123 multimedia SASU and the Capital Recovery Test Behind Local Network Control

Summary

• 123 multimedia SASU is a Toulouse publisher of community, chat and dating services, not a conventional retail Internet provider. Its public offer combines consumer brands such as Tchatche and Babel with a white-label social platform, round-the-clock operation and moderation, advertising, subscriptions and services for media and telecom companies.
• The current legal company was incorporated in October 2013 after Philippe Pisani led a EUR 200,000 purchase of Index Multimedia's business assets. It inherited brands and operating know-how from a much older business, but that lineage should not be confused with continuous ownership of the predecessor's balance sheet, contracts or infrastructure.
• The company is a RIPE NCC member and holds 185.40.100.0/23, equal to 512 IPv4 addresses, plus 2a01:4f20::/32 in IPv6 space. The IPv4 block is publicly routed by Eurofiber France's AS39405 with valid route-origin authorisation; the IPv6 allocation had no visible public route at review time. No autonomous-system number was found under the company's RIPE organisation record.
• That footprint provides useful administrative control and potential portability, but it is not evidence that 123 multimedia owns fibre, sells transit or operates a self-sufficient network. The company's own legal notice names FullSave, now part of Eurofiber France, as its hosting provider, making supplier execution and contract terms central to the continuity claim.
• The last detailed public accounts cover an 18-month period ending June 2019: EUR 1.01 million of revenue, EUR 123,857 of operating profit, EUR 64,408 of net profit and EUR 938,089 of assets. Accounts through June 2025 have been filed with confidentiality declarations, so current revenue, cash flow, capital expenditure, customer concentration and profitability remain undisclosed.
• The provisional judgment is that the address resources are defensible continuity infrastructure, but the capital-recovery case is unproved. A positive conclusion requires current evidence that recurring gross contribution covers hosting, staffing, moderation, security and renewal; that the white-label business retains customers; and that direct control produces measurable availability or migration benefits unavailable from a managed cloud setup.

Toulouse is an operating boundary, not a moat

Geography matters here because it defines both the advantage and the constraint. 123 multimedia SASU is based in Toulouse, and the company's public documents, RIPE membership entry and French corporate record all converge on that location. Its registered office moved to 41 impasse de la Flambere in December 2024. Its declared activity is other telecommunications, with a more descriptive statement covering community services, games, digital products, and supervision of the network supporting those services. The address places the company inside a substantial regional technology economy, close to a local data-centre and fibre supplier. It does not give the company exclusive access to its users or customers.

The service itself crosses geography. A chat session does not become more valuable because its servers are close to the publisher's office. Users care about whether the service loads, whether enough other people are present, whether conversations feel safe, and whether the free or paid experience is preferable to the next application on the phone. A media group considering a white-label community service can procure software and hosting from outside Occitanie. A telecom company can build, buy or abandon a branded community product. Global cloud providers offer French regions, while French providers sell compute and storage by the hour. Locality can improve response times when engineers or equipment need attention, but the customer can still compare a Toulouse arrangement with national and global substitutes.

That makes scale an economic constraint before it becomes a strategic choice. A local operating team can be close to its supplier and retain direct knowledge of a mature code base. It also has to spread security, moderation, software maintenance, app-store support and infrastructure expense across a smaller revenue base than the largest dating groups. The cost of one senior engineer, one overnight moderation shift or one redundant hosting environment does not fall proportionally because the publisher is small. Unless those fixed commitments protect a valuable audience or a sticky business contract, control becomes an expense in search of a return.

The company's RIPE membership can easily be misread in this setting. It is real evidence that 123 multimedia administers Internet number resources. It is not evidence of a local access network. There is no public tariff for household broadband, no disclosed fibre footprint, no visible claim that the company sells IP transit, and no autonomous-system number linked to its RIPE organisation entry. The commercial boundary shown by the company's own materials is a digital-service publisher and platform operator. Any valuation of its network position has to begin there.

A EUR 200,000 restart inherited brands, not certainty

The legal company is much younger than the 123 Multimedia name. French corporate data identify SIREN 798 073 375, a single-shareholder simplified joint-stock company incorporated in October 2013 with EUR 50,000 of capital and Philippe Pisani as president. The company's own account says Pisani, previously an executive of Index Multimedia, led the October 2013 acquisition with former employees.

Court records add the price and the limits of continuity. The Toulouse Commercial Court approved a EUR 200,000 sale of Index Multimedia's business assets on 17 October 2013, after Index had entered court-supervised restructuring and then liquidation. The new single-shareholder company signed its statutes six days later. A later ownership dispute concerned money advanced to finance the acquisition and working capital; the Toulouse Court of Appeal rejected the claim that 90% of the new company's shares had to be transferred, and the Court of Cassation dismissed the subsequent appeal.

The history matters because it explains how a small legal entity could possess recognisable brands, a large legacy code base and a technically demanding service from its first days. The predecessor business had moved from telematics and premium-rate services into mobile content and online communities. Tchatche dates from the earlier era. In a 2015 interview, Pisani said the inherited service was handling 15 million exchanges a day and that infrastructure and applications required substantial renovation. Those statements are old and management-supplied, but they make the transaction legible: the buyer did not start with an empty company and then discover network operations. It acquired a distressed operating estate whose value depended on keeping traffic, software and commercial links alive.

That purchase price should not be mistaken for replacement cost. EUR 200,000 was the court-approved price for distressed business assets, not a market appraisal of a newly built community platform, a modern moderation operation or a redundant hosting estate. The later court record says a Luxembourg company advanced EUR 550,000 for the purchase and working capital, which was later repaid. The additional funding illustrates the difference between acquiring assets and financing continued operation. Software has to be rewritten, user support staffed, suppliers paid and products refreshed even when the legal title has already changed hands.

Nor should the predecessor's past scale be carried into the current company without evidence. The old business once reported tens of millions of euros in revenue and belonged to a listed Japanese group. The new company is a different legal entity formed around a much narrower asset purchase. Its later filings, staff indicators and public product set describe a small business. Historical brand recognition may lower the cost of attracting some users and reassure old commercial partners. It does not prove that today's revenue, audience or margins resemble those of the predecessor.

The restart nevertheless created one valuable strategic option: the company could preserve old brands while making the infrastructure fit a smaller base. A distressed estate can produce attractive returns if the acquirer strips out overhead, keeps the loyal audience and avoids rebuilding everything. It can also become a maintenance trap if legacy components absorb skilled labour while modern competitors reset user expectations. The 2013 acquisition makes capital discipline the right lens. Every inherited asset must still earn its present cost of care.

The product is community infrastructure sold in two directions

123 multimedia's public business has two sides. The first is direct-to-consumer publishing. Its site presents Tchatche and Babel as owned community brands, and its Tchatche Google Play listing describes a French-language chat and dating application with advertising and in-app purchases. The same developer page points to additional brands, including Amitie and Cybermen. Babel remains available as an anonymous chat service. This portfolio gives the company several entry points into the same underlying capabilities: profiles, location, messaging, content review, advertising inventory and paid features.

The second side is business-to-business. The company website markets a white-label social platform to organisations that want to keep their audiences within a branded community. It lists development, user experience, integration, 24-hour operation, multilingual content validation, advertising, premium payments and audience analysis as parts of the offer. It names NRJ, Trace TV, La Depeche du Midi, Bouygues Telecom, Orange and Telecoming as customers or references.

That list demonstrates positioning, not current contract value. The page does not date the relationships, name current annual fees, distinguish live deployments from historical work or disclose renewal terms. Some press links on the same site date from 2014 to 2017. A buyer should therefore treat the names as evidence that the company has served significant media and telecom brands, while withholding any assumption that each relationship remains active or material in 2026.

The two directions can reinforce each other. Consumer brands give the company live operating experience. They expose moderation rules, messaging performance, device compatibility and monetisation decisions to real users. White-label customers can then buy a platform shaped by those lessons rather than commissioning a fresh product. Business contracts can provide steadier revenue than advertising, while consumer traffic gives the platform technical credibility.

They can also compete for scarce resources. A feature that raises subscription conversion on Tchatche may not matter to a media customer. A bespoke integration for a large client can consume engineering time that would otherwise improve the owned applications. Content standards differ by brand and audience. A consumer service can tolerate a product experiment that a telecom customer rejects. The economic benefit appears only if the underlying platform remains standard enough that each new customer adds contribution rather than another privately maintained version.

The company says its platform is standardised and claims service quality above 99.9%. It does not publish the measurement window, exclusions, service credits or independent monitoring behind that figure. At 99.9%, simple arithmetic still permits roughly 8 hours and 46 minutes of downtime in a 365-day year. More important, an overall availability percentage can hide the difference between a brief quiet-period fault and an outage during peak conversation time. A serious white-label buyer would need contractual definitions, incident history, recovery objectives and evidence of physical and supplier diversity, not only the headline.

Free attention must become paid contribution

The consumer model begins with free access. Tchatche's terms say the main functions are free and that advanced features are sold through automatically renewing packs on the mobile applications. Prices and pack contents can vary with connection method, promotions and loyalty offers, and the final price appears during payment. Paid voice functions are also available. The Google Play listing confirms advertising and in-app purchases; in a June 2024 response to a user, the company said an advertising-free option cost less than EUR 2 a month at that time.

That low visible price is revealing even though it is not a complete tariff. A EUR 2 monthly feature produces at most EUR 24 a year before consumption tax, app-store charges, payment effects, support, fraud, moderation, hosting and marketing. At a hypothetical 10,000 continuously paying users, gross billings would be below EUR 240,000 a year. At 50,000, they would be below EUR 1.2 million. These are illustrations, not estimates of current subscribers. They show why conversion, retention and product mix matter more than a download count.

Advertising broadens the paying base because free users can produce revenue without subscribing. In 2015, Pisani said advertising and freemium represented about 80% of receipts, with operator and media relationships then described as marginal. He also cited roughly two million monthly users and hundreds of millions of page views. Those figures are useful history but too old for a current valuation. They predate major changes in mobile privacy, advertising identifiers, app-store rules, European platform regulation and the competitive dating market.

The current public audience signals are narrower. Google Play showed more than one million cumulative downloads and about 15,800 reviews for Tchatche at review time, with a rating around 3.3 out of five in the French-facing listing. The application was updated in December 2025. Cumulative downloads confirm that the product has reached a meaningful number of devices over its lifetime; they do not reveal installations still active, daily users, geographic mix, retention, advertising impressions or paying conversion. Reviews are self-selected and can lag the current version.

Unit economics turn on a chain of ratios the company does not publish. The first is acquisition cost per new active user, including advertising and app-store discovery. The second is the share who complete a credible profile and find enough relevant people to remain. The third is advertising revenue per free user and paid conversion per retained user. The fourth is churn after a subscription begins. The fifth is the cost of reviewing content, handling reports, supporting payments and preventing abuse for each active cohort.

A dating and chat service has a particularly unforgiving feedback loop. More genuine users improve the chance of a useful conversation, which improves retention and attracts more users. Fake profiles, harassment, broken image uploads or thin local participation do the reverse. Cutting moderation can reduce expense this month while damaging the network value that earns next month's revenue. Spending heavily on acquisition can inflate downloads while bringing users who never form a durable habit. Visible growth is not value creation unless the incremental cohort produces more lifetime gross contribution than it costs to acquire and serve.

The white-label side has different unit economics. A business customer may pay an implementation fee, a recurring platform charge, usage-linked fees, or a share of advertising and premium revenue. The company does not publish that structure. The attractive version is a standard platform with modest setup work, a multi-year contract and shared improvements across customers. The unattractive version is a sequence of bespoke projects priced to win recognisable names but burdened with permanent support commitments. Contract-level gross margin, renewal rates and engineering hours would distinguish the two.

The last open accounts show profit, not present earning power

The latest detailed public financial statements are too old to settle the case, but they establish a useful baseline. Accounts summarised by Le Figaro cover the 18 months ended 30 June 2019. They show EUR 1,012,682 of net revenue, EUR 1,441,875 of total operating income, EUR 1,318,018 of operating charges, EUR 123,857 of operating profit and EUR 64,408 of net profit. The operating margin on net revenue was about 12.2%, while net margin was about 6.4%.

The 18-month period complicates comparison with normal annual results. It also contains EUR 429,193 of operating income beyond reported net revenue, a substantial amount whose recurring character cannot be assumed from the summary. The company was profitable, but the public numbers do not show whether ordinary customer sales alone supported the same margin.

The balance sheet ended the period with EUR 938,089 of assets, including EUR 598,818 of fixed assets and EUR 339,271 of current assets. Equity was EUR 487,869 and debt EUR 450,220. Fixed assets represented about 63.8% of the total, while debt represented about 48.0%. For a digital publisher with five reported employees, that is not a purely weightless business. The fixed-asset line could include software, equipment or other long-lived items, but the summary does not provide enough detail to allocate it confidently.

The staffing indicators point to a lean legal employer. Le Figaro reports five employees in 2019; Pappers gives a band of three to five for 2022. The company website says it mobilises about 100 subcontractors, mainly in France. The claim is undated and unaudited, yet it fits a model in which a small central company buys substantial development, moderation or support capacity from outside providers.

That structure can protect cash. Contractors can be scaled by shift, language, product and customer demand. The company avoids carrying every specialist as fixed payroll. It can also weaken operating leverage if suppliers capture much of the margin, if quality varies or if the company must keep paying for 24-hour coverage before revenue arrives. A three-to-five-person employer cannot personally staff every hour, language and technical function described on the website. The quality of supplier agreements is therefore part of the business, not an incidental purchasing detail.

Newer filings do not resolve the question. Pappers records annual-account deposits for years through 30 June 2025, but the recent filings carry confidentiality declarations available to qualifying small companies under French rules. Public readers cannot see current revenue, profit, cash, debt or assets. Confidentiality is not evidence of weakness. It simply prevents claims that the 2019 margin survived later changes in advertising, mobile distribution, privacy requirements, supplier pricing and competition.

The missing cash-flow bridge is especially important. Accounting profit does not show spending on software renovation, servers, security work or prepaid supplier commitments. It does not show whether advertisers pay slowly, whether app-store receipts settle quickly, or whether white-label clients fund implementation in advance. It does not show capitalised development. For capital recovery, operating cash after necessary product and infrastructure renewal matters more than the reported profit of an 18-month period seven years ago.

The address footprint is real and deliberately limited

The clearest current infrastructure evidence comes from RIPE records. The RIPE NCC member page lists 123 multimedia SASU as a French member at the Toulouse address. The underlying organisation record, ORG-MS231-RIPE, carries the same legal registration number and identifies the company as a Local Internet Registry.

An inverse lookup on that organisation record returns two allocations. The IPv4 allocation is 185.40.100.0/23, a block of 512 addresses. The IPv6 allocation is 2a01:4f20::/32, an address space vast enough for many internal subnet assignments. Both are provider-aggregatable resources allocated through the RIPE system. The organisation record was created in November 2013, within weeks of the new company's formation.

At the review date, RIPEstat showed the IPv4 /23 visible through origin AS39405, Eurofiber France. The route was seen by 326 of 327 relevant RIPE RIS peers. A separate route-origin validation query returned valid, with authorisation for AS39405 to originate the /23 and routes as specific as /24. The equivalent IPv6 routing query showed no visible origin.

Those records support several conclusions and reject several others. The company controls an allocation large enough to assign stable public addresses to separate services, hosts or security zones. It has maintained the registry relationship since its formation. The active IPv4 route has wide public visibility, and the valid route-origin status reduces one class of accidental or malicious routing rejection. The unannounced IPv6 allocation preserves future addressing capacity but does not demonstrate current IPv6 service.

The same evidence does not establish ownership of routers, fibre, racks or data-centre space. The company does not originate its own route. It depends on AS39405 for public reachability, and the RIPE role record includes Eurofiber France as a maintainer alongside 123 multimedia. No autonomous-system number appears in the organisation lookup. The footprint is best described as customer-controlled address space carried by an upstream infrastructure provider.

That is not a defective arrangement. Many companies want stable address administration without operating a public autonomous system or negotiating multiple transit relationships. The upstream can provide routing expertise, denial-of-service protection, physical connectivity and on-call operations at greater scale. The customer retains an address asset that may simplify server grouping, allow-listing, reputation management and a future migration.

But partial control must be priced as partial control. If Eurofiber's route, facility or support fails, ownership of the allocation alone does not restore service. If the hosting agreement makes the addresses difficult to move, the theoretical portability may not become practical portability. If the application already sits behind a content-delivery or protection service, most users may never see the allocated block directly. If only a small share of the 512 IPv4 addresses is active, the unused capacity has option value but not necessarily current earnings.

The 2026 RIPE NCC annual contribution is EUR 1,800 per LIR account, before any applicable additional resource charges. That fee is modest beside engineering and hosting, but it is a recurring commitment. The larger cost is administrative care: accurate contacts, abuse response, route authorisation, security, monitoring and migration planning. The economic test is not whether 512 scarce IPv4 addresses sound valuable. It is whether administering them reduces the total cost and risk of delivering the company's products.

FullSave supplies the route from local control to the wider Internet

The company's legal notice names FullSave as the provider responsible for direct and permanent storage. That disclosure aligns with the network evidence: AS39405 is the FullSave lineage now registered to Eurofiber France, and it originates 123 multimedia's IPv4 block. The supplier relationship is therefore visible at both the hosting and routing layers, although the public documents do not reveal the contract, capacity, redundancy or exact facility used.

FullSave was founded in Toulouse and built regional fibre, hosting and cloud operations. Eurofiber announced its acquisition of FullSave in 2020, describing a 600-kilometre fibre network expected by the end of that year and a 1,600-square-metre data centre in Toulouse. This gives 123 multimedia access to an infrastructure owner much larger than itself while preserving local proximity.

That supplier can create genuine economies. Routing a /23, running resilient power and cooling, maintaining upstream connectivity and staffing a facility are scale activities. 123 multimedia avoids replicating them. A local supplier can also shorten the distance for equipment visits and make escalation more direct than a purely remote arrangement. The company retains its product and address knowledge while buying the heavy physical layer.

Concentration remains the downside. FullSave appears in both the hosting disclosure and the route origin. The public record does not show a second origin, second hosting provider or active IPv6 path. Logical reach through a well-connected supplier may be strong, but it is not the same as independently controlled physical diversity. A single contract, operational team or commercial transition can affect several layers at once.

Two recent supplier developments make that risk concrete without proving harm to 123 multimedia. First, Eurofiber France disclosed a November 2025 security incident affecting its telecom ticketing platform and cloud customer portal. It said a software vulnerability was exploited and data associated with the affected platforms were taken, while telecom and cloud services continued. There is no public evidence that 123 multimedia data were involved. The incident matters because supplier support portals and account records are part of the attack surface even when the hosted application remains online.

Second, in June 2026 Eurofiber and ETIX announced exclusive negotiations over the intended sale of four Occitanie data-centre sites, including Toulouse assets. Eurofiber said its French colocation activities were not large enough to be structurally competitive on a standalone basis and proposed a strategic partnership with ETIX. That statement validates the central economic problem: even a much larger regional infrastructure group must test whether local assets have sufficient scale to earn their keep.

The proposed transaction could improve facility scale and resilience under ETIX. It could also introduce contract, support or migration changes. No impact on 123 multimedia has been disclosed, and the exact hosting location is not public. The correct conclusion is a due-diligence question, not an outage prediction: does the company have change-of-control protection, tested recovery elsewhere and a clear route-migration plan if supplier ownership or service terms change?

Managed cloud is cheaper to start and harder to compare honestly

The alternative to a dedicated allocation and regional hosting relationship is not simply "put it in the cloud." A real substitute includes compute, storage, databases, load balancing, content delivery, abuse protection, monitoring, backup, recovery, support and staff capable of operating them. It also includes migration risk for a mature, high-message-volume application. The comparison must use total cost and service outcomes, not a server's advertised hourly rate.

Still, the entry price of managed infrastructure is a powerful constraint on 123 multimedia's economics. OVHcloud's French public-cloud tariff offers compute by the hour, multiple French locations, load balancers, storage and included standard denial-of-service protection. A small general-purpose instance is listed at a few euro cents an hour, while larger capacity scales in visible increments. AWS operates a Paris region with three availability zones, offering a different route to multi-site resilience without owning local hardware.

These menus do not prove that a hyperscale deployment would be cheaper. Chat is stateful, latency-sensitive and operationally noisy. Message histories, profiles, media, presence information and moderation queues create storage and database demands. Moving a mature application can require code changes, parallel operation and data transfer. Managed services can replace capital with variable bills that rise quickly with traffic, logs, databases and premium support. A provider can offer three zones while the customer still deploys everything in one.

The cloud alternative nevertheless changes the burden of proof. 123 multimedia does not need to show that its current arrangement is cheaper than every cloud service in every scenario. It needs to show that the arrangement produces a useful advantage: lower steady-state cost at known traffic, faster recovery, better control over sensitive data, stable addressing, more predictable bandwidth, or a white-label proposition that buyers value. If no such benefit is measured, the address allocation and local hosting relationship risk becoming inherited habits.

A hybrid approach may be economically strongest. The company can retain its IPv4 and IPv6 resources, product expertise and data-governance choices while moving commodity functions to managed services. Static media, backups or burst capacity need not share the same architecture as real-time messaging. A secondary recovery environment need not run at full scale every day. The address block can remain useful for stable endpoints and migration even if not every workload sits behind it.

That strategy requires discipline. Hybrid estates create more interfaces, more monitoring and more failure modes. Savings appear only if each component has a clear role and the company removes the cost it replaces. Adding public cloud backup while keeping every existing server, contract and support obligation can improve resilience but not efficiency. Management should compare three fully costed cases: continue the regional setup, modernise it with selective managed services, or migrate the primary estate. Each case needs traffic assumptions, recovery targets, staffing, supplier fees and exit costs.

Buyers and users hold more bargaining power than the company

The company's customer set contains three groups with different leverage. Free users can switch at almost no monetary cost. Paying users can cancel through the mobile platform and retain access until the end of the billing period. Advertisers and their intermediaries can move budgets across applications and formats. White-label customers can compare a renewal with an in-house build, a global community product, a specialist vendor or no community feature at all.

123 multimedia has some defences. Tchatche is an established French brand. An anonymous, open chat experience differs from services built mainly around swiping or existing contacts. A mature moderation operation and a reusable white-label platform are not trivial to reproduce. Existing users, conversation history and local participation create a network effect. A media or telecom customer already integrated with the platform faces migration and audience disruption.

Those strengths do not remove price pressure. The company's own 2015 account said advertising and freemium generated most receipts. Advertising prices are set in a market where a small publisher has limited leverage over demand platforms, consent rules and mobile identifiers. The current privacy policy says data may be shared with advertising partners, advertisers and data brokers subject to the stated legal bases and choices. Any reduction in addressable advertising can lower free-user revenue while the cost of hosting and moderation remains.

The app stores control another gate. Tchatche's terms direct purchases, cancellations and refunds through Apple or Android services. The company can set product tiers, but it does not own the payment relationship end to end. Store rules affect fees, disclosures, renewals and application distribution. A product defect or policy dispute can therefore affect acquisition and revenue even when the Toulouse infrastructure works perfectly.

Large white-label customers can exert the opposite kind of pressure. A recognisable media or telecom brand is valuable to a small supplier, which can make the supplier willing to accept custom work or aggressive pricing. The customer knows that. If a contract requires separate moderation rules, bespoke interfaces and round-the-clock commitments, headline revenue can grow while contribution falls. The company has not published customer concentration, contract length, backlog, renewal rates or the share of work that is standard rather than custom.

The correct concentration measure is therefore not only the largest invoice. It is the largest dependency across revenue, acquisition and infrastructure. One telecom group might be a white-label customer, a distribution channel and a connectivity counterparty. Google Play is both a discovery channel and payment gate. Eurofiber is visible in hosting and routing. An advertising intermediary can influence fill rate across several brands. Diversification should be tested by function, not merely by counting company names.

Global rivals prove the market and raise the price of relevance

The dating market is large enough to support valuable businesses, but size has not made growth easy. Match Group's 2025 results show about USD 3.5 billion of annual revenue, 14.2 million payers and revenue per payer of USD 20.09, while total payers fell 5%. Its portfolio includes Tinder, Hinge, Meetic and other established brands. Match spent at a scale that a small Toulouse company cannot imitate, yet it still had to manage payer decline.

Bumble's 2025 annual report tells a similar story. Revenue fell from about USD 1.07 billion to USD 965.7 million as total paying users declined from 4.15 million to 3.67 million. Average revenue per payer rose, but the user base contracted. The company explicitly identifies restrictive mobile-payment policies, alternative products, product appeal and slower market growth as risks.

These figures are not direct valuation comparables for 123 multimedia. The product portfolios, geographies, brands and accounting bases are far apart. They are useful because they reveal the category's economic variables. Scale brings marketing data, cross-brand learning and product investment, but it does not guarantee user growth. Raising price or revenue per payer can offset volume decline for a time. Trust and safety work may reduce short-term engagement while improving the long-term service.

123 multimedia's realistic strategy is not to outspend these groups. It is to occupy segments where its age, language focus, open-chat format or white-label capability produces cheaper retention. The Google Play listing emphasises French speakers, simple registration, nearby profiles and unlimited free chat. Babel offers an anonymous entry. Cybermen and Amitie target other communities. This segmentation can make a small audience more relevant to its members than a much larger general service.

The downside is fragmentation. Each brand needs enough active people in the same place and time to feel alive. Separate applications split marketing, reviews, updates and moderation. Shared technology can lower the technical cost, but it cannot manufacture local liquidity. A portfolio should be judged by incremental contribution and cross-brand reuse, not by the count of icons in a store.

Free messaging and social products are another substitute. A person seeking conversation can use established networks without downloading a dating application. A media company can direct its audience to an existing social channel instead of operating a branded community. These alternatives may not offer the same anonymity or commercial control, but their zero incremental price to many users is a hard ceiling on monetisation. 123 multimedia has to make discovery, safety and community identity valuable enough to justify both attention and, for a minority, payment.

Moderation and privacy are operating costs, not legal footnotes

The company sells interaction among strangers, sometimes using location and profile information. That makes trust work part of the product. The website advertises round-the-clock multilingual content validation. Tchatche's terms refer to moderation and monitoring teams, user reports and account termination for breaches. The Google Play listing is rated for adults and says the application may collect location, personal information and several other data categories.

The published privacy policy says minors are not authorised, identifies collection of identity, demographic, geolocation, authentication, connection and browsing data, and describes advertising, audience analysis and commercial communications among the purposes. It says some data may go to service providers and advertising partners, sets a maximum period of two years from the user's last initiated interaction for the stated user data, and offers access, deletion, objection and portability rights. The page also says the company participates in the IAB Europe consent framework.

The policy's displayed last-modified date is July 2018. The general terms show March 2021. Old dates do not by themselves prove that actual practices are unlawful or unchanged. They create a diligence requirement because the product, mobile platforms, suppliers and European rules have evolved since publication. A current buyer would want to reconcile the notices with the live application, consent screens, software components, retention settings, vendor list and incident response.

The EU Digital Services Act adds obligations relevant to online platforms, including transparency around advertising and recommendation, notice mechanisms and restrictions on certain profiling. Precise duties and exemptions depend on the service's classification and size. The point for economics is broader: user reporting, reasoned decisions, records, appeals, age controls, advertising choices and privacy safeguards require people and software. They cannot be treated as optional overhead when the service's value depends on safe interaction.

Moderation has a two-sided cost. Too little allows harmful or deceptive content to drive away genuine users and exposes commercial customers to brand risk. Too much, badly executed, can frustrate users, create appeals and remove legitimate engagement. Automation can lower review effort, but edge cases and language context still require judgment. The company claims multilingual, continuous coverage; the public evidence does not disclose review volumes, response times, false-positive rates or cost per active user.

Security has the same shape. Stable address resources and local hosting can help monitoring and incident response, but they do not secure the application by themselves. Account takeover, vulnerable code, supplier portals, payment disputes and misuse of location data sit above the routing layer. The 2019 accounts do not separate security spending, and the company does not publish audit results or recovery tests. A high-availability claim should be paired with application security and data-recovery evidence.

Geopolitical risk is less direct than for a cross-border carrier, but it is not absent. The company serves an online audience through global mobile stores, advertising markets and Internet transit. Rules on data transfer, platform content and consumer subscriptions can diverge. A local French hosting relationship may support buyer preferences for domestic data location, yet the surrounding service chain can still include foreign platforms and advertising counterparties. Sovereignty is a contract-by-contract property, not a postcode.

Public signals show a living product, not a measured franchise

The strongest positive unofficial signal is continued product activity. Tchatche's Google Play page showed a December 2025 update, more than one million cumulative downloads and thousands of reviews. The company website and Babel service remained reachable during review. Recent company responses to store reviews discuss advertising, profile features and support. These observations make it less likely that the brands are merely dormant intellectual property.

The rating is more ambiguous. A score around 3.3 out of five is neither evidence of failure nor a strong endorsement. Individual reviews complain about photographs, location or profile quality; others praise the free service. The sample is self-selected, spans product versions and can reflect rejected profiles or device-specific faults. It is useful for identifying diligence questions, not estimating churn or present service quality.

The public corporate footprint is similarly mixed. Annual accounts continue to be filed, the registered address was updated, the RIPE organisation record was updated in May 2026 and key trademarks such as Tchatche and Babel have been renewed into the 2030s. Those are signs of continuing administration. They do not reveal the size of the audience, revenue or staff commitment behind each brand.

The company's own claims should also be separated by evidentiary weight. Millions of historical downloads, 99.9%-plus service quality, around 100 subcontractors and named major customers are plausible and commercially relevant. They are not accompanied by dates, measurement methods or current contract disclosures. The correct use is to form questions: Which customer relationships are active? How many contractors work in an average month? What is the measured availability over the last year? How many monthly active and paying users remain?

Absence from public routing under its own autonomous-system number is not a negative signal because the company has not claimed to be a transit operator. The visible route through Eurofiber is consistent with outsourced infrastructure. The missing IPv6 announcement is more useful as an execution question. The company has held a /32 since 2013; if it remains unused publicly, management should explain whether migration is unnecessary, blocked by legacy code or customer equipment, or simply not visible in public collectors.

Capital recovery depends on contribution, not on resource scarcity

The case for retaining local network control has four potential returns. First, stable address space can reduce disruption when servers or suppliers change. Second, direct registry control can improve abuse handling, route authorisation and address planning. Third, local hosting can produce predictable latency and hands-on support. Fourth, a controlled platform and infrastructure story may help win white-label buyers concerned about continuity and data location.

Against those returns sit at least six costs. The company pays RIPE membership and upstream infrastructure fees. It must maintain the application and address records. It needs security, monitoring, backup and recovery. It carries moderation and user support. It must fund product updates and customer integrations. It must acquire and retain enough users for advertising and premium revenue to cover all of the above.

The public evidence proves the existence of the assets and activities but not their return. The IPv4 route is active and properly authorised. The consumer application is live. The business has historical profitability and recognised customers. Yet there is no current financial statement, traffic figure, subscriber count, contract mix, gross margin, supplier price, capital budget or availability history. The answer cannot honestly be a clean yes.

It is also too early for a clean no. The annual RIPE fee is small. The company has operated the resource footprint since 2013, which suggests the arrangement has survived several supplier and market changes. The address block can be valuable for a mature service with many allow-lists, reputation histories and integrations. A lean employer using specialist suppliers may keep fixed payroll below what a larger in-house operation requires. Brand longevity and a still-active application indicate that some operating capability has endured.

The judgment is therefore conditional and moderately sceptical. 123 multimedia appears to own useful control points without owning the whole network. That can be economically rational. It becomes value creation only if the company can demonstrate that recurring gross contribution after app-store charges, advertising costs and direct service expense pays for engineering, moderation, security and necessary infrastructure renewal, while producing an acceptable return on the capital and attention tied up in legacy products.

Pricing power is the first test. Evidence of stable or rising revenue per active user without deteriorating retention would support the case. A low-cost advertising-free option alone is unlikely to carry the platform. Higher-value packs, voice features, advertising and white-label fees must combine into a durable mix. If prices rise only because the company narrows free features and users leave, nominal revenue per payer can improve while franchise value falls.

Fixed-cost absorption is the second. The company should disclose cost per million messages, moderation cost per active user, hosting cost per peak concurrent session and contribution by brand. Those measures would show whether traffic scale improves economics. If one platform serves all brands and customers with little incremental engineering, the estate may have strong operating leverage. If each brand and client requires a separate team and stack, visible reach may conceal poor returns.

Supplier resilience is the third. A documented secondary environment, tested recovery time, route migration procedure and clear contract after the proposed data-centre transaction would materially improve the judgment. So would active IPv6 deployment and evidence that the /23 can move without prolonged disruption. Conversely, a single-site arrangement, untested backups or contract dependence on one upstream would make the address ownership look more symbolic than operational.

Customer quality is the fourth. Current multi-year white-label contracts, diversified advertising demand, low consumer churn and a growing cohort of payers would show that users and buyers value the service. Heavy dependence on one old customer, falling active use or continued app maintenance without meaningful conversion would argue for simplifying the estate and buying more managed infrastructure.

Cash conversion is the final test. Current operating cash after software and equipment renewal, not the 2019 net-profit figure, would show whether the footprint earns its keep. A business can report profit while deferring renovation or capitalising development. It can also produce strong cash with modest accounting profit if customers prepay and the platform is mature. Without that bridge, the capital-recovery question remains open.

The facts that would change the judgment

A more positive view would require a small set of hard disclosures. The most important would be current annual recurring revenue split among advertising, subscriptions, white-label contracts and other services; gross margin by line; active and paying users by brand; twelve-month retention; top-five customer share; and operating cash after necessary infrastructure and product spending. Evidence that a common platform supports several brands and clients with limited incremental engineering would confirm reuse rather than duplication.

Infrastructure evidence would also matter. The company should show the number of active addresses, primary and secondary hosting locations, physical path diversity, backup frequency, tested recovery time, major incidents, IPv6 plan and the terms under which the /23 can be moved. A contractual availability record that supports the public quality claim would turn marketing into an investable operating fact.

A negative change would be evidence of declining active users, weak paid conversion, customer losses, rising moderation cost, unresolved security incidents or dependence on a single supplier without tested recovery. Material custom-development obligations on low-margin contracts would weaken the white-label case. A large gap between reported availability and user experience would weaken the value attributed to local infrastructure.

The strategic alternative should remain explicit. If the dedicated footprint does not improve unit cost, continuity or contract wins, management should retain only the resources that support portability and migrate commodity work to managed services. If it does provide measurable value, the company should invest enough to make that control resilient rather than leaving it dependent on one visible route and an unannounced IPv6 allocation.

123 multimedia's history encourages neither nostalgia nor dismissal. The company recovered brands from a failed predecessor, rebuilt a small operating business and preserved a live address footprint for more than a decade. That is evidence of endurance. The next question is harder: whether endurance has become an adequately funded, cash-generative advantage. Until current contribution, resilience and customer data answer it, local network control remains a credible option whose return is still to be proved.