The 12 roasts, with receipts.

So what is this technological moat? According to their own product page: sensors monitor TDS, flow, pressure and pump current. Data travels over TCP/IP or GSM. Alerts go through SMS, app or dashboard.

Guys. That is useful. That is industrial telemetry. That is remote monitoring. But unless there is significantly more technology behind the curtain, that is not exactly CERN.

The site even calls these “Predictive Alerts.” Yet its own explanation says the sensors flag things like membrane fouling or pressure drops before they interrupt operation. There is no published explanation of predictive models, historical training, remaining-useful-life estimation, anomaly models, failure probabilities, machine learning, digital twins — or even the name of an algorithm.

This may be the best line on the entire website. Twenty percent more what? Water recovery? Electrical efficiency? Membrane efficiency? Throughput? Flux? Hours between someone hitting the machine with a spanner?

And which incumbents? There is no benchmark, no baseline, no reference design, no test methodology, no comparison table. Nothing.

That is an enormous commercial claim. Yet the page provides no competitor comparison, no BOM economics, no consumables comparison, no maintenance schedule, no membrane replacement assumptions, no operating-hours assumption, no electricity comparison, no water recovery calculation, no service cost model, no TCO period, and no case study.

Maybe the claim is completely defensible. But the website gives a procurement manager virtually nothing with which to defend it internally.

It genuinely looks as though the web designer created twenty logo placeholders and everybody collectively decided: “We’ll fill those after Series A.”

What’s particularly strange: in a 2026 Delhi High Court proceeding, Althion’s counsel stated the company had more than 40 installations — naming AIIMS campuses, NIMS Hyderabad, CMC Vellore, Tata Memorial Hospital and Sir Ganga Ram Hospital. That’s Althion’s assertion recorded in a judgment rather than independent verification — but if those references are publishable and current, they’re vastly more compelling than an almost-empty customer section.

Then: ultrapure water. Then: kidney dialysis. Then: laboratory. Then: a “portfolio” of hemodialysis monitoring and heart health monitoring. But the actual public product catalogue today is essentially: Big RO. Small RO. Lab RO.

There is nothing wrong with owning a narrow vertical — medical water purification is a technically demanding and commercially meaningful niche. The weird part is trying to sound broader than the evidence presented.

The Government of India’s DST says the products were developed based on technology licensed from CSIR-IICT. The product page says “Patented Technology: Jointly with CSIR-IICT.” Separately, in a 2026 Delhi High Court matter, Althion’s counsel stated Althion holds a 2021 patent for a compact softener-free hemodialysis RO system.

Those statements may all be reconcilable — licensed IICT membrane IP plus Althion’s own subsequent system patent. But the website never explains that.

Meanwhile they prominently display the 2020 National Award. So visually the message becomes: Award achieved six years ago ✅. European/Indian product markings: working on it ⏳.

Medical-device certification pathways take time and vary by product — that doesn’t necessarily mean anything is wrong. But prominently advertising pending certifications is unusual unless accompanied by a roadmap or current regulatory status.

Ironically, this is where the roast has to give them credit. A Delhi High Court judgment records Althion’s representation that it holds ISO 13485 certification, CDSCO Class C licensing for the dialysis-grade RO device, DSIR recognition of its R&D centre and a 2021 patent. India’s Department of Pharmaceuticals has listed Althion as a manufacturer of portable RO plants. That is actual substance.

Which makes the website even more frustrating. Where are the certificate numbers? Downloadable certificates? Test reports? Validation protocols? Installation references? Water-recovery curves? Service SLAs? Real dashboards? Uptime data? Documented before/after numbers?

To Althion’s credit, the AT250 page actually provides useful parameters: 100/250 LPH, feed requirements, power, rejection rate, filtering stages and water-quality specifications. Then it suddenly switches back into marketing mode.

99.6% rejection — okay, under what feed chemistry? Temperature? Pressure? Recovery? Membrane age? AAMI compliant — great, show representative validation reports. Remote monitor — good, show the dashboard. Predictive alerts — show the prediction.

The genuinely interesting part of Althion does not appear to be IoT. It is the water-treatment architecture: licensed CSIR-IICT membrane technology, elimination of resin softening, reduced salt requirement, and potential water/electricity savings. That is interesting. That’s physical-world IP. That’s measurable. That’s defensible. That’s something a hospital can economically understand.

Instead of hammering that engineering differentiation, Althion repeatedly reaches for: I O T   E N A B L E D.

Aster Technologies and Controls LLP is a reputed Pune manufacturer that has sold industrial RO controllers for years: Astero NXT, RO Matic NG, Astero-Mini. Their spec sheets read like Althion’s IoT page — conductivity/TDS, flow, pressure and level inputs; pump and flush-valve control; RS-485 for SCADA/BMS; Bluetooth pairing; ‘IoT and communication ready’. The difference is that Aster publishes the model number, the I/O list and a price on request. Althion publishes the word ‘proprietary’.

Don’t take our word for it — drag the slider below. Left is the front panel from Althion’s own AT250 product photo; right is Aster’s Astero nXT catalogue shot. Same Menu/Select/arrow cluster, same round Start/Stop button, same System Auto/Manual and Floaty Auto/Bypass indicators, same RWT–RWP–FILTER–MF–LPS–HPP–HPS–RSV–TWT schematic in the same order. And in the top-left corner of Althion’s own marketing photo, in Aster’s own typeface: ‘ASTERO nXT 11’. Nobody even peeled the label. We are not saying Althion did anything wrong by buying a good controller — we are saying the word ‘proprietary’ is doing a lot of unpaid work, and Althion has never published a spec sheet, a firmware note or a teardown that would settle it.

Two possibilities, both bad for the pitch deck. Either the ‘IoT platform’ is a stock industrial RO controller with a sticker and a GSM modem — in which case it is procurement, not R&D. Or Althion independently re-engineered a commodity controller that a Pune LLP already sells off the shelf — in which case it is R&D spent rebuilding something with zero moat. Either way, ‘proprietary’ is doing a lot of unpaid work in that sentence.

A patent is a legal document. A moat is a business fact. Althion has spent five years treating the first as proof of the second. The filing is real — Patent No. 379691, a compact double-pass RO arrangement that swaps in hydrophilized polyamide membranes so the softener can go. Read the claims yourself (links above): it is an arrangement of commodity RO components and a membrane choice. That is the kind of IP every mid-size Indian water company could design around in a quarter, and most never bother to file because customers don’t buy patents, they buy water.

Follow where the word ‘patented’ actually appears: the funding press, the award citations, the LinkedIn funding alert, the pitch to BioAngels. Follow where it does not appear: a licensing deal, a royalty line in the financials, a competitor forced to change a product, a tender won on the strength of exclusive technology. Nine years of ‘IP-led’ and the IP has never once shown up as revenue. It shows up as a slide.

Meanwhile the technology the company is most proud of — the membrane chemistry — is licensed in from CSIR-IICT, which means the genuinely defensible science belongs to a government lab, not to Althion. What Althion owns is the packaging. Filing a patent on the packaging and calling it a moat is not fraud; it is just the oldest trick in Indian deep-tech fundraising: the namesake patent, filed to be mentioned, not to be enforced.