# Komatsu Doubles Down on Autonomy — This Time With Tier IV
Komatsu is now running two autonomy partnerships at once. Tier IV and its subsidiary EARTHBRAIN have begun joint development of autonomous driving systems for three Komatsu dump truck models, targeting quarrying and civil engineering sites in Japan and North America, with commercial introduction targeted for 2027. The company is also working with Silicon Valley’s Pronto on Smart Quarry Autonomous for the North American market. Here’s what this really means: Komatsu isn’t betting on one horse — it’s buying options. That’s either prudent hedging or an admission that nobody has fully cracked off-road autonomy yet. Given my line of work, I’d say both.
The hardware lineup tells you where the money is. The partnership covers the HD785 rigid dump truck (93.9 t capacity) and HD605 (64.1 t) — mining workhorses — plus the smaller HM400/460 articulated dump trucks at 42 t, which serve quarries, civil engineering sites, and smaller mining operations. Tier IV brings the autonomy stack; Komatsu brings vehicle design and control technology. Together they plan to validate systems for the full cycle: loading, hauling, dumping. That last part matters more than the press releases admit. Getting a truck to drive itself on a haul road is the easy demo. Handling the messy, unscripted chaos of a live quarry — load sequences, spot-on dump timing, weather, human drivers wandering around — is where autonomy programs have historically bled money and timelines.
The EARTHBRAIN angle deserves attention. By tying the autonomous system into a site-wide control platform, the pitch moves beyond individual vehicles to full operation management. This is the correct framing, frankly. Autonomy as a standalone feature is a gadget. Autonomy stitched into fleet orchestration is a productivity play — and productivity plays are the ones that eventually show up in earnings. The stated goal of accumulating operational knowledge through field demonstrations before scaling is the sober path. It’s also the slow one, which is exactly why you should discount any 2027 revenue projections heavily. This isn’t the first time we’ve seen this circus; autonomy timelines have a habit of slipping two to three years past the announcement deck.
The labor math is the real driver here, and it’s not debatable. Experienced haul truck operators are aging out and hard to replace in both Japan and North America. Mines and quarries don’t have the luxury of waiting for a perfect labor market. Automation that reduces labor dependency and improves safety is a structural demand story, not a cyclical one — assuming the technology actually works at commercial reliability. That assumption is where I keep my skepticism loaded. Tier IV made its name in open-source autonomous driving research, largely on public roads. Off-road hauling is a different animal: no lane markings, brutal terrain, enormous payloads where a software error costs serious money. It’s a harder problem in some ways and easier in others — controlled sites, predictable routes, no pedestrians (in theory). The gurus will tell you it’s a slam dunk. It isn’t. It’s a multi-year grind of demonstration tests, edge cases, and incremental validation, which is precisely what these two companies say they plan to do. Fine. Good. That’s the honest approach.
The dual-track strategy is worth watching. Komatsu pairing Pronto’s proven North American quarry autonomy with Tier IV’s stack for Japan and broader applications suggests they’ll let field results decide which platform wins where. For customers, that’s potentially confusing — two systems, two integration paths, one OEM. For Komatsu shareholders, it’s sensible risk management in a sector where being late to autonomy means losing fleet refresh cycles to Caterpillar, which has been running autonomous haulage at scale for over a decade. Komatsu isn’t pioneering here; it’s catching up. Nothing wrong with that, but don’t let anyone frame this as revolutionary. It’s competitive necessity dressed up in a press release.
My take for those watching this space: the investable theme isn’t the autonomy software — most of it isn’t publicly traded anyway. The theme is capital intensity. Mines and large quarries facing labor scarcity will spend real money retrofitting and replacing fleets over the next decade, and OEMs with credible automation offerings will capture margin on hardware, software subscriptions, and site-management platforms like EARTHBRAIN. The 2027 target should be treated as a milestone marker, not a revenue event. Watch the demonstration test results, watch customer adoption rates at pilot sites, and ignore the headline word count of partnerships announced. In this sector, traction is measured in trucks running autonomous hours, not in press releases. Until then, this is a story about positioning, not proof.
