycles turn quickly.

The work of building defensibility, integration and exit readiness isn’t something you do in a downturn — it’s what determines whether you survive the next one. Have a great week!