A 10-Year-Old Startup Just Raised $230M in a Single Round

01 /The Concept

This startup has been operating since 2013 and raised less than $20M over its entire history. Then, in a single round, it just raised $230M – from KKR, one of the world's most prominent private equity funds.

The distinction between private equity and venture capital matters here. Venture funds invest in startups to sell or list them later and pocket the gain. Private equity funds typically invest to collect dividends over time from a profitable business.

When a fund like KKR writes a $230M check, the implied expectation is a very substantial ongoing profit stream. Whatever XOi is doing, the economics are serious.

XOi built a platform that connects field technicians who repair equipment with the information and tools they need to do that job faster and better.

The entry point is a single photo. A technician takes a picture of the label on a piece of equipment they're working on – the app reads the text, identifies the exact model and variant, and pulls up everything relevant: user manuals, diagnostic guides, repair procedures, internal schematics, and related materials.

The core problem XOi solves: technicians currently spend an average of 2.5 hours per day hunting for information online before they can properly diagnose or fix an unfamiliar piece of equipment. XOi compresses that to minutes.

When the technician can describe the problem in words, the AI takes it further: it analyzes all available documentation for that device and assembles a step-by-step repair plan – listing likely failure causes, how to verify each one, and the recommended repair sequence.

The results are measurable. Companies using XOi see a 40% improvement in first-visit resolution rates – meaning 40% fewer callbacks from customers saying "the problem isn't fixed." The same 40% reduction applies to repeat dispatch trips caused by arriving without the right tools or parts. Time saved on each job becomes capacity for additional jobs and the revenue they generate.

Accurate diagnosis also benefits customers directly: companies using XOi extend 30% less payment credit because repairs come in cheaper, with less rework.

There's an additional revenue angle built into the product. When XOi identifies that equipment is near end-of-life – based on the model label and the nature of the fault – it flags the situation for the company's sales team with a notification labeled "Revenue Opportunity." The prompt: the technician is already on-site with a customer whose equipment is about to fail. Ideal time to propose a replacement.

XOi has built a substantial database covering equipment in HVAC, plumbing, electrical, construction tools, and commercial kitchens. About half the $230M round went toward acquiring Specifx, a startup that had been building a comparable database for a similar product. That database is now merged into XOi's.

Specifx was founded in 2022 and had raised just $3.5M. Its founders and investors sold for what appears to be well over $100M – confirming that this space is lucrative even for early players.

02 /Why It Matters

XOi is effectively "ChatGPT for field repair." If office workers – programmers, analysts, accountants – have been using AI assistants to do their jobs faster, there's no reason tradespeople and field technicians shouldn't have the same capability.

The addressable market is enormous. The global equipment maintenance and repair market was approximately $1.5 trillion in 2024 and is projected to reach nearly $2.5 trillion by 2029. That umbrella covers automotive, consumer electronics, medical devices, and more – each a potential home for its own XOi-style platform.

Commercial and industrial equipment maintenance alone was roughly $350B in 2024 and is expected to grow to $500B by 2029. Several startups are already active in adjacent segments:

Squint ([related review](/review/700-millionov-chelovek-kotorym-obychnye-platformy-ne-podojdut)) and DeepHow ([related review](/review/63-milliarda-dollarov-na-700-millionov-uchenikov)) both built platforms that let workers point their phone at industrial machinery to pull up task-specific instructional videos, raising $19M and $37.1M respectively.

HVAC – heating, ventilation, and air conditioning – is another large sub-market: approximately $78B in 2023, heading toward $110B by 2029. Multiple startups are building digitization tools here, with the most interesting business model belonging to Pipedreams ([related review](/review/makdonalds-dlja-uslug)).

Pipedreams builds the equipment database and digital tools – but doesn't sell them. Instead, it acquires small service companies in the space and connects them to its platform, capturing the efficiency gains as profit rather than licensing fees. The company buys at a price proportional to historical revenue and earns substantially more after digitization takes hold. Pipedreams raised $35.7M, and at its most recent round of $25.5M stated it no longer plans to raise equity capital – only debt. It's self-funding growth from the returns.

Another interesting business model comes from Scription ([related review](/review/jetu-biznes-model-mozhno-ispravit)), which has a contrarian take: the goal shouldn't be to fix equipment faster, it should be to prevent it from breaking in the first place. Its database of common failure modes powers a predictive maintenance service that helps companies schedule preventive work before things go wrong.

This shifts the entire business model: clients move to a maintenance subscription rather than paying per repair call. Clients benefit from less downtime; the service company benefits from predictable recurring revenue – getting paid regularly to keep things working rather than reactively fixing things that broke. Scription recently narrowed its focus to restaurant equipment and raised a new round in January, amount undisclosed.

03 /Opportunities

The broad direction: build a "ChatGPT for repair" in a specific vertical – narrow or broad.

Specialization is probably non-negotiable. The real asset in this space is a comprehensive database of equipment models and the specific failure patterns associated with each one. That database is the moat. The AI is only as good as the data it's trained on, which means generic models won't match a deep specialist.

What makes this space particularly attractive is that you can win in two different ways: build for long-term dividend-style returns, or build to be acquired.

Today's XOi story confirms both are live possibilities simultaneously – the private equity round reflects the long-term earnings potential, while the Specifx acquisition at a massive multiple proves that smaller players in the same space can command serious exit prices.