In the first half of 2026, buyers chose the Honda Prologue over the Chevrolet Blazer EV by a margin of 2.6 to 1. Under the sheet metal, these two midsize electric SUVs are virtually identical: both are built on General Motors’ Ultium platform, sharing the same chassis, battery packs, and electric powertrains, rolling off the exact same assembly line.
What created this staggering 165.5% sales divergence was a single software distinction. Honda retained Apple CarPlay and Android Auto smartphone mirroring; General Motors intentionally stripped the feature out. In doing so, the automotive industry inadvertently staged one of the cleanest A/B tests in its history. Two years of hard sales receipts have now attached a painful price tag to GM’s ambition of building a proprietary software walled garden.
Turning One-Off Hardware into Recurring Subscriptions
In 2023, General Motors announced a controversial strategy that defied consumer intuition: starting with model year 2025 vehicles, it would phase out third-party smartphone projection entirely across its new electric lineup. All new Chevrolet EVs would ship exclusively with a custom native infotainment system built on Google’s Android Automotive OS. Aside from basic Bluetooth phone calls, owners would no longer be able to mirror their smartphone apps onto the central touchscreen.
The core motivation behind stripping out a phone projection interface that hundreds of millions of drivers use by muscle memory is cockpit data control and recurring revenue. The global auto industry is racing to reinvent its business model, shifting from one-off vehicle manufacturing margins toward high-margin, recurring software services. Wall Street prizes average revenue per user (ARPU) and predictable subscription streams. To satisfy investor appetites for recurring revenue, automakers feel compelled to trap drivers inside their proprietary billing channels.
Photo: The Chevy Blazer EV. Source: a.wholelottanothing.org
As long as CarPlay remains the default interface, drivers stream music through Apple Music or Spotify and navigate using cellular data from their phones at no cost to the carmaker. By cutting that cord, GM saw an opportunity to erect a tollbooth—demanding monthly subscription fees for in-vehicle cellular connectivity and premium navigation. It was a high-stakes gamble: shut down an open interface to engineer a recurring monthly bill.
Deep Integration Fails to Mask Everyday Friction
From the perspective of GM’s leadership, developing an in-house operating system had defensible engineering logic. GM’s public defense centered on data privacy and holistic vehicle integration. The company argued that reclaiming OS control prevented driver telemetry from flowing to Apple, while deep hardware-level access enabled features that smartphone projection simply could not deliver.
Architecturally, native vehicle software can access hardware sensors that a projected smartphone cannot touch. A deeply integrated system can project turn-by-turn navigation directly onto an Augmented Reality Head-Up Display (AR-HUD) in the driver’s line of sight. When planning routes to fast-charging stations, the native system can automatically pre-condition the battery pack in advance to slash charging wait times. It can also unify cabin controls under global voice commands, letting drivers adjust climate settings with a single spoken phrase.
These native capabilities do provide marginal utility in specific scenarios. However, those niche advantages were quickly wiped out by severe friction in high-frequency everyday tasks. To use native navigation, drivers must log into separate in-car accounts and tolerate disconnected address books that don’t sync with their phones. Compared to the occasional need for battery pre-conditioning, having your podcast or playlist resume seamlessly the moment you step into the car is the real daily user experience that matters.
The Walled-Garden Impulse Transcends Borders
The urge to turn the vehicle cabin into a captive traffic pond is not unique to legacy Detroit automakers. In China’s fast-evolving new energy vehicle (NEV) market, leading EV brands have embraced remarkably similar software enclosure strategies. Most domestic EV startups and market leaders deliberately restrict or omit Apple CarPlay support.
Their underlying playbook is identical to GM’s: invest heavily in proprietary in-car operating systems, establish in-house application ecosystems, and turn the cockpit into an independent digital hub to capture the user interaction lifecycle.
The critical difference lies in software execution. Chinese EV makers developed exceptionally rich local app ecosystems, characterized by fluid interfaces and rapid over-the-air update cadences that largely offset the loss of phone projection. GM’s initial native system, by contrast, offered little compelling advantage over CarPlay while magnifying the pain of a severed habit. The contrast between both markets proves a fundamental rule: a closed ecosystem only works if your standalone product vastly outperforms the open alternative. Enclosing users by sheer mandate inevitably triggers a revolt.
Muscle Memory and Safety Beat Flashy Screens
Beyond business models and software ecosystems, smartphone projection addresses a fundamental driving safety issue. Tech writer Matthew Haughey pointed out that CarPlay’s greatest strength is blind-touch interaction born of deep familiarity. Drivers can use voice commands to send texts, check messages, or pull up regular routes without ever taking their eyes off the road.
At highway speeds, every extra second of cognitive friction is an acute safety hazard. Smartphone operating systems have been refined over more than fifteen years, and users have developed near-instinctive muscle memory for their layout and behavior. Forcing drivers to abandon that subconscious fluency in favor of an unfamiliar, idiosyncratic automaker touchscreen UI directly increases driver distraction.
On car forums and owner communities, the absence of familiar smartphone mirroring has become a dealbreaker. Countless prospective EV buyers openly state that they couldn’t care less about native infotainment specs—the moment they learn a vehicle lacks CarPlay, it is immediately crossed off their shopping list. Consumer demand for digital continuity has proven far more inelastic than automakers anticipated.
The Ledger Quantifies the Cost of a Walled Garden
The Honda Prologue provided the auto industry with a stark quantification of what happens when you fight user habits. Built with identical chassis architecture, battery capacity, and motor specs, Honda’s sole meaningful distinction was listening to consumers and retaining smartphone projection. Since both models debuted in 2024, their sales trajectories have diverged into a massive scissor gap:
| Period | Chevy Blazer EV | Honda Prologue | Honda Advantage |
|---|---|---|---|
| 2024 Full Year | 23,115 | 33,017 | +42.8% (+9,902 units) |
| 2025 Full Year | 22,637 | 39,194 | +73.1% (+16,557 units) |
| 2026 First Half | 3,166 | 8,407 | +165.5% (+5,241 units) |
In the first year, a 42.8% deficit might have been dismissed as brand equity differences. But by 2025, the gap widened to 73.1%—market discipline taking effect as real-world word-of-mouth spread. By the first half of 2026, the 165.5% gap delivered a definitive verdict. Two vehicles rolling off the same assembly line produced drastically opposite outcomes, proving that GM’s multibillion-dollar software ambitions were routed on the sales floor by an invisible software interface.
Chart: U.S. sales comparison (Honda in orange vs. Chevy in blue, 2024–2026H1). Source: a.wholelottanothing.org
Automakers vastly overestimated the indispensability of their proprietary software, while severely underestimating consumers’ unwillingness to compromise their digital routines. No matter how powerful in-car chips become, a vehicle remains a semi-stationary digital island that cannot replace the phone as the central thread of daily life. A car that deliberately fights its driver is perceived as fundamentally flawed. By attempting to build a subscription tollbooth behind a wall, GM ultimately locked out its own customers.
Reference Links:
- Original article on a.wholelottanothing.org
- Hacker News discussion (item?id=49590225)
