A small business does not discover the quality of its internet line when the speed test looks good. It discovers it at 4.30pm, when the lights go off, the battery in the point-of-sale terminal starts beeping, WhatsApp orders keep arriving, and a customer rings a voice number that has quietly become a data product. In South Africa that moment has not been a rare disaster. It has been part of the operating environment. ICASA's 2025 state-of-the-sector report says telecoms licensees cut battery spending from R2.59bn in 2023 to R173.75m in 2024 as load-shedding eased, while generator spending fell from R930.21m to R211.47m; the point is not only the fall, but the scale of the bill when electricity was unreliable (https://www.icasa.org.za/uploads/files/The-State-of-the-ICT-Sector-Report-of-South-Africa-2025.pdf). For a fixed wireless and fibre provider such as Sonic Telecoms, uptime is not a slogan. It is a cost line, a support promise and a test of whether the network still behaves like a utility when the grid does not.
Sonic's own tariff pages make this unusually concrete. A fibre-to-the-business table sells SME, Business and Enterprise tiers from 10Mbps to 300Mbps, with monthly prices from R3,795 for a 10Mbps SME line to R19,895 for a 300Mbps Enterprise line; the same table separates 4:1, 2:1 and 1:1 contention and moves the priority window from 6am-6pm to 6am-10pm to all day (https://www.sonictelecoms.co.za/fttb-campaign/). The wireless pages make the same promise in another form: SME packages advertise 97% uptime, Business packages 98%, and Enterprise packages 99%, with different contention ratios, public IP allocations and service levels (https://www.sonictelecoms.co.za/packages_sme/, https://www.sonictelecoms.co.za/packages_business/, https://www.sonictelecoms.co.za/packages_enterprise/). A 99% monthly uptime promise still leaves roughly seven hours of possible downtime in a 30-day month. For a household that may be annoying. For a pharmacy, security room, call-heavy estate office or restaurant that has shifted calls and card payments onto IP, seven hours can be the difference between a communications service and an uninsured interruption.
That is the lens through which Sonic Telecoms should be read. It is not merely another local ISP name in a directory. Nor is it, on the public record, a national carrier whose value can be judged by fibre kilometres alone. Sonic is a Cape Town-origin wireless and fibre access business, now bound into the Herotel group, with a public website that still markets its own tower construction, 17GHz microwave backhaul, 5GHz last-mile links, point-to-point fibre, VoIP, private networks and disaster-recovery links (https://www.sonictelecoms.co.za/). The economics sit between local field work and group-scale resilience. Sonic has to install and maintain last-mile equipment close enough to the customer to matter, but it also depends on upstream capacity, undersea-cable diversity, data-centre interconnection, group procurement and the quality of support processes that customers do not see until something fails.
The company therefore belongs to a category that is easy to underestimate. South African broadband competition is often described as a race between fibre network operators, national ISPs and mobile data. Yet many customers buy something more prosaic: a line that can be installed where fibre has not arrived, a wireless link that works because a tower has line of sight, a business service that separates data from voice, and a support team that can answer when Eskom, municipal power, a roof right, a radio alignment or a wholesale handoff breaks the normal rhythm of the day. Sonic's public evidence says it has tried to monetise that middle layer. The question is whether the evidence supports a durable operator story or only a local brand surviving inside a larger group.
The business is local, but the control story is now larger
Sonic's public origin story is a Cape Town wireless one. ITWeb reported in October 2017 that HeroTel acquired Sonic Telecoms, describing Sonic as the largest wireless internet service provider in Cape Town and saying it expected to exceed 5,000 clients that year (https://www.itweb.co.za/article/herotel-acquires-cts-biggest-wisp/XnWJadMbkEdqbjO1). The same report said HeroTel bought 100% of the business, that Sonic was based in Montague Gardens, and that its customer mix included residential, SME and large corporate clients across the CBD, southern suburbs, Atlantic Seaboard, West Coast, northern suburbs, Paarl and Somerset West. Telecompaper's contemporary summary similarly described the acquisition as a 100% HeroTel purchase and said Sonic had a healthy residential, SME and corporate mix (https://www.telecompaper.com/news/herotel-acquires-sonic-telecoms--1216984).
That matters because Sonic's value cannot be judged as if it were still an isolated founder-led WISP. The Herotel context is now part of the product. The Competition Tribunal announced in December 2025 that Vumatel would acquire control of Hero Telecoms, subject to conditions, and described Herotel as operating across national fibre infrastructure, last-mile fibre infrastructure, retail internet access services and fixed wireless access (https://www.comptrib.co.za/info-library/press-room/Merger-Alert%3A-Vumatel-%28Pty%29-Ltd-and-Hero-Telecoms-%28Pty%29-Ltd). Herotel Business, meanwhile, presents wholesale and business infrastructure offerings that include fibre, fixed wireless, voice, co-location, dark fibre, Layer 2 connectivity to Teraco data centres, more than 2,000 towers and 200 fibre points of presence nationwide (https://herotelbusiness.com/affiliates/). For Sonic customers, the brand may still be local; the control surface is increasingly a group surface.
This creates a double edge. Group ownership can bring procurement scale, spares, bandwidth contracts, tower-sharing knowledge, capital discipline and a broader support bench. It can also make accountability harder to read. A Cape Town customer who signs with Sonic, sees Herotel contacts in registry records, uses infrastructure connected through larger backhaul arrangements and relies on a public website that still carries Sonic-specific pages is buying both a local relationship and a group architecture. The evidence points to continuity rather than disappearance. But it also means a buyer or lender cannot analyse Sonic by looking only at the Sonic domain. The relevant question is how Sonic's installed base, network assets, support staff and brand obligations sit inside Herotel's wider network and, after the Vumatel-Herotel transaction, inside a larger South African fibre consolidation story.
The website sells ownership of the difficult edge
Sonic's website is clearest when it talks about the physical edge. The homepage says the company builds its own towers, links them with direct fibre optic connections, uses 17GHz microwave point-to-point links for backhaul and 5GHz microwave links for last-mile connections (https://www.sonictelecoms.co.za/). The about page repeats the same infrastructure language and says Sonic covers the Cape Peninsula, Northern Suburbs, Helderberg, Cape Winelands and West Coast, listing areas from Bellville, Bloubergstrand and Claremont to Paarl, Somerset West and Woodstock (https://www.sonictelecoms.co.za/about/). A WhichVoIP directory page, updated in June 2026, describes Sonic as a Cape Town-based wireless provider founded in 2010 that operates its own 17GHz microwave tower network and offers packages from R499 a month (https://whichvoip.co.za/listing/sonic-telecoms/).
The ownership claim is commercially important, even if it still needs site-by-site verification. A reseller can buy access and resell a bill. A wireless access operator that owns and maintains towers has to manage rooftop agreements, power, equipment refreshes, interference, line-of-sight constraints, storm damage, technician dispatch, backhaul and customer-premise devices. Those duties make the business harder, but they also give it more control over the bottlenecks that customers actually notice. In dense Cape Town suburbs, wireless is not a romantic rural substitute for fibre. It is often a pragmatic way to reach premises that lack usable fibre, need temporary service before fibre lead time clears, require backup diversity, or want a provider that can install without waiting for a larger civil build.
Sonic's service list points to that role. It advertises broadband wireless internet, point-to-point microwave links of up to 2Gbps, private network infrastructure for urban, rural or mining projects across Africa, wireless VPN, point-to-point fibre links of up to 10Gbps, fibre splicing and installation, VoIP and Wi-Fi, offsite disaster-recovery links, network cabling and CCTV backhaul (https://www.sonictelecoms.co.za/). The commercial bundle is not only "internet access". It is a small infrastructure contractor and ISP combination: radio, fibre, cabling, voice, private link, monitoring and support. The attractive margin is in solving awkward local access problems that a national retail tariff does not solve neatly.
There is a caution. The same website has old copyright text on some pages, inconsistent branding dates and contact pages that are not as clean as an institutional buyer would want. That does not negate the operating evidence. It does mean the website should be treated as a service catalogue and lead-generation surface, not as an audited asset register. A serious customer should confirm tower ownership, power arrangements, service boundaries, escalation routes, fibre-provider dependencies and current legal contracting party before relying on the brand for critical service.
Prices reveal how Sonic tries to charge for certainty
The wireless tariff ladder shows Sonic's core pricing logic. In SME wireless, a basic 2Mbps service is listed at R997 per month on a 36-month contract, with 1Mbps upload, 8:1 contention, one public IP, a dedicated VLAN for VoIP, Silver SLA and 97% uptime; the 10Mbps SME tier rises to R1,797 per month on a 36-month contract (https://www.sonictelecoms.co.za/packages_sme/). Business wireless changes the shape: the 10Mbps tier shows 5Mbps upload, 4:1 contention, five public IPs, Gold SLA, 98% uptime and R2,696 per month on a 36-month term (https://www.sonictelecoms.co.za/packages_business/). Enterprise wireless is a different product again: 10Mbps symmetric, 1:1 contention, 16 public IPs, Platinum SLA, 99% uptime and R6,997 per month on a 36-month term (https://www.sonictelecoms.co.za/packages_enterprise/).
That ladder is a compact lesson in access economics. The customer does not pay linearly for megabits. The customer pays for less sharing, more upload, more public addressing, a better support promise and the implied right to complain with greater urgency. The same 10Mbps headline can therefore be a consumer-like SME service, a business access service or an enterprise line. Sonic's page puts that fact in numbers. Contention falls from 8:1 to 4:1 to 1:1. Uptime moves from 97% to 98% to 99%. Priority hours expand. Installation fees change with term length. The product is not only bandwidth; it is the allocation of operational attention.
Fibre-to-the-business pricing repeats the same segmentation. Sonic's FTTB campaign lists 10Mbps at R3,795 for SME, R3,995 for Business and R5,195 for Enterprise, while 300Mbps moves from R14,595 to R15,295 to R19,895 across the same three classes (https://www.sonictelecoms.co.za/fttb-campaign/). The Century City Connect page offers a starker comparison: 100Mbps on a contended 5:1 line is R12,595, while 100Mbps on a 1:1 line is R38,395; both are symmetrical and designed for businesses operating 24/7, but the price difference is the cost of not sharing capacity in the same way (https://www.sonictelecoms.co.za/fttb-uncapped-century-city-connect/).
For Sonic, this is the correct way to expose value, but it also exposes risk. If customers buy only on download speed, Sonic will be compared with larger fibre ISPs and mobile fixed-wireless offers that can appear cheaper. If customers understand that a 1:1 service with public IPs, priority support, voice separation and a local installer solves a different problem, Sonic has a more defensible niche. The entire business therefore depends on educating buyers without overpromising. "Fast internet" is a crowded phrase. "The line that keeps calls, card machines, cameras and staff reachable when conditions are bad" is a more valuable product, but it is harder to prove.
Voice turns a data outage into lost trade
Voice is the small line in the package table that changes the economics. Sonic's SME packages include a dedicated VLAN for VoIP and show a 256Kbps VoIP LAN option as free, with higher VoIP LAN speeds priced at R600, R1,000 and R1,500 per month for 512Kbps, 1Mbps and 2Mbps respectively (https://www.sonictelecoms.co.za/packages_sme/). Business and Enterprise wireless packages split VoIP and data into VLANs, with higher service levels and more public IPs (https://www.sonictelecoms.co.za/packages_business/, https://www.sonictelecoms.co.za/packages_enterprise/). The website also advertises VoIP and Wi-Fi solutions, wireless VPN and private infrastructure (https://www.sonictelecoms.co.za/).
This matters because voice has moved from a standalone telephone service to an application that rides on the same access path as card payments, bookings, remote support and security cameras. If a Cape Town dentist's reception desk loses IP voice during an outage, the loss is not simply the price of a call. It is missed appointments, staff time, patient frustration and reputational damage. If an estate security office loses a CCTV backhaul and voice path at the same time, the failure is operational. If a restaurant can still cook on gas but cannot receive delivery calls or process online orders, the bottleneck is communications rather than food.
The economics of VoIP over wireless or fibre are therefore closer to insurance than to commodity bandwidth. Separating voice from data in VLANs is a technical way to protect call quality, but it does not solve power by itself. The radio, customer router, PoE injector, fibre ONT, switch, handset or PBX, and upstream network all need power or backup. Sonic's uptime promise must be read through the entire chain. A customer with backup power at the premises but a powerless rooftop device still loses service. A powered customer device with an upstream tower on exhausted batteries also fails. A powered wireless link with an upstream fibre break or congested failover path degrades. Sonic can price the promise only if it has the field discipline to know which part of the chain is likely to fail.
That is why load-shedding changed the product. South Africa's crisis forced telecoms providers to turn batteries, generators, security, spares and technician routing into parts of the service. ITWeb reported from ICASA's 2024 sector figures that telecoms providers spent R2.5bn on batteries and R930m on generators in 2023 and bought 150,415 batteries and 3,268 generators to counter load-shedding (https://www.itweb.co.za/article/icasa-counts-load-shedding-costs-for-telcos/8OKdWMDXyl9MbznQ). The numbers are sector-wide, not Sonic-specific. But they show the cost pool from which Sonic's uptime promises must ultimately be funded.
A concrete failure scenario
Imagine a 30-seat medical practice in Bellville using Sonic for a business wireless link, VoIP phones, card terminals and cloud appointment software. The practice has a small inverter for laptops and the router, because the owner learned during 2023 that the waiting room cannot go dark every afternoon. At 3pm, load-shedding hits the area. The router stays on. The phones still have power. For ten minutes everything works. Then call quality drops, reception hears only every second word, and a queue of patients starts forming because the card terminal cannot settle cleanly. The receptionist phones support from a mobile. Sonic's support desk has to decide whether the problem is the customer's inverter, the customer radio, the tower battery, congestion on a failover route, a voice VLAN issue, a fibre handoff fault or a wider Herotel upstream event.
This is where the business is won or lost. The customer will not care whether the root cause sits in a microwave hop, an IP address allocation, a private VLAN, an undersea-cable failover, a Teraco interconnect, a Herotel aggregation link or a device on the practice roof. It bought one relationship. If the provider can see the link, explain the failure, route voice traffic, dispatch a technician if needed and restore service before the afternoon's calls are lost, the premium is justified. If the provider can only recite an uptime percentage and ask the customer to reboot, the premium collapses into resentment.
The failure is also financially asymmetric. The monthly difference between an SME and Business or Enterprise plan looks large in the tariff table. But one afternoon of failed calls can cost more than the difference if the business is appointment-driven or transaction-heavy. Conversely, if the customer is a household streaming video after work, paying for business-grade support may be irrational. Sonic's challenge is segmentation: sell the expensive service to customers whose downtime is expensive, and do not let low-margin consumer expectations consume high-cost support resources.
This is why the "power and uptime" lens is more useful than a simple company-history lens. Sonic's strategic problem is not whether it can tell a nice story about fast wireless. It is whether it can convert operational resilience into priced tiers, prove that resilience during failures, and keep support costs from eating the premium.
The network record is real, but it points to Herotel
Peering and registry evidence supports a real network history, but also shows why Sonic should be read in Herotel context. PeeringDB lists Sonic Telecoms as AS37417, also known as Fusion Wireless (PTY) Ltd, with website https://www.sonictelecoms.co.za, IRR set AS-SONICWIRELESS, network type Cable/DSL/ISP, 30 IPv4 prefixes, one IPv6 prefix, 10-20Gbps traffic, mostly inbound ratio and African geographic scope (https://www.peeringdb.com/net/7469). The PeeringDB API record also shows no listed public exchange connections and no listed facilities, with the same traffic and prefix data (https://www.peeringdb.com/api/net?asn=37417).
AFRINIC RDAP is more revealing about current control. The AS37417 RDAP record has handle AS37417, status active, registration event dated February 14 2022, last changed April 2 2026, and registrant HERO TELECOMS (PTY) LTD, with Herotel contacts and addresses in Stellenbosch (https://rdap.afrinic.net/rdap/autnum/37417). A sampled RDAP IP record for 154.66.248.0/24 shows HERO TELECOMS as registrant, Herotel abuse and net-ops contacts, active status, country ZA and a last changed date of June 30 2026 (https://rdap.afrinic.net/rdap/ip/154.66.248.0). That does not erase the Sonic brand. It says the address and autonomous-system evidence now belongs to a Herotel-controlled network environment.
Routing visibility adds another caveat. Public routing sites have recently shown mixed signals around AS37417. Search-visible BGP tools identify AS37417 as HERO TELECOMS (PTY) LTD and connect it to the Sonic website, while Hurricane Electric search snippets have indicated that AS37417 has not been visible in the global routing table since April 5 2024 (https://bgp.he.net/AS37417). That should not be overstated; routing visibility can change, route objects can be historical, and operational traffic may sit under other ASNs in a group. But it matters for diligence. If a customer or buyer cares about Sonic's live autonomous-network independence, PeeringDB alone is not enough. They should ask which ASN currently originates customer traffic, which prefixes are used, which upstreams carry the service, what route diversity exists and how failover is tested.
The correct public judgement is therefore balanced. Sonic has stronger evidence than a bare domain: a public service catalogue, package prices, address, acquisition history, PeeringDB record and AFRINIC records. But the network-control evidence increasingly points to Herotel. Sonic's economics are best understood as a local access and support brand within a larger Herotel infrastructure and procurement system, not as a standalone backbone.
Accountability sits in the terms, not only in the mast
The legal and regulatory boundary is less visible than the radio equipment, but it affects what a customer is really buying. Sonic's public pages link to terms and conditions, debit-order terms, complaints procedures and a code of conduct, and several package pages make a debit order mandatory for monthly service fees while stating that pricing excludes VAT and that equipment remains Sonic's property on cancellation (https://www.sonictelecoms.co.za/; https://www.sonictelecoms.co.za/commercial-wireless-business/). These small clauses are commercially meaningful. They tell the customer that the provider is not merely shipping a router and walking away. It retains equipment control, expects recurring payment certainty and defines the complaint route through published service terms.
Historical licence evidence also points to a regulated communications-service role, though it should be handled with care. A 2022 class-license list mirrored by Ellipsis from ICASA materials shows Fusion Wireless (Pty) Ltd, associated with James Wilkinson and a Milnerton address, with C-ECNS and C-ECS entries (https://www.ellipsis.co.za/wp-content/uploads/2023/07/List-of-Class-Licensees-2022-Updated-13-April-2022-rev1.pdf). That is not a substitute for checking the current live licence position, especially after Herotel integration and later Vumatel control approval. But it supports the view that Sonic/Fusion Wireless has been part of the formal South African electronic-communications licensing universe rather than only a web brand.
For buyers, this changes the diligence task. The first question is not whether Sonic can legally sell internet in an abstract sense. The more useful question is which licensed entity signs the contract today, which terms govern a Sonic-branded customer, how Herotel group obligations flow through to the customer, and whether advertised SLA categories are enforceable service commitments or marketing shorthand. A 99% uptime promise is worth more when the contract defines measurement, exclusions, remedy and escalation. It is worth less if the customer can only find a tariff table and a generic support number.
The same applies to debit-order and equipment terms. Mandatory debit orders reduce collection friction and protect cash flow, but they can irritate customers if service quality is poor or cancellation is messy. Provider-owned customer equipment gives Sonic more control over the installed base and helps recover radios or routers after churn, but it also creates inventory, maintenance and retrieval obligations. A lender looking at Sonic's cash flow would ask how much capital is tied up in customer-premise equipment, how often equipment is recovered and reused, and whether installation fees cover enough of the upfront cost. An acquirer would ask whether contracts, equipment records and service addresses match cleanly inside Herotel systems. In access networks, paperwork and inventory can be as important as signal strength.
Backup is more than undersea cables
Sonic's homepage says it has international failover capacity with multiple undersea cables and other local internet providers at Teraco to offer speed, scale and quick connection (https://www.sonictelecoms.co.za/). The FTTB campaign page names SEACOM, WACS and SAT3 as undersea cable systems used for SME, Business and Enterprise packages (https://www.sonictelecoms.co.za/fttb-campaign/). Those are meaningful claims because cable diversity and data-centre interconnection matter in South Africa. International capacity, local peering and Teraco presence can reduce the risk that one wholesale path becomes the bottleneck.
But cable diversity is not the same as service resilience at the customer edge. A business in Montague Gardens can have access to traffic that ultimately rides multiple submarine systems and still lose service if a rooftop radio has no power, a mast battery is stolen, a local fibre handoff is delayed, a wireless line of sight is blocked by construction, or a technician queue is overwhelmed after a storm. The first mile and last mile are often the fragile miles. Sonic's website is sensible because it combines "bold backup" with local tower and microwave language. The danger would be selling international failover as if it solved local power.
South Africa's power data makes this distinction unavoidable. CSIR's 2024 utility statistics report says actual load-shedding decreased by about 76% in 2024 compared with 2023 and that there were 281 consecutive days without load-shedding by December 31 2024, but it also records 6,948 load-shed hours in 2023 and 1,656 in 2024 (https://www.csir.co.za/sites/default/files/2025-09/Utility%20Statistics%20Report_Jan%202025_Final.pdf). OECD's 2025 survey says load-shedding eased to only 69 days in 2024, after a severe 2023, but warns that the system remains fragile and that early 2025 brought a return of load-shedding (https://www.oecd.org/en/publications/oecd-economic-surveys-south-africa-2025_7e6a132a-en/full-report/reforming-south-africa-s-electricity-sector_05fdccb6.html). Eskom said in November 2024 that load-shedding had been suspended for 226 consecutive days since March 26 2024 and that diesel savings reached R14.6bn year on year (https://www.eskom.co.za/loadshedding-remains-suspended-as-investments-in-the-generation-recovery-plan-continue-to-pay-off-driving-efficiencies-and-supporting-economic-growth-diesel-savings-reached-r14-6-billion-year-on-ye/).
For Sonic, the improvement lowers emergency cost but does not remove the buyer memory. Customers who bought inverters, changed routers, complained about towers and learned which providers failed during cuts will not forget quickly. The commercial question becomes: can Sonic keep enough backup capability and support readiness after the crisis eases, without carrying the full 2023 cost base into a more competitive 2026 market?
Competition is not only price; it is who pays for standby capacity
ICASA's 2025 report shows why Sonic operates in a market with demand but limited room for mistakes. Fixed internet access at home was highest in the Western Cape at 40.1% in the 2023 household data, far above the national average of 14.5%, while fixed internet and data revenue grew 14.62% in 2024 to R34.97bn (https://www.icasa.org.za/uploads/files/The-State-of-the-ICT-Sector-Report-of-South-Africa-2025.pdf). The same report says fixed broadband subscriptions jumped from 1.4m to 2.7m, driven mainly by fibre, and that fibre-to-the-home and fibre-to-the-building subscriptions rose from 1.0m to 2.4m. That is a demand tailwind for fibre and business connectivity. It is also a competitive threat to fixed wireless in urban Cape Town.
The threat is simple. If fibre reaches a street with a cheap, stable, well-supported service, wireless loses some of its urgency. Sonic's own 2017 acquisition story anticipated this. ITWeb quoted James Wilkinson saying HeroTel would bring new skills, capital, national partnerships for fibre to the home and business, and LTE-A packages for clients that could not access high sites (https://www.itweb.co.za/article/herotel-acquires-cts-biggest-wisp/XnWJadMbkEdqbjO1). That was the right strategic direction. A WISP that refuses fibre becomes trapped as fibre expands. A WISP that adds fibre, LTE/5G alternatives, voice, private links and managed support can remain useful even as the access medium changes.
The cost problem is that standby capacity has to be paid for by somebody. Business customers want 98% or 99% uptime, all-day priority, uncapped service, low contention, public IPs and voice quality. They also compare prices with consumer fibre ads. Sonic's tariff ladder is an attempt to make the trade-off explicit: if a customer wants 1:1 contention and all-day priority, the monthly fee rises. If the customer accepts 4:1 or 8:1 contention and narrower support hours, the monthly fee falls. The risk is adverse selection. Customers with high downtime costs may underbuy and then demand premium support during failures. Customers with low downtime costs may overbuy and churn when they realise cheaper fibre is enough. Sonic's sales discipline therefore matters as much as its radio engineering.
Herotel's broader consumer positioning makes the contest sharper. Its public site emphasises local teams, no long-term contracts, affordable internet, daily support and community presence (https://herotel.com/). That is a mass-market trust proposition. Sonic's older pages, by contrast, expose a more technical business-access proposition. The best outcome is for Herotel to use Sonic's Cape Town technical edge while giving customers a cleaner group support and billing experience. The weaker outcome is brand confusion: Sonic pages suggest one promise, Herotel systems another, and the customer is left unsure who owns the fault.
Labour and support are the hidden cost of local trust
Local support is expensive because it cannot be fully automated. Sonic's homepage and about page both say it provides online and in-person support (https://www.sonictelecoms.co.za/, https://www.sonictelecoms.co.za/about/). Herotel's consumer site leans heavily on local support teams and seven-day support (https://herotel.com/). Herotel Business says its fibre and fixed wireless products offer 24/7 support, performance reports and a single provider experience (https://herotelbusiness.com/affiliates/). These are not decorative claims. In fixed wireless and business fibre, labour is part of the product.
The support load rises during stress. A load-shedding schedule, a storm, a mast outage, a fibre cut, a stolen battery or an upstream routing event does not create one ticket. It creates dozens or hundreds, many of which look different from the customer side. The provider has to triage. Is this a known area fault? Is the customer-power device down? Is it a line-of-sight issue? Is it a fibre-provider lead-time problem? Is voice degraded because data traffic is saturating the link? Is a backup route carrying traffic but with higher latency? Sonic's promise of "fast and effective" support is worth money only if the team can answer those questions quickly.
The public record also includes a labour-risk signal. Search-visible summaries of the Moses and Herotel/Fusion Wireless litigation describe Fusion Wireless, trading as Sonic Telecoms, as a Herotel subsidiary involved in 2020 retrenchment litigation, with financial difficulty, churn and competition from fibre cited in public summaries of the dispute (https://www.saflii.org/za/cases/ZALCCT/2024/5.html; https://www.saflii.org/za/cases/ZALAC/2025/42.html). The accessible search snippets should be treated carefully, not as a full operating audit. But the signal is consistent with the sector economics: fixed wireless operators faced fibre competition, load-shedding operating costs and consolidation pressure. Labour planning is not peripheral; it affects service quality and continuity.
For an acquirer, this is where diligence should go beyond tariffs. How many field technicians cover Sonic's Cape Town footprint? How many sites require battery maintenance? What is the truck-roll rate per active customer? Which faults are resolved remotely? What is average restoration time during area power cuts? How many high-value business customers depend on voice VLANs? What customer-premise equipment remains owned by Sonic and collected at cancellation? Sonic's public pages say equipment remains Sonic's property on some commercial wireless terms (https://www.sonictelecoms.co.za/commercial-wireless-business/). That creates asset control, but also recovery, refurbishment and support obligations.
Public signals are useful, but thin
Unofficial market signals around Sonic are modest. WhichVoIP's 2026 listing is generally positive about Sonic's Cape Town tenure, owned 17GHz network and starting price, but it also tells buyers to confirm coverage, site-survey results, SLA terms and whether business packages fit their address (https://whichvoip.co.za/listing/sonic-telecoms/). HelloPeter's Sonic Telecoms page showed one review in the last 12 months and a low TrustIndex snapshot when reviewed, which is too small a sample to prove service quality but sufficient to say the public complaint surface is thin rather than richly reassuring (https://www.hellopeter.com/sonic-telecoms). LinkedIn's public company page identifies "Herotel Sonic" in Cape Town with more than 2,000 followers, again confirming brand continuity more than operational quality (https://za.linkedin.com/company/sonic-telecoms).
Thin signals cut both ways. A bad review sample of one is not a reliable indictment. A polished website is not a service audit. A directory listing may be commercially generated. A LinkedIn page can lag organisational reality. For Sonic, the best evidence remains the combination of its own tariff pages, network registry records, acquisition history and sector data. They show a real service proposition and a plausible economic mechanism. They do not reveal current subscriber count, churn, tower-level uptime, gross margin, backup-power stock, customer concentration, exact Herotel integration state or current route table.
That uncertainty should shape the judgement. Sonic is investable as a local reliability story only if the operating metrics support the promise. The brand's public claims are specific enough to test. A buyer can ask for active customer count by product tier, ARPU by wireless versus fibre, churn by suburb, outage minutes by cause, battery autonomy by high site, fault-resolution time, voice-ticket incidence, business customer concentration and the share of service delivered over owned towers versus aggregated fibre. A lender can ask whether cash flow is resilient when electricity is stable and when it is not. A regulator can ask whether complaints, licence obligations, service terms and customer communications match the advertised SLA tiers.
The one public fact that would most change the judgement is not another slogan about speed. It is verified current uptime by product tier, split between customer-premise fault, Sonic access fault, upstream fault and power-related fault, over at least 12 months. If Sonic can show that its 98% and 99% tiers actually behave differently during stress, the premium is credible. If the tiers are mostly packaging with similar failure experience, the business is more exposed to price competition.
What a buyer or lender would pay for
A buyer, lender, large customer or regulator would not pay much for the Sonic name alone. They would pay for active customers with low churn, tower and rooftop rights that are transferable, clean customer contracts, working field teams, a documented power-backup plan, public-IP and voice customers who cannot easily switch, and Herotel/Vumatel group access that lowers bandwidth, equipment and resilience costs. They would discount stale web pages, unclear legal contracting identity, uncertain live ASN independence, weak public customer evidence, any mismatch between advertised uptime and measured restoration, and any reliance on high sites without adequate power or security. They would refuse to underwrite a critical-service promise without proof of battery autonomy, upstream diversity, escalation ownership and ticket-resolution data. The evidence supports a cautious but real operating thesis: Sonic's value is in local uptime execution inside a larger group, not in being a standalone Cape Town backbone.
Public evidence register
The Sonic website supports the service-catalogue claims: own towers, 17GHz microwave backhaul, 5GHz last-mile links, private networks, VoIP, disaster-recovery links and Teraco/undersea-cable failover language (https://www.sonictelecoms.co.za/). The Sonic about page supports the Cape Peninsula, Northern Suburbs, Helderberg, Cape Winelands and West Coast coverage claims and the Montague Gardens address (https://www.sonictelecoms.co.za/about/). Sonic's SME, Business and Enterprise wireless package pages support the 97%, 98% and 99% uptime tiers, contention ratios, public IP allocations, VoIP VLAN treatment and monthly pricing examples (https://www.sonictelecoms.co.za/packages_sme/, https://www.sonictelecoms.co.za/packages_business/, https://www.sonictelecoms.co.za/packages_enterprise/). The FTTB campaign and Century City Connect pages support the fibre pricing, cable-system, SLA-window and 1:1 versus contended price analysis (https://www.sonictelecoms.co.za/fttb-campaign/, https://www.sonictelecoms.co.za/fttb-uncapped-century-city-connect/).
PeeringDB supports the AS37417, Fusion Wireless alias, traffic, prefix and peering-policy evidence (https://www.peeringdb.com/net/7469; https://www.peeringdb.com/api/net?asn=37417). AFRINIC RDAP supports the Herotel registrant and contact evidence for AS37417 and sampled IP space (https://rdap.afrinic.net/rdap/autnum/37417; https://rdap.afrinic.net/rdap/ip/154.66.248.0). ITWeb and Telecompaper support the 2017 Herotel acquisition record and the reported client and coverage context (https://www.itweb.co.za/article/herotel-acquires-cts-biggest-wisp/XnWJadMbkEdqbjO1; https://www.telecompaper.com/news/herotel-acquires-sonic-telecoms--1216984). The Competition Tribunal and Herotel Business pages support the later Vumatel-Herotel control context and the wider Herotel infrastructure surface (https://www.comptrib.co.za/info-library/press-room/Merger-Alert%3A-Vumatel-%28Pty%29-Ltd-and-Hero-Telecoms-%28Pty%29-Ltd; https://herotelbusiness.com/affiliates/).
ICASA, ITWeb, CSIR, OECD and Eskom support the South African load-shedding and telecom-cost context: sector battery and generator spending, fixed internet revenue growth, broadband subscription growth, load-shedding days and hours, and the 2024 suspension of scheduled power cuts (https://www.icasa.org.za/uploads/files/The-State-of-the-ICT-Sector-Report-of-South-Africa-2025.pdf; https://www.itweb.co.za/article/icasa-counts-load-shedding-costs-for-telcos/8OKdWMDXyl9MbznQ; https://www.csir.co.za/sites/default/files/2025-09/Utility%20Statistics%20Report_Jan%202025_Final.pdf; https://www.oecd.org/en/publications/oecd-economic-surveys-south-africa-2025_7e6a132a-en/full-report/reforming-south-africa-s-electricity-sector_05fdccb6.html; https://www.eskom.co.za/loadshedding-remains-suspended-as-investments-in-the-generation-recovery-plan-continue-to-pay-off-driving-efficiencies-and-supporting-economic-growth-diesel-savings-reached-r14-6-billion-year-on-ye/). WhichVoIP, HelloPeter and LinkedIn are used only as market-signal context, not as proof of operating quality (https://whichvoip.co.za/listing/sonic-telecoms/; https://www.hellopeter.com/sonic-telecoms; https://za.linkedin.com/company/sonic-telecoms).
The judgement
Sonic Telecoms is most interesting where South Africa's electricity problem meets a very local access business. The company does not need to be a national champion to matter. It needs to keep a Cape Town customer reachable when power, fibre lead times, roof rights, radio paths and voice quality collide. Its tariff pages show a rational attempt to charge for certainty. Its public network records show real history but a Herotel-controlled present. Its acquisition history explains why scale may now come from group ownership rather than standalone expansion. Its biggest commercial opportunity is to turn local support and backup discipline into a paid premium for customers whose downtime is expensive.
The risk is that the premium becomes hard to defend as fibre spreads, mobile fixed-wireless improves, and customers forget the worst of load-shedding. When the grid is stable, a 99% promise can look like an unnecessary surcharge. When the grid fails, it can look cheap. Sonic's next phase will depend on whether it can keep proving, not merely advertising, that difference. In a market where the lights have taught customers to understand uptime, the winning operator is not the one with the loudest speed claim. It is the one whose support desk, tower batteries, fibre handoffs and voice paths still make sense when the afternoon goes dark.