Published on 15/05/2026 — Last updated on 01/09/2026. We have since obtained the full test reports from Nano Filter and have analysed them in detail: the update of 1 September, at the bottom of the article, explains what they contain and on which filtered volume the measurements were made. Two previous updates, from 14/07/2026 and 23/07/2026, remain available in the same place. The original text below has not been modified.

 

Nano Filter highlights very fast filtration, an accuracy announced at 0.002 μm, a patented Swiss technology and a single cartridge supposed to filter 4,000 liters. In the APS promotional email dated Friday, May 15, 2026, Nano is presented as a "HUGE revolution", "50 times more accurate" than Coldstream, with a flow rate of 18 liters per hour, or "the equivalent of a large glass of water per minute".

 

It's a very strong commercial argument. It's visual, simple, easy to understand.

 

But in water purification, speed is not enough. The real issue is not which filter fills a glass the fastest. The real issue is what remains in the water at the end.

There's no point in running, but you have to arrive with really controlled water.

When marketing runs faster than evidence

In the market for gravity water filters, there are two opposing approaches.

• The first is to publish complete tests, name the laboratories, indicate the standards used, specify the contaminants tested, the volumes filtered, the detection limits and the exact conditions of the tests.

• The second is to talk about "revolution", "Swiss nanofiltration", "0.002 μm", "50 times more accurate", "bacteria and viruses destroyed", "TFA eliminated", then to show water flowing very quickly.

Nano Filter clearly plays this second card.

The brand claims two patented technologies, PlasmaShield-NanoMatrix and Nanoporous Adsorber Fleece, which are supposed to allow a level of filtration inaccessible to conventional gravity filters. On its technology page, Nano claims that its membrane destroys pathogens and that its system reaches 0.002 μm, "50 times more accurate" than conventional gravity filters.

Note: these names do not appear to be unique to Nano Filter. Swissreg lists PLASMASHIELD-NANOMATRIX and NANOPOROUS ADSORBER as trademarks associated with Evodrop AG, a Swiss company active in water treatment. This suggests that Nano is likely relying on a technology, license, or vocabulary from this Swiss ecosystem. This is not, in itself, evidence of ineffectiveness. But it is also not proof of performance under real-world gravity conditions: a trademark, proprietary technology, or data sheet is not a substitute for a full lab report on the final product.

The same page goes so far as to write: "PFAS, pesticides, drug residues, heavy metals: nothing gets through."

This is a massive promise.

But a massive promise requires massive evidence.

➜ And this is precisely where the Nano file becomes fragile.

To remember

Flow rate is a criterion of comfort. Not proof of purification.

A filter can fill a glass very quickly while leaving the essential questions open:

• Which contaminants were actually tested?
• By what method?
• On how many liters?
• With what detection limit?
• And with which comprehensive reports are available ?

Nano highlights spectacular speed. But in the end, when you compare the available public evidence, Coldstream FTO+ remains much more solidly documented: tests according to recognized standards, identified laboratory, wide range of contaminants, results at 3,000 L, and even a clearly named virus test.

Speed does not purify water

Nano's central argument is simple: it goes fast.

In the APS email dated 15/05/26, The flow rate is presented as "unparalleled", at 18 liters per hour. The message insists on the visual comparison with a transparent tank, as if the speed was enough to demonstrate the superiority of the filter.

On its product page, Nano advertises a cartridge at 0.3 L/min, a lifespan of 4,000 L or 12 months, and "15x faster" filtration. The same page also highlights "1 glass of filtered water per minute".

All right. That's fast.

But a fast flow does not prove better purification.

The flow rate answers a practical question: how long should you wait?

The tests answer a health question: what am I really drinking ?

To confuse the two is to take a stopwatch for a laboratory report. Water can pass through a cartridge quickly without it showing anything about TFA, viruses, pesticides, nanoplastics, heavy metals or drug residues.

In filtration, speed is an advantage of use. But safety is proven by tests, not by a speed video.

  

The 0.002μM: A spectacular figure but not universal proof

Nano claims to achieve a filtration of 2 nanometers, or 0.002 μm, and presents this as "50 times more accurate" than a conventional gravity filter. The brand even writes that "PFAS, pesticides, drug residues, heavy metals: nothing goes through".

It's a very strong sentence. Too strong.

Dissolved chemical contaminants do not filter like sand through a sieve. PFAS, drugs, pesticides or heavy metals are not simply "particles" that are blocked because they are larger than a pore. Nano indirectly acknowledges this by explaining that PFAS are captured by hydrophobic interaction, pesticides by molecular affinity and heavy metals by electrostatic attraction or complexation.

So if Nano wants to prove effectiveness on these contaminants, the "0.002 μm" is not enough.

Contaminant-by-contaminant testing is required, with specific conditions, inlet concentrations, detection limits, flow rate, filtered volume, replicates, and complete reporting.

Otherwise, "0.002 μm" becomes above all a slogan. A nice slogan, certainly. But a slogan nonetheless.

TFA: a claimed figure, not verifiable proof over time

Nano claims a result on TFA, trifluoroacetic acid: 0.738 μg/L before filtration, "n.d." after filtration, reduction displayed at 100%. This result is associated with Bureau Veritas Report 424-1242/A.

On paper, it's impressive.

But we have to look at what it really proves.

• First, "n.d." does not mean "zero". It means "not detected below the instrument's detection limit". However, without the full Bureau Veritas report, the consumer cannot verify the exact detection limit, the method used, the water matrix, the pH, the TDS, the flow rate, the repeats, the number of cartridges tested or the volume actually filtered before the test.

• Secondly, Nano itself specifies that an "additional test" on performance at 5,000 liters is still underway and will be published after receiving the results.

So even if we accept the Nano synthesis as is, the TFA is not publicly proven at 4,000 or 5,000 liters.

It is claimed in a painting. It is not the same thing.

And this is a major point, because TFA is precisely one of the most difficult PFAS to remove. The TZW indicates that TFA cannot be eliminated by natural or conventional technical treatments of drinking water, and that ozonation, activated carbon or chlorination are practically ineffective; the TZW also indicates that the reduction of TFA in drinking water treatment is only really possible by reverse osmosis, with strong reservations about its generalization. The German agency BfR/UBA also describes TFA as highly persistent and highly mobile, with low affinity for sediment and activated carbon filters.

Conclusion

Nano claims a result on the TFA. But without a full Bureau Veritas report, a clearly published detection limit and an end-of-life TFA test, this claim is not enough to prove sustainable performance over 4,000 liters.

A TFA promise without a full report is like a racing car without a technical inspection: it can shine, but you don't know if it holds up.

Virus: one claim but no visible virus test

Nano explicitly talks about viruses.

On its technology page, the brand claims that upon contact with the membrane, "a bacteria, virus or any other pathogen" would have its envelope pulverized and its DNA broken. On the product page, Nano also claims that the cartridge "destroys bacteria and viruses on contact with it".

The APS email, dated 15/05/26, repeats this promise by stating that Nano filters "bacteria, viruses and pathogens", and that "all filtration performance has been validated by independent laboratories, accredited ISO 17025".

The problem? In the Nano public register, we don't see any specific virus test

The organisms listed are: Total coliforms, E. coli, P. aeruginosa and S. aureus. They are bacteria. The results are expressed in bacterial units, not in viral units. We do not see coliphagous MS2, no rotavirus, no surrogate norovirus, no PFU unit, no identifiable viral protocol.

It's a very serious shift in vocabulary.

Testing bacteria does not prove effectiveness on viruses. They are not the same organisms, not the same sizes, not the same structures, not the same validation protocols.

So yes, it's suspicious. And technically, it's not serious.

When a brand talks about viruses in its commercial promise, it must display a virus test. Not a bacteria table with the word "virus" in the title.

 

By way of comparison, Coldstream FTO+ explicitly displays a "Virus" line with Rotavirus spp., an input concentration of 2.4642 × 10⁷ PFU/L and a reduction >99.99% at 3,000 litres.

There, at least, the viral contaminant is named. The protocol is identifiable. The unit is coherent. The customer is not asked to guess.

Nano talks about viruses. Coldstream shows a virus test.

This is not a detail. It is a major difference.

 

Microplastics and nanoparticles: another slip

Nano claims to be effective on microplastics and also mentions nanoplastics. The APS email even goes so far as to speak of "microplastics, including nanoparticles".

But in the Nano public tests, the published results are on particles >5 μm and >10 μm. And Nano itself specifies that an additional test on nanoplastics smaller than 1 μm is underway.

So here again, we must distinguish between promise and proof.

Saying "we tested for particles >5 μm" is one thing.

To suggest that nanoparticles or nanoplastics are demonstrated in the laboratory is another.

As long as the nanoplastics test <1 µm n’est pas publié, cette promesse reste non démontrée publiquement.

 

Full reports: a summary sheet is not a certification

Nano displays results and cites references: SGS Report NBF24-0013449-01, SGS Report NBF25-0023352-02 and Bureau Veritas Report 424-1242/A. The brand also claims that the tests are carried out by SGS and Bureau Veritas, two independent laboratories accredited to ISO/IEC 17025.
As of the date of this analysis (15/05/26), we have not found a link on the public pages of Nano to download the full SGS and Bureau Veritas reports. Only report references and summary tables are visible. If these full reports are published at a later date, the analysis will naturally need to be updated.

But on the public pages viewed, the consumer does not see the complete original reports with all the pages, the laboratory header, the exact address of the test site, the scope of accreditation, the detailed methods, the contaminant-by-contaminant detection limits, the water conditions, the repetitions and any disclaimers.

And this is precisely what you should look at before buying a filter. Our guide recalls that a reliable filter must indicate:

• Who makes the product
• Which lab tested
• If the laboratory is ISO/IEC 17025 accredited
• What standards were followed
• If the product has an official certification

He also points out that a simple "SGS test" without details is not a sufficient guarantee.

Another point to be clarified : Nano cites SGS references of type NBF. However, SGS public reports with the same type of NBF prefix are issued by SGS-CSTC Standards Technical Services Co., Ltd. Ningbo Branch, China. This does not prove that the Nano report NBF24-0013449-01 is identical or issued under the same conditions, but it does make it essential to publish the full report to know the exact SGS site, the scope of accreditation and any limitations of use.

Coldstream, on the other hand, has a more classic, less sexy, but much more readable file: tests according to NSF/ANSI 42, 53 and P231, by IAPMO R&T Laboratory in New Jersey, with ISO 17025/NELAC compliance. The Coldstream sheet also indicates a nominal capacity of 1,500 L, a recommendation for replacement every six months, and results of up to 3,000 L on several families of contaminants.

Nano has a summary sheet.

Coldstream has a much stronger normative record.

 

The test panel: Nano tests little, Coldstream documents extensively

Nano displays results on several families of contaminants. But when you look at the details, the panel is very targeted.

• For pesticides, Nano publishes two main molecules: dimethoate and carbofuran.

• For drugs, Nano publishes two antibiotics: roxithromycin and oxytetracycline.

• For heavy metals, Nano highlights three metals: lead, mercury and cadmium.

That's not nothing. But it's not a wide range.

Coldstream FTO+, documents many more families and many more molecules: heavy metals, volatile organic compounds, semi-volatile organic compounds, pesticides, herbicides, pharmaceuticals, particulate matter and HAA5. In particular, the fact sheet displays many pesticides and herbicides at 3,000 L, with many reductions of more than 99%. It also displays a wide range of pharmaceuticals and endocrine disruptors at 3,000 L, as well as HAA5 results.

The difference is simple:

Nano selects a few contaminants that sell very well.

Coldstream documents a much broader behavior. Nano speaks broadly.

Coldstream documents widely. It's not the same.

The lifespan: 4000 liters announced, but not all the promises are proven at 4000 liters

Nano announces a cartridge for 4,000 liters or 12 months. The product page even claims constant effectiveness for 4,000 liters, "without progressive degradation".

But we have to look at it family by family.

On certain contaminants, Nano shows results after 5,000 liters: heavy metals, pesticides, drugs, chlorine, bacteria, particles >5 μm and >10 μm.

But on short PFAS and TFA, Nano says that the additional test at 5,000 liters is still ongoing.

On nanoplastics <1 µm, Nano indique aussi que le test complémentaire est en cours.

On viruses, no specific virus test is visible in the public register.

So we can't seriously write that all the great Nano promises are proven on 4,000 liters.

We can write that some families are tested at 5,000 liters.

But TFA, nanoplastics and viruses remain major weak points.

Coldstream takes a different approach: the sheet indicates a nominal capacity of 1,500 L and recommends a replacement at least every six months, while publishing results up to 3,000 L. On certain parameters such as fluoride, nitrate or chlorine, the sheet clearly shows that performance decreases with volume: this is not always flattering, but it is precisely what makes the case more credible.

Coldstream does not claim to do everything for 4,000 liters.

He shows what he's doing, and also where it's falling.

It's less salesy. It's more serious.

 

"Swiss technology": okay, but where is the cartridge made?

Nano strongly promotes a "patented Swiss technology" and a Swiss development. At this stage, we have not identified any public patent clearly attached to the Nano Filter product.

Based on the original product elements we have, the Nano cartridges are most likely made in China.

The problem is not that a product is made in China. Very good products can be made in China.

The problem is the discrepancy between a very "Swiss" oriented communication and the lack of clear public transparency about the actual manufacturer, the country of manufacture of the cartridge, the factory, the OEM and the complete reports.

When a brand puts so much emphasis on the Swiss flag, the consumer must be able to know what exactly is Swiss: the company? the patent? the design? the cartridge? the membrane? the factory? the assembly ?

Without this precision, "Swiss technology" becomes a marketing cover rather than traceability information.

Conversely, the Coldstream plug clearly states that the CF163W filter is manufactured by KLT Filtration Ltd in King's Lynn, UK, and mentions "Made in the United Kingdom". Again, it's not about the flag. It's about transparency.

Targeting the installed base: Nano doesn't just sell a filter, it wants to replace existing cartridges

The APS email, dated May 15, 2026, does not just recommend a complete new system.

It is also aimed at people who already have a gravity tank. The message tells them to keep their tank, then to "upgrade to the next technology" by replacing their existing cartridges with Nano's. It also claims that the cartridge is compatible with tanks from all brands on the market.

The Nano product page explicitly lists several compatible systems, including Pure Filters, Berkey, British Berkefeld, ProOne, Purewell, Phoenix and Weeplow.

So it's not just a new competitor.

It is a strategy for capturing the installed base.

And when a brand says that all you have to do is change the cartridge to move to a "superior technology", it has to provide superior proof.

Not just better throughput.

 

Summary Comparison Table

Criterion	Nano Filter	Coldstream FTO+
Main argument	Velocity, 0.002 μm, Swiss technology, 4,000 L	Documented tests, standards, identified laboratory
Flow rate	Very fast: 0.3 L/min	Slower
Flow rate = health proof ?	No	No, but Coldstream does not base its proof on speed
Reports	SGS/Bureau Veritas references displayed, but full original reports not publicly visible on the pages viewed	Detailed sheet with standards, laboratory, contaminants, volumes
Standards	Announced ISO/IEC 17025 Tests	NSF/ANSI 42, 53, P231, IAPMO, ISO 17025/NELAC
TFA	Claimed result, but 5,000 L test still in progress	Doesn't claim what isn't documented
Viruses	Commercial claim, but no specific virus test visible on the public register	Rotavirus spp. tested at 3,000 L, reduction >99.99%
Nanoplastics	Promise mentioned, but test <1 µm en cours	More framed claim on particles/microplastics
Pesticides	2 main molecules displayed	Wide range of pesticides/herbicides at 3,000 L
Medications	2 antibiotics displayed	Wide pharmaceutical panel at 3,000 L
Origin	Swiss communication, cartridge made in China according to our product elements	Clearly identified UK manufacturer
Transparency	Strong promise, partial evidence	More complete and verifiable record

 

Simple questions to ask before buying

Before purchasing Nano Filter, one should ask:

• Where exactly is the cartridge manufactured ?

• What is the name of the manufacturer or factory ?

• Are the full SMS reports NBF24-0013449-01 and NBF25-0023352-02 available ?

• Is the full Bureau Veritas 424-1242/A report available ?

• What is the exact limit of detection for TFA ?

• Is the TFA/PFAS test after 4,000 or 5,000 liters published ?

• Is there a complete virus test with MS2, rotavirus or other viral indicator ?

• Is there a nanoplastics test <1 µm publié ?

• Does the product have NSF, IAPMO, or WQA certification, or just spot testing ?

A transparent brand must be able to answer these questions simply

Conclusion: Nano is fast. Coldstream arrives with its proofs.

Nano Filter has a very good storytelling.

The brand talks about nanofiltration, Swiss technology, 0.002 μm, 18 L/h, TFA, viruses, nanoplastics and 4,000 liters.

But when we compare the publicly available evidence, several major promises are not demonstrated at the level at which they are claimed.

⮕The TFA is displayed as "not detected", but the full Bureau Veritas report is not publicly visible and the 5,000-liter PFAS test is announced as still in progress.

⮕Viruses are claimed, but the public board shows bacteria, not an identifiable virus test. No MS2. No rotavirus. No surrogate norovirus. No PFU unit. For a microbiological promise, this is a real problem.

⮕ Nanoplastics are being talked about, but Nano says that testing <1 µm sont encore en cours.

Pesticides and drugs are tested on a few molecules, while Coldstream publishes a much wider panel.

⮕ And on traceability, Nano highlights Switzerland, while the actual manufacture of the cartridge must be clearly assumed and documented.

The fastest filter is not necessarily the safest.

In water purification, it's not the speed that counts at the end: it's the contaminants actually tested, the methods used, the volume filtered before testing, the detection limits, and the transparency of the reports.

Based on the available public evidence, Coldstream FTO+ remains today the most solidly documented filter of this comparison.

 

Update 14/07/2026

Since the publication of this article, Nano Filter has put online a much more detailed "Evidence" page: Consulted on 14/07/2026, with a downloadable PDF by family of contaminants (PFAS, heavy metals, pesticides, pharmaceutical residues, microplastics, chlorine, bacteria).

Let's start with what has actually changed, with the supporting documents — and what that change actually proves, or doesn't.

What is now public, and was not in May:

• Out of several families, Nano now publishes two measurement points: when the cartridge is started, and after 5,000 liters, i.e. "1,000 liters beyond the guaranteed duration" in their own words. Sources: Heavy metal ratio (ref. NBF25-0023352-02 / NBF24-0013449-01), Pathogen Report (same ref.), Pesticide Report, , Pharmaceutical Residues Report and Chlorine and by-product ratio.
• A detailed figure on TFA is now visible, with five short- and ultra-short-chain substances, LC-MS/MS method and detection limits indicated. Source: short-chain PFAS report, reference 424-1241/A.

But making visible is not the same as producing. On closer inspection, these PDFs do not document new essays: they reformat already existing reports.

The heavy metals report and the pathogens report cite exactly the same two references, NBF25-0023352-02 and NBF24-0013449-01. The long-chain PFAS report also cites NBF25-0023352-02. In other words, several of these "contaminant-by-contaminant" sheets, presented as separate documents, come from the same two laboratory reports — simply reclassified by contaminant family on separate PDFs. This is not a new test campaign. It is the same material, cut differently.

Another clue: our May article cited, for the TFA, the reference "Bureau Veritas Report 424-1242/A". The new July PDF, on what seems to be the same test, indicates "424-1241/A" — a number of discrepancies. An isolated typo would prove nothing in itself, but it points in the same direction as the rest: this republication was obviously done quickly, not as part of a new rigorous test campaign.

In digging into these same documents, several additional elements deserve to be noted.

Still no viral test, despite a virus promise maintained

The new section is entitled 'Bacteria and micro-organisms' (See the page, the corresponding section, and the Pathogen Report which is related to it). Three organisms are tested there: Escherichia coli, , Pseudomonas aeruginosa, , Staphylococcus aureus. They are bacteria. No results in PFU units, no coliphagous MS2, no rotavirus or recognized viral substitute appear in the document.

However, the Technology page de Nano still claims that upon contact with the membrane, "a bacterium, a virus or any other pathogen" has its envelope destroyed. The brand had the opportunity to fill this point by completely redesigning its evidence page. It did not do so. The word "virus" remains a commercial promise, without a test table that corresponds to it.

The new figures show a deterioration, while the discourse promises the opposite

By publishing results at 5,000 liters, Nano made public a piece of information that could not be verified in May: the performance of the cartridge really decreases with the filtered volume, on several parameters.

Contaminant	Startup Reduction	Reduction after 5,000 L	Source
Staphylococcus aureus	> 99.99%	96,33 %	Pathogen Report
Pseudomonas aeruginosa	> 99.99%	99,40 %	Pathogen Report
Escherichia coli	99,98 %	99,69 %	Pathogen Report
Cadmium	> 99.60%	98,84 %	Heavy metal ratio
Carbofuran (pesticide)	> 97.87%	95,00 %	Pesticide Report
Chloroform	> 99.93%	99,15 %	Chlorine and by-product ratio

All these values are taken as they are from the documents published by Nano Filter itself; none of them are an estimate or extrapolation on our part.

In Staphylococcus aureus, we go from a reduction of more than 4 log to a reduction of just over 1.4 log at the end of life announced. This is a real decrease, not a rounding detail.

This data directly contradicts the promise of a performance that would remain the same from the first to the last liter, as displayed on the Maintenance page of the brand, which speaks of a "constant efficiency and flow from the first to the last liter". You can't both publish a measured drop and continue to sell the absence of degradation as a differentiating argument.

Still no original lab report published

Each PDF posted online by Nano bears the same mention: the document was prepared by Nano Filter SA, it faithfully reproduces results issued by a third-party laboratory, and the original report "can be consulted on request" (identical formula on the six documents: PFAS Short Chain, , PFAS Long Chain, , Heavy metals, , Pesticides, , Pharmaceutical residues, , Pathogens). As of the date of this update (14/07/26), no original laboratory reports, including letterhead, scope of accreditation, detailed methodology, and complete test conditions, are publicly available on the pages visited.

This is objectively more detailed than in May. But it is still not a searchable laboratory report — it is an in-house synthesis that claims to be faithful to an original that you cannot verify yourself.

An inconsistency between the "certifications" page and the TFA report itself

The certifications section of the Nano page (block "ISO/IEC 17025 accredited analysis reports") states: "All tests are carried out by SGS, an independent ISO/IEC 17025 accredited laboratory."

The TFA and other short-chain PFAS PDF, referenced 424-1241/A, however, indicates something else. The "Analysis" field mentions an "independent Swiss laboratory, ISO 17025 accredited", without naming SGS. And the field "Commissioned by" reads: "Commissioned by: Evodrop AG, on behalf of Nano Filter SA."

Two things to learn from it, cautiously:

• On the contaminant presented by Nano itself as the most difficult to filter, the mark announced on its own document is not the one displayed on the certification page.
• This is the first time that we find, in black and white on a document published by Nano Filter, confirmation that Evodrop AG is directly involved in their trials — not just a hypothesis built from a trademark application to Swissreg, as mentioned earlier in this article.

This does not allow us to conclude that the results are false: we do not have access to the full report, so we cannot say anything about its validity. On the other hand, it allows us to say, factually, that Nano's communication on its laboratories is not consistent from one document to another.

The TFA remains without a test of resistance over time

Unlike the other contaminant families, which now have two measurement points (starter and 5,000 L), the short-chain PFAS table shows only one of them, without specifying at what stage of the cartridge's life cycle the sample was collected — compare with the PFAS Long Chain Report, which specifies at least "measurements made at the start of the filter's life cycle". The contaminant that Nano itself presents as the most difficult to retain is also the one for which no end-of-life performance data is still available.

In summary

Nano Filter hasn't relaunched any new trials since May: it has republished, in a more extensive format and more contaminants displayed, data from the same reports already cited in May — at least two of them visibly recycled from one sheet to another. More visible numbers don't mean more evidence: that's still zero original lab reports that can be consulted, zero viral tests despite a virus promise being maintained, a documented inconsistency between what Nano says about its labs and what its own TFA report says, and now a report reference that no longer matches the one cited in May. The new 5,000-liter data also shows a real degradation in performance on several parameters — which contradicts the promise of an efficiency that would never decrease.

According to the public evidence available to date, Coldstream FTO+ remains the most solidly documented filter of this comparison.

Update 23/07/2026

Nano's "Evidence" page states that original lab reports "can be viewed upon request." We took the brand at its word.

Method. Two independent individuals, under two separate email addresses, contacted Nano Filter's customer service two days apart (July 21 and 23, 2026), with a neutrally and generally worded question, without citing any laboratory or report references: "You mention lab tests performed on your filters, would it be possible to receive full reports of these tests?"

Result. Both times, Nano responded quickly and forwarded the exact same two documents:

• A Labor Veritas AG report No. 424-1242/A, on five short- and ultra-short-chain PFAS substances (including TFA).
• An SAS Accreditation Certificate for Labor Veritas AG — a document that certifies that the laboratory is ISO/IEC 17025 accredited, but does not contain any test results or performance data on any product.

In both cases, there is no trace of the two SGS reports (referenced NBF24-0013449-01 and NBF25-0023352-02) that the Nano page cites by name for heavy metals, pathogens, long-chain PFAS, pesticides, pharmaceutical residues and chlorine. Of the seven families of contaminants publicly displayed by the brand, six are based on these two SGS reports. Neither of them has been sent to either application.

What the report transmitted is really worth. A laboratory report is not serious simply because it contains figures. It is limited to a sample "before filtration" and a sample "after filtration" — a single pass, a single measurement point on each side. No volume of filtered water is indicated: it is impossible to know whether the "after" sample corresponds to 1 litre, 50 liters or 500 liters passed through the cartridge. No replication either: no duplicate, no triplicata, no indication of variability. And above all, no shelf life level — nothing at 1,000 L, 4,000 L or 5,000 L, while Nano has just published this type of level for other families of contaminants. On the contaminant that the brand itself presents as the most difficult to retain, there is therefore no data on its resistance over time.

It is not a filter validation protocol in the sense of the NSF/ANSI or IAPMO standards, or even the type of step-by-step test that Nano now applies to its own heavy metals or pathogens tables. It is a chemical feasibility test in a single pass: can the membrane reduce these five molecules in a given sample once? The answer is yes. But this format does not allow us to draw any conclusions about the capacity of the cartridge, its announced lifespan, or its actual performance in use.

This is confirmed by the document, once consulted live. The report states that the device under test is called "Filtrationsgerät EVOdrink", commissioned by Evodrop AG (for the attention of Fabio Hüther), sampled on 05/11/2024 and reported on 29/11/2024. The name "Nano Filter" does not appear anywhere in this document. The only test report that the brand actually submitted, on two separate applications, therefore relates to a product that is not its own, from a single-pass test, without indicated volume, carried out almost twenty months before this update.

Which we cannot conclude. Two similar requests, over two days, do not support the assertion that Nano routinely refuses to transmit SMS reports to anyone who requests them. A different wording, or an identified business address, might get a different response. What these two requests show with certainty is that when faced with a neutral and general question, to date, the promise of "reports available on demand" has not been honored in its entirety — and that the only document actually transmitted is, methodologically, a one-off test without volume or replication, on a device that does not bear the name of the brand that brandishes it.

Since the publication of this article, we have requested and obtained the full test reports of Nano Filter. We have read them in full. The following is not based on any interpretation: only on what these documents themselves indicate.

Two pieces are at stake. On the one hand, a chemical test report from the Finnish laboratory Measurlabs, reference 29505, commissioned by Nano Filter SA, samples received on July 3, 2026 and report issued on August 10, 2026. On the other hand, a document dedicated to microorganisms, presented as containing SGS results.

The chemical report relates to five filtered litres

This is the central point, and it is written in black and white, in each section of the Measurlabs report: "with a filtered volume of 5 L".

Five liters. For the 13 pharmaceutical residues, for the 20 PFAS, for the 157 pesticides and metabolites, for the metals, for free chlorine. The entire analytical matrix put forward by the brand was measured on five litres of filtered water.

The cartridge is announced for a guaranteed 4,000 liters, with a claimed endurance of 5,000 liters. Five liters therefore represents about a thousandth of the product's announced lifespan. There are no intermediate measurements, no steps, no verification of what happens to the performance when the cartridge ages.

This is not a long-term performance test. It is a feasibility test on a new cartridge. Both have their uses, but they do not answer the same question, and it is not the one asked by a user who keeps his cartridge for several months.

Samples were collected by the customer

The report also states, in full: "Sampling was performed by the customer."

Measurlabs is an accredited laboratory, and we don't question that. But in this configuration, thehe laboratory did not mount the filter, did not circulate the water, did not control the flow rate or the conduct of the test. He has analysed many of thesamples sent to him, according to a procedure provided by the client, which is reproduced in the report and which refers to an installation video.

This is different in nature from a performance test conducted entirely in the laboratory, where the cartridge is installed in the laboratory bench and the water is circulated under its control, as is the case with the Envirotek and QFT reports on which Coldstream relies.

The water used is pure doped water

For pharmaceutical residues, PFAS and chlorine, the report indicates pure water spiked with contaminants. For metals, spiked tap water.

This deserves an explanation, because the point is technical but decisive. NSF/ANSI performance protocols require a defined challenge water: pH, TDS, turbidity, and above all a total organic carbon between 1.0 and 2.0 mg/L. This requirement is not decorative. The organic matter naturally present in real water competes with contaminants to occupy the carbon adsorption sites. Pure water is, from this point of view, the most favorable condition that exists for a filter.

The Measurlabs report does not characterize the test water. The Envirotek reports used for Coldstream cartridges indicate temperature, pH, TDS, turbidity, and TOC at each of the measurement steps.

The trial schedule tells the same story

The report indicates that the samples were received on July 3, 2026 and that the report was issued on August 10, 2026. A little more than five weeks, most of which corresponds to laboratory analysis time and writing.

This delay is perfectly consistent with what the report describes. Filtering five liters takes a few hours. There is no campaign to drive, no cycles to repeat, no cartridge to age.

A step-by-step campaign does not work like this. Several thousand litres must be passed through the cartridge at a controlled rate, with use and rest phases, and sampling at each stage. The reference report for the CF163W cartridge includes measurements at 500, 1,000, 1,500, 2,000, 2,500 and 3,000 litres. Comparable reports for other ceramic cartridges have eleven sampling points up to 3,030 litres. This type of campaign is counted in months, not weeks.

The time frame for a trial is therefore not administrative information. It is a fairly reliable indicator of what has actually been done.

What their own figures show

We do not dispute the published values. We read them.

Of the PFAS, filtered water still contains 0.0393 μg/L PFBS, 0.012 μg/L PFBA and 0.009 μg/L PFOA from the fifth litre onwards. The sum of the 20 PFAS is 0.060 μg/L. As a benchmark, EU Directive 2020/2184 sets the limit for the sum of 20 PFAS at 0.100 μg/L in water intended for human consumption. In other words, over five litres and in pure water, the effluent already reaches 60% of this value, and it is the short-chain compounds that pass through, which is consistent with the known behaviour of activated carbon towards these molecules.

On TFA, which is often presented as 100% withdrawn, the report indicates doping at 1 μg/L and a result below the limit of quantification of 0.000025 mg/L. This corresponds to a reduction of more than 97.5%, not 100%. The wording "100%" reflects an undetected result, which is not the same thing. And again, on five litres.

On pesticides, two compounds show a measurable passage from this stage, chlormequat and picolinafen.

On metals, the table in the report is entitled "removal / leeching", i.e. removal and leaching, and leaching is indeed present: phosphorus goes from 35 to 184 μg/L after filtration, iron from 12 to 24 μg/L, and sodium, potassium and calcium also increase. Several reductions remain partial, including uranium at about 50%, arsenic at about 41%, antimony at about 40% and manganese at about 23%.

We point out these figures without drawing excessive conclusions: they are measured over five litres, and they say nothing about what happens next. That is precisely the problem.

The Microorganisms document is not the SMS report

On the form, we have to be precise, because the distinction counts.

The document that is circulating is not a report issued by SGS. It is a document formatted by Nano Filter SA which reproduces results attributed to SGS. The document itself says so in its footnote: it was prepared by Nano Filter SA, it reproduces the analytical results of an original test report issued by SGS, and this original report can be consulted on request. We do not doubt the good faith of this approach, but the primary document is not the one that is distributed.

This document also bears two separate references in the header, NBF25-0023352-02 and NBF24-0013449-01. These are two different test campaigns, of two different financial years, the results of which are presented side by side in the same table. The input concentrations differ between the column "at start-up" and the column "after 5 000 L", which confirms that this is not a continuous test followed over time on the same cartridge, but two measurements close together a posteriori.

Finally, the methods cited are ISO 6222 and ISO 9308. These are methods for counting microorganisms in water intended for human consumption. These are excellent analytical methods, but they are not purifier challenge protocols. A protocol like NSF/ANSI P231 defines something else: the minimum concentration of the inoculum, the composition of the challenge water, the cycles of use and rest, the number of units tested in parallel, the duration of the test and the success criteria expressed in logarithmic reduction.

What this document shows despite everything

It's the most interesting piece of the file, because it has two life points and it documents what happens in between.

At the beginning, the results were excellent: more than 99.99% on Pseudomonas aeruginosa and Staphylococcus aureus, 99.98% on Escherichia coli.

After 5,000 litres, the same parameters give 99.69% for E. coli, 99.40% for Pseudomonas and 96.33% for Staphylococcus aureus.

96.33% corresponds to approximately 1.4 log reduction. To put it in perspective, the microbiological reference protocols applied to gravity purifiers require 6 logs on bacteria. The residual values indicated at the filter outlet after 5,000 liters are 98 CFU/mL for E. coli, 1,100 CFU/mL for staphylococcus and 4,500 CFU/250 mL for Pseudomonas.

These are their figures, in their document. They show a significant drop in microbiological performance between the new cartridge and the announced end of life.

No virus tests appear in the documents provided

We noted this in the initial version of this article based solely on public information. The full reports confirm it: three bacteria were tested, and no viruses.

This is an important point, because viral retention is the most demanding criterion for a gravity filter, and because it cannot be deduced from an announced pore size.

For comparison, the Coldstream FTO+ cartridges have data obtained according to the NSF P231 protocol on rotavirus, with an influent of 24,642,000 units and effluents measured between 80 and 610, as well as on Klebsiella with an influent of 328,646,000 CFU and effluents measured between 74 and 360. These measurements were carried out in increments up to 3,000 litres.

Why everyone posts width and so few people publish duration

There is a simple economic reason, and it explains a large part of what we see in this market.

Expanding a list of analytes costs almost nothing. Going from fifty to one hundred and fifty pesticides sought is the same injection on the same device, with a longer list of compounds. The figure displayed on the product sheet doubles, the cost of the trial increases little.

Adding bearings is expensive. You need a test bench that has been mobilized for weeks, considerable volumes of water, cartridges sacrificed, and repeated sampling and analysis at each stage.

This is why a very wide table measured only once is not equivalent to a narrower table measured eleven times. The number of contaminants tested indicates the extent of the scan. The number of litres indicates reliability over time. These are two different pieces of information, and only the second corresponds to the actual use of a domestic filter.

What we publish on our end, including limits

The Coldstream FTO+ cartridges correspond to the CF163W model. The available tests were conducted by Envirotek and QFT, accredited laboratories, according to NSF/ANSI protocols 42, 53, 401 and P231, with measurements carried out at 500, 1,000, 1,500, 2,000, 2,500 and 3,000 liters, on two cartridges tested in parallel and with a challenge water characterized at each level.

We also publish what is less favourable, because a comparison in which a single product is good on everything would have no value. On the reference ratio, the reduction of fluoride is low, that of nitrates and nitrites remains partial, and total arsenic does not allow for a strong claim. We therefore do not present them as performances.

On PFAS, our position deserves to be explained in detail, because it illustrates exactly the method defended by this article.

We do not currently have a PFAS report prepared on behalf of Coldstream and specifically relating to the FTO+ cartridge. A test campaign is expected. As long as it is not available, we do not make any PFAS claims on this cartridge.

However, we do have PFAS data, and we publish them without any ambiguity about their origin. This is the QFT Laboratory report number 20-548, on the ProOne G2.0 cartridge, a third-party product with related technology, the same cartridges under a different trade name. This test has six measurement points spread over the lifetime, up to 200 gallons, or about 757 liters, with reductions generally greater than 90%. We also publish the limits: performance varies according to the compounds and according to the steps, GenX for example going down to 60% at an intermediate stage. We also specify that the retained input concentration, 0.1 μg/L, associated with a limit of quantification of 0.01 μg/L, mechanically caps the demonstrable reduction at 90%.

We therefore present this document for what it is: a useful technical point of comparison, drawn up on behalf of a third party, and not a test specific to the FTO+. We do not turn it into a commercial argument on our own cartridges.

This is a less comfortable position than a percentage displayed without origin. It is also the only one that allows the reader to judge for himself. And the difference in method remains telling: even this report, which we do not claim to be ours, is based on six measurements spread over nearly 760 liters.

The only question to ask, regardless of the brand

This case is not just an opposition between two trademarks. It illustrates a reading method that any buyer can apply, including to our own documents.

In front of a performance table, four questions are enough.

On how many liters was the measurement made, and at how many different points.

Who conducted the test, and who collected the samples.

What water was used, and are its characteristics indicated.

What performance standard was applied, not just what method of analysis.

A laboratory accredited to measure a molecule is not the same as a performance test conducted according to a recognized protocol. Both are legitimate. They do not prove the same thing.

The reports mentioned here can be consulted from their respective publishers. We have our own reports available on request.