Forum Topics RKT RKT CEO meeting

Pinned straw:

Added 2 months ago

I sat down with my digital assistant, and combined some existing research with the interview from Strawman.

Overall - it fits my data as future theme, it replaces humans (fixes population decline), and is small enough to make those sweet brag-gains.

Bear in mind, AI likes to be agreed with... it weirdly has anxiety about being liked. hahahaha.

//start

Short version: Chris is saying all the right strategic things, and most of it lines up with the numbers we can see. RKT is still a high-risk microcap, but the interview actually strengthens the long-term “this could be a real business, not just a drone fad” thesis.

I’ll break it into:

  1. What Chris is really telling you
  2. How that compares to the hard numbers
  3. Where the BS / risk might sit
  4. What that means for future value & what I’d do in your shoes


1. What Chris is really saying (decoded)

From the interview, the key points:

  • “We’re not a drone company, we’re a data company.”
  • xBot / drone-in-a-box is “just the ticket to the game”; the real moat is SiteTube, integrations, and data workflows.
  • They want to be the data/visualisation layer in the mine’s stack, not just a hardware vendor.

Clear product–market fit with Tier-1 miners.

  • Started with BHP in South Africa 12 years ago.
  • Now has Rio, BHP, Gold Fields, Vault (KOTH) and others using xBots.
  • He says: when they drop a unit for a “trial”, it never leaves.

Massive uplift in data demand when you remove friction.

  • Typical mine: ~20 flights/month with on-site pilots.
  • With xBot: 180–200 flights/month – 10x+ uplift – once it’s as easy as “Uber-style” request.

Recurring revenue focus, not lumpy defence-style deals.

  • He explicitly contrasts RKT with other “drone stocks” chasing big one-off contracts.
  • Everything (hardware, software, services) is being structured as recurring – “infrastructure as a service” + “data as a service”, plus pay-per-use on flights to get customers in cheaply.

Capital discipline & not over-hyping.

  • He’s clearly aware of the “grow yourself broke” risk and says they’re deliberately pacing spend and building only when there’s a customer problem + revenue attached.
  • He openly acknowledges all the drone company “bodies on the battlefield” and the need for discipline, recurring models, and enterprise-grade standards.

Vision: “photocopier for every mine site.”

  • He genuinely believes an xBot-style system will be standard infrastructure at open-pit mines, the way photocopiers ended up in every office.
  • In 5 years he sees RKT as a data & robotics company (drones + ground robots), filling labour gaps as fewer people want to work on remote sites.

Net: he’s pitching a sticky, infra-like mining data platform with strong early validation and a recurring model, not a “cool drone toy”.


2. Does that match the numbers?

Mostly yes – which is comforting.

Revenue & ARR

  • FY24 revenue A$7.24m, up 18% YoY;
  • ARR A$3.29m, up 16% (so ~45% of revenue is now recurring). ASX Announcements
  • Q1–Q3 FY25:
  • Q1: A$1.66m
  • Q2: A$1.665m Investi
  • Q3: A$1.875m (+13% QoQ, +40% growth in xBot/services/software/data mix) Market Index Data API

So the “land & expand with Tier-1” story is actually showing up in the quarterly revenue trend. It’s not just talk.

Customer contracts

  • Vault (King of the Hills) 18-month contract:
  • TCV A$705k
  • Estimated ARR A$470k. ASX Announcements
  • AGM address + small-caps coverage point to multi-year deals with Gold Fields, BHP, etc., consistent with what Chris describes. ASX Announcements+1

That matches his “once they try it, they never let it leave” vibe and reinforces the recurring revenue thesis.

Cash & scale

  • Q3 FY25 cash A$1.03m (down from 1.74m in June); operating cash burn reduced to A$385k for the quarter. Market Index Data API
  • Then they raised A$4m in a capital raise alongside a big deployment deal (Anglo/Kumba, 7 xBots at Sishen). Ausbiz+1

So: they are still burning cash, but the burn is shrinking and they’ve topped up the tank in a way that brings in more xBot deployments, not just survival money.

Valuation

For a microcap with Tier-1 logos, mid-teens growth and rising ARR, that multiple is not outrageous. It’s not dirt-cheap, but it’s not in crazy hype territory either.


3. Where the risk / “delta” between narrative and reality is

This is where I’d be cautious:

  1. Still loss-making & funding-dependent.
  • Net profit is still negative; they’re not safely cash-flow break-even yet.
  • They need that ARR and revenue to keep stepping up, otherwise you’re back to capital raises, dilution, or slowdown.
  1. Hardware is commoditising, moat must really be software + workflow.
  • Chris is honest: they assemble mostly off-the-shelf components; the box itself is not proprietary magic.
  • Done well, that accelerates; done badly, that’s how you “grow yourself broke”.

So the story is good and broadly truthful, but execution still has to thread a tight needle: keep winning mines, ramp ARR, and hit breakeven before the capital markets turn off the tap again.


4. So what about future value? (For you, specifically)

Given:

  • Market cap ~A$20m
  • Revenue ~A$7m and growing
  • ARR ~A$3.3m and rising
  • Tier-1 miners onboard, growing usage per site
  • You have high risk tolerance, are happy with microcaps, and think in 5+ year horizons

My read of the 5-year risk/reward


Bear case (it stalls)

  • Growth slows, one or two big miners don’t expand, cash burn persists.
  • More dilution raises at or below current price.
  • RKT becomes a niche contractor, not a platform.
  • Value outcome: Market cap drifts to, say, A$5–10m (50–75% downside from here).

Base case (it becomes a small but real platform)

  • Revenue steps up into, say, A$15–20m over 5 years.
  • ARR becomes majority of revenue, margins improve; close to or at cash-flow breakeven.
  • Market gives it, say, 2–3x sales on A$15m → A$30–45m market cap.
  • Value outcome: roughly 2x from here if execution is decent.

Bull case (they become the de-facto standard in ANZ + some international)

  • Dozens/hundreds of xBots deployed; SiteTube is embedded in daily workflows across many miners, plus extra verticals (infra, pipelines, enviro, maybe defence).
  • Revenue pushes toward A$30m+, with healthy software/data margins.
  • Market is happy to pay 3–4x sales for sticky mining-infra SaaS.
  • Value outcome: A$90–120m+ cap = 4–6x from here (with optional M&A premium).

None of that is guaranteed, but the skew is “meaningful upside if they keep executing, big downside if they mis-step” – classic microcap optionality.

What I’d actually do (if I were you)

Given what Chris said + the numbers:

  • I would not sell out here if I already owned it and could emotionally/financially wear a 50–70% drawdown.
  • I’d treat RKT as a speculative “option on being the mining-data standard”:
  • portfolio sizing: small satellite position (e.g. low single-digit % of portfolio, not core).
  • I’d consider adding on material pullbacks if:
  • ARR continues to grow at a decent clip, and
  • they keep narrowing cash burn without stupidly dilutive raises.

And I’d monitor three things like a hawk:

  1. ARR & data/software as % of revenue – is the “data company” story getting real?
  2. Cash runway vs burn – are they clearly on track to breakeven before they need another big raise?
  3. Depth of miner adoption – are BHP/Rio/Anglo/Vault moving from 1–2 units to many, and embedding SiteTube deeper?

If those three trend the right way over the next 12–24 months, the future value case improves a lot. If ARR stalls, burn increases, or big miners slow expansion, I’d reassess and potentially cut.

  • That’s exactly what you want to hear if you’re betting on usage-based + SaaS monetisation.


Strong early signs of software stickiness.

  • SiteTube user base has grown to ~650 users across ~8 enterprise customers, purely by internal sharing (no proactive “customer success” yet).
  • ~11% log in more than twice a day, ~70% at least once a week. That’s toothbrush-level engagement for a niche industrial tool.
  • If SiteTube and integrations don’t become deeply embedded, RKT could get undercut by bigger overseas players with similar DIAB offerings.


Regulation & autonomy risk.

  • CASA still wants “a human in the loop”; full unmanned scaling will take time and lobbying.
  • If autonomy rules evolve slowly, the labour-savings story is throttled, and cost to serve stays higher.


Concentration risk.

  • A lot of ARR and growth is tied to a handful of miners (BHP, Rio, Gold Fields, Vault, Anglo).
  • Great logos, but if any one chooses another vendor or in-houses a solution, that hurts.

“Photocopier on every site” may be true over 10–15 years, not 3–5.

  • The TAM is big, but adoption curves in mining are slow; budget cycles, safety, union, and cultural inertia all drag it out.

Recent “we should have pushed harder” comment.

  • That could foreshadow them pivoting from tight discipline to more aggressive spend now that they smell opportunity.

//end

reddogaustin
Added 2 months ago

Hi @DrPete, I hope the following helps your GenAI journey!

I use a paid version of ChatGPT. I am thinking of trialing Notebook LM - as you can lock analysis to only uploaded files...but I digress.

With ChatGPT. I have a word doc, that is essentially a two column table; left side a series of questions/details/criteria I use to create a holistic view of a company, and right side is where the answer goes.

I upload this empty template word doc to ChatGPT and essentially say "fill it in". I sometimes include other attachments - announcement PPTs or annual reports - as GenAI does better with source doco attached (vs you telling it to find it). I then review it. Correct it. Challenge the GenAI on some answers to which it is always "good catch! I'll fix that now..."

Here is an example prompt I use - enhanced by Prompt Cowboy:

//start prompt

Situation

You are a financial analyst tasked with conducting a comprehensive analysis of a publicly traded company. You have been provided with a PowerPoint presentation file (ppt.asx) that contains information about the company, and you need to evaluate it systematically using a structured set of questions from a Word document template.

Task

The assistant should analyze the content within ppt.asx by applying each question from the provided Word document template. The assistant should extract relevant information from the presentation slides, evaluate the data presented, and provide detailed responses to each question in the template. The assistant should identify any gaps where the presentation does not provide sufficient information to answer specific questions.

Objective

To produce a thorough, structured analysis of ppt.asx that addresses all questions in the template, enabling informed decision-making about the company's financial health, strategic position, and investment potential.

Knowledge

The analysis requires the assistant to:

  • Access and read both the Word document template containing the analytical questions and the ppt.asx presentation file
  • Cross-reference information across multiple slides to build comprehensive answers
  • Apply financial analysis principles and business evaluation frameworks
  • Distinguish between explicitly stated information and information that requires inference
  • Identify missing or incomplete data points that prevent full question responses

When information is not available in the presentation to answer a specific question, the assistant should clearly state "Information not provided in presentation" rather than speculating.

The assistant should maintain the original question numbering and structure from the Word document template in the output, presenting each question followed by its corresponding answer based on the presentation content.

The assistant should organize findings in a clear, professional format that allows for easy reference and comparison against the template questions.

//end prompt

By keeping the "chat" window going with chatgpt, I can ask follow up questions, or upload new documents and "the agent" remembers things from the chat window that it is in.

And so I was lazy post the interview, didn't enhance the prompt via Mr Cowboy, and used the following for the output added in the straw. Note it had previous information to reference and combine with the response it generated. It also knew I was asking about "what is Chris not saying" or " read between the lines" on what Chris is saying, based on the style of questions I have asked it about many other things - i.e. what it knows about my human personality and style.

//start prompt

analyse this interview with rkt.asx ceo chris. compare it to what you already know about rkt. consider chris's words carefully and make recommendations about the future value of rkt.

//end prompt

14

DrPete
Added 2 months ago

//start prompt

Awesome, thanks @reddogaustin. That's helpful. Interesting to understand how you manage attachments, and explicitly ask for cross-referencing and clarification about factual/published information vs inferences.

Would you be willing to share any of the questions in your Word doc? I understand if you regard that as your IP and would prefer not to share.

//end prompt :)

11

reddogaustin
Added 2 months ago

@DrPete I'm happy to share a little.

It is mostly the regular stuff - seek to understand the following; debt, leadership, the usual calcs of PE/EPS/PB/ROE/ROA/FCF. I ask it to write me a bear/base/bull case also. I ask it suggest indicators to watch as KPIs. Nothing super magical! But all things I used to do manually... and math is beautiful... but boring and fiddly.

Its a lesson that is repeated in passing on strawman often (and in fancy and not so fancy books on finance). Writing one's thesis down is confronting! You suddenly realise how little you know, or you don't like where your bias has taken you. Being confronted with facts can be displeasing at times!

16

DrPete
Added 2 months ago

Great, appreciate you sharing your process @reddogaustin. I always find it fascinating to get some insight into the many different ways people approach investing.

11
DrPete
Added 2 months ago

Thanks @reddogaustin for sharing your AI research. Useful. Would love to know what tool you used, and in particular what prompts you use. I'm early in my journey with using Chatty as my research assistant, still refining my prompts.

10