Summary

A merchant's monthly bill is the right unit of analysis

Imagine a small cosmetics seller in Belo Horizonte who has outgrown Instagram direct messages. She needs a domain, email, a storefront, payment links, Pix, card acceptance, shipping labels, a support number and a place for a WordPress site or a lightweight server. Her measurable unit is not "digital transformation." It is a bill that starts at a few dozen reais and can become several hundred reais a month once a site, store, checkout, shipping, email, monitoring and support are added. Locaweb advertises hosting plans from R$5.90 per month with domain, email, SSL, daily backup, migration and 24-hour support (https://www.locaweb.com.br/hospedagem-de-sites-com-dominio-gratis/). Its own educational commerce page describes a Loja Virtual offer at R$29.90 per month on an annual plan (https://www.locaweb.com.br/conteudos/plataforma-de-vendas-online/). Vindi, the payments brand inside the group, shows a card rate from 2.75% for 30-day receipt, a 14-day card rate from 3.25%, Pix at 0.95% with a minimum of R$1.60, and boleto at R$2.99 (https://vindi.com.br/precos/).

The substitute is just as concrete. Hostinger's Brazil price page advertises hosting from R$5.99 per month and a separate VPS page shows a KVM plan from R$29.99 per month with renewal at R$59.99 (https://www.hostinger.com/br/precos and https://www.hostinger.com/br/servidor-vps). HostGator Brazil advertises site hosting from R$10.09 per month, servers in Brazil, free migration, email and 24-hour support (https://www.hostgator.com.br/hospedagem-de-sites). AWS Lightsail offers global cloud primitives in dollars, with a Linux bundle at $5 per month for 1 GB memory, 2 vCPUs, 40 GB SSD and 2 TB transfer, while its pricing page also shows smaller and larger bundles in USD (https://aws.amazon.com/lightsail/pricing/). The merchant does not compare these products in a spreadsheet the way an analyst would. She asks which provider will help her sell this week, recover a password at night, keep email working, accept Pix, print a shipping label and avoid a surprise bill.

That is the lens for Locaweb. The company can lose the hosting comparison and still win the account if its bundle lowers the buyer's total friction. It can win the hosting comparison and still lose margin if support, refunds, cancellations, cloud inputs or payment financing eat the spread. Brazilian hosting margin after software, payments and cloud stopped being separate is therefore a question about the joints between products. The first sale is cheap. The economic sale is the second, third and fourth product, kept alive by support and made profitable by operational leverage.

Locaweb's public identity also shows the same shift. The company presents its purpose as helping businesses be born and prosper through technology, with an integrated portfolio for different stages of the business journey (https://ri.lwsa.tech/en/). Its official customer-facing billing FAQ says the current corporate name is LWSA S/A, the trade name is Locaweb, and the CNPJ is 02.351877/0001-52 (https://www.locaweb.com.br/ajuda/wiki/nova-razao-social-lwsa-sa/). The directory article should therefore treat "Locaweb" as the public brand and the merchant relationship, while the filings and market data come from LWSA S/A. That distinction matters because the brand sells trust at the front door and the listed company reports the economics behind it.

Commerce now carries the growth that hosting once promised

The latest numbers make the company much more commerce-led than a casual hosting reference suggests. In the 1Q26 earnings release, LWSA reported consolidated net revenue of R$362.8 million, up 10.0% year over year on a comparable basis. Commerce net revenue was R$262.1 million, up 14.3%, while BeOnline/SaaS net revenue was R$100.7 million, up only 0.1%. Platform subscription net revenue inside Commerce was R$145.6 million, up 18.9%, and ecosystem net revenue was R$116.5 million, up 9.1% (https://api.mziq.com/mzfilemanager/v2/d/a8a11432-9651-4cc3-9064-ff466658c119/9521a0df-bf57-f304-c2f1-7690dd212003?origin=2). A rough mix calculation puts Commerce at about 72% of 1Q26 consolidated revenue.

That is not a rounding-point change. LWSA's 4Q25 financial statements show the same full-year pattern: 2025 consolidated net revenue reached R$1.4885 billion, up 10.3% on a comparable basis; Commerce net revenue was R$1.0732 billion, up 15.3%; and BeOnline/SaaS net revenue was R$415.4 million, down 0.8% (https://api.mziq.com/mzfilemanager/v2/d/a8a11432-9651-4cc3-9064-ff466658c119/402e6a52-0b7b-db4d-268e-3233ed509720?origin=2). The full-year mix is again roughly 72% Commerce. Hosting and BeOnline remain relevant, but the growth proof has moved toward commerce subscription revenue, payments, logistics, marketplace integration and enterprise e-commerce.

The company says it serves more than 700 thousand clients across online presence, e-commerce, management and financial services (https://api.mziq.com/mzfilemanager/v2/d/a8a11432-9651-4cc3-9064-ff466658c119/9521a0df-bf57-f304-c2f1-7690dd212003?origin=2). The operating indicators sharpen the point: platform subscribers reached 211.0 thousand in 1Q26, up 7.7%; ecosystem GMV reached R$20.29 billion, up 11.5%; TPV in payments reached R$2.2275 billion, up 10.1%; and own-store GMV was R$1.6168 billion, up 7.0%. BeOnline/SaaS clients at period end were 381.0 thousand, down 2.4% from the prior-year period. The hosting-era base is still large, but the expansion logic sits in converting merchants into platform subscribers and then monetizing their activity.

Adjusted EBITDA also points to the same internal migration. Consolidated adjusted EBITDA was R$91.0 million in 1Q26, up 28.4%, with margin at 25.1%. Commerce adjusted EBITDA was R$64.9 million, with a 24.8% margin, while BeOnline/SaaS adjusted EBITDA was R$26.0 million, down 7.5%, with margin slipping to 25.8% (https://api.mziq.com/mzfilemanager/v2/d/a8a11432-9651-4cc3-9064-ff466658c119/9521a0df-bf57-f304-c2f1-7690dd212003?origin=2). The absolute BeOnline margin is still respectable, but the direction matters. Locaweb's old hosting surface can be profitable yet not enough to carry the equity story.

Management's explanation is explicit. Platform subscription growth was driven by subscriber additions and ARPU expansion; the release says merchants using more ecosystem tools and selling more naturally migrate to plans with greater GMV capacity, data and functionality. It also says payments and logistics penetration helped sustain ecosystem revenue. That is a software-and-transaction model, not a simple rent-a-server model. The margin therefore depends on whether customer success and product attachment raise revenue per merchant faster than support, cloud inputs, marketing, payment financing and churn consume the gross profit.

Hosting remains the trust layer even when it is no longer the whole story

The reason Locaweb still matters in a cloud-service directory is that hosting is not dead. It has become the trust layer underneath the cross-sell. The company's own history page begins with Locaweb's 1998 founding as a straightforward website solution, says it reached 1,000 clients quickly, reached 10,000 clients by the end of 2000, began offering data center and email services in 2003, built its first own data center in 2006 with capacity for 4,000 servers, launched cloud computing in 2008, and launched a second data center in 2010 with capacity for 25,000 servers (https://ri.lwsa.tech/en/company/history/). That history is economically useful because it explains why Brazilian buyers still remember Locaweb as infrastructure, not only as commerce software.

The ecosystem page frames the current Locaweb brand as digital-presence solutions for entrepreneurs and developers, with SSL, hosting, email marketing, support for developers and agencies, an easy panel, unlimited email accounts and SSD databases (https://ri.lwsa.tech/a-companhia/nosso-ecossistema/). KingHost is described on the same page as high-performance hosting with 100% Brazilian servers and technical support. The business-units page lists Bagy, Bling, Cplug, KingHost, Locaweb, Melhor Envio, Octadesk, Tray, Vindi and Wake (https://ri.lwsa.tech/en/company/our-business-units/). That list is not just brand inventory. It is the map of the merchant journey LWSA is trying to own.

Hosting gives that journey a low-friction starting point. A merchant can enter through a domain, email or WordPress site and later be offered a virtual store, payment hub, shipping integration, customer-service automation or ERP. The company's public hosting page is designed for that entry point: "No Brasil, em reais" is more than marketing copy because it directly addresses the buyer's fear of dollar-denominated infrastructure bills (https://www.locaweb.com.br/hospedagem-de-sites-com-dominio-gratis/). The same page advertises 30 days to ask for a refund and free migration. Those promises are expensive if the customer base is difficult to support; they are valuable if they reduce buyer hesitation and raise conversion.

The billing FAQ adds another margin detail. It says monthly service values can be adjusted annually by IPCA, Brazil's consumer price index (https://www.locaweb.com.br/ajuda/wiki/duvidas-frequentes-cobranca/). This gives Locaweb some inflation pass-through on real-denominated plans. It does not fully solve FX exposure, imported hardware, third-party cloud bills or competitive discounting, but it gives the company a local pricing mechanism that global cloud primitives do not automatically offer to a small Brazilian buyer.

The trust layer is also where product simplicity matters. A business owner who buys a cheap global VPS may get a low sticker price but still need DNS configuration, SSL, email deliverability, backups, security patches, monitoring, checkout integration and Portuguese support. Locaweb's site makes the bundle look familiar: domain, hosting, WordPress, site builder, email, SSL, daily backup and support. The buyer pays for the comfort of not assembling every piece. That comfort is the margin opportunity. It is also the support burden.

Support is not a side cost; it is part of the product sold

Locaweb's support promise is central to the economics because its typical buyer is not a hyperscale engineer. The company's support page says customers can use WhatsApp, chat and phone, lists Sao Paulo and other-region numbers, and states that technical support is available 24 hours a day, seven days a week, while billing support runs during business hours (https://www.locaweb.com.br/atendimento/). A help article about faster technical support tells customers to collect evidence from tests and screenshots and reiterates that the support team is available 24/7 (https://www.locaweb.com.br/ajuda/wiki/coletando-informacoes-para-um-atendimento-mais-agil-com-o-suporte-tecnico/). Another page describes the status blog as the place customers can use to follow service status (https://www.locaweb.com.br/ajuda/wiki/conheca-quais-sao-os-canais-oficiais-de-comunicacao-da-locaweb/).

Those pages are not neutral operating details. They show that support is part of the price proposition. A R$5.90 hosting plan cannot be understood without the possibility of a human support contact. Every forgotten password, DNS migration, email-deliverability issue, SSL failure, site outage, invoice dispute, chargeback question or marketplace integration problem has a labour cost. If Locaweb can automate enough of that work and reduce repeat contacts, support becomes a retention tool. If the support surface grows faster than ARPU, low-ticket hosting becomes a margin trap.

Public complaint surfaces are useful here as market signals, not as a verified technical audit. Reclame Aqui's Locaweb page aggregates complaints and reputation metrics for the brand (https://www.reclameaqui.com.br/empresa/locaweb/). Its Locaweb "sobre" page says the main complaint themes in the 2026 period include customer-service quality and improper billing (https://www.reclameaqui.com.br/empresa/locaweb/sobre/). Downdetector's Locaweb page tracks user-reported outages and says the service is associated with website hosting, internet services and cloud computing (https://downdetector.com.br/fora-do-ar/locaweb/). None of these pages proves a systemic failure. They do show where customers feel pain: billing, responsiveness and availability.

That pain is economically relevant because the company's model asks merchants to put more of their operating stack under the same umbrella. If a merchant only buys a domain, a support problem is annoying. If she buys hosting, email, store software, payment acceptance and logistics integration, the same support problem threatens revenue. The greater the cross-sell, the higher the trust requirement. Locaweb's own 1Q26 release says AI onboarding tools produced 30% more stores going live within the first 30 days and reduced average time to first sale by 15%; management connected that to lower churn and higher monetization as merchants sell more (https://api.mziq.com/mzfilemanager/v2/d/a8a11432-9651-4cc3-9064-ff466658c119/9521a0df-bf57-f304-c2f1-7690dd212003?origin=2). That is the right target because it moves support from reactive cost to earlier customer success.

The underwriting question is whether the public metrics start showing this at scale. A rising platform subscriber count with stable or falling support complaints would suggest leverage. Rising GMV and TPV with worsening complaint signals would suggest support debt. Flat BeOnline/SaaS revenue and falling BeOnline clients already show that the old base is not automatically expanding. Locaweb has to keep the support promise economically bounded while using it as the reason buyers do not self-assemble a cheaper stack elsewhere.

Churn is fought before the first sale, not after cancellation

Churn in Locaweb's model begins before a customer clicks cancel. It begins when the merchant fails to publish the first site, cannot configure email, cannot import products, cannot accept Pix, cannot print labels, cannot make the first sale or cannot reach support when the store breaks. This is why the 1Q26 release's onboarding numbers matter more than their technology label. Thirty percent more stores live within the first 30 days and a 15% shorter time to first sale are not cosmetic improvements. They attack the period when a small merchant is most likely to decide the tool is too hard (https://api.mziq.com/mzfilemanager/v2/d/a8a11432-9651-4cc3-9064-ff466658c119/9521a0df-bf57-f304-c2f1-7690dd212003?origin=2).

The Locaweb model needs that early success because the first product is often too small to justify a heavy human rescue. A R$5.90 hosting plan, a R$29.90 store plan or a low-volume payment account has little room for repeated support contacts. The company needs product design, onboarding, help content and automation to get merchants to revenue quickly. Once the merchant is selling, the economics can change. Payment volume, shipping volume, additional features, higher GMV limits, marketing tools and customer-service automation can raise revenue per account without adding the same amount of support cost.

The public numbers make this visible. Platform subscribers rose to 211.0 thousand in 1Q26, and platform subscription revenue rose 18.9%. LWSA says that growth came from both customer-base expansion and ARPU expansion, including merchants moving to plans with higher GMV capacity, data and functionality as they sell more (https://api.mziq.com/mzfilemanager/v2/d/a8a11432-9651-4cc3-9064-ff466658c119/9521a0df-bf57-f304-c2f1-7690dd212003?origin=2). That is positive churn math: keep the merchant, help the merchant grow, then earn more because the account needs more tools.

The counter-signal is BeOnline/SaaS. Clients at period end fell 2.4% to 381.0 thousand, and net revenue was essentially flat. There are benign explanations: portfolio changes, intentional pruning, fewer low-value accounts, product migration or less emphasis on legacy hosting. There is also a less benign reading: small-business presence products remain competitive and support-intensive. The filings do not disclose gross churn, net revenue retention or cohort economics for Locaweb hosting and commerce brands, so the public analyst has to infer from subscriber growth, ARPU, support signals and segment margins.

The IPCA adjustment clause in the billing FAQ helps protect price, but price increases can also expose churn when a merchant has cheap substitutes (https://www.locaweb.com.br/ajuda/wiki/duvidas-frequentes-cobranca/). A customer who is not selling much will compare the invoice with Hostinger, HostGator or a global cloud bundle. A customer whose store is working, whose payments are clearing and whose support experience is acceptable will compare the invoice with the cost of disruption. Locaweb's job is to move as many buyers as possible from the first comparison to the second.

Cross-sell is the margin engine Locaweb is trying to prove

Cross-sell is not a slogan in the Locaweb case; it is the difference between a low-margin hosting account and a higher-value merchant account. The official ecosystem page divides the portfolio across financial services, management and ERP, enterprise commerce, SMB commerce and BeOnline SaaS, and it describes financial services as payment automation, end-to-end payments and credit solutions (https://ri.lwsa.tech/a-companhia/nosso-ecossistema/). The same page describes Vindi as an intelligent payment hub for recurring payments, e-commerce and credit lines, with multi-acquirer billing automation, tokenization, checkout, payment links, split payments and integrations with ERPs, CRMs and e-commerce platforms. Bling, Tray, Wake, Melhor Envio, Octadesk and other brands turn the customer relationship from "host my site" into "run my online operation."

The 1Q26 release shows why this matters. Ecosystem GMV reached R$20.29 billion, TPV reached R$2.2275 billion, platform subscription revenue was R$145.6 million, and ecosystem revenue was R$116.5 million (https://api.mziq.com/mzfilemanager/v2/d/a8a11432-9651-4cc3-9064-ff466658c119/9521a0df-bf57-f304-c2f1-7690dd212003?origin=2). Locaweb is trying to attach itself to merchant activity, not only merchant presence. GMV itself is not revenue, but it is the base on which subscription tiers, payment penetration, shipping revenue and value-added tools can grow.

The enterprise side also matters. The 1Q26 release says Wake made progress acquiring and bringing new large-client accounts live, and it names brands such as Haight, Yamaha, Loja Coty and Farmacia Super Popular as examples of the diversity and complexity of clients served. It also says Wake's average ticket rose 21.8% year over year. That is important because enterprise e-commerce has different economics from a small hosting account. It can involve heavier implementation, integrations and support, but it also offers higher tickets and stronger switching costs if the platform becomes operationally embedded.

M&A created much of this surface. The history page lists the 2012 acquisition of Tray, the 2018 move into payments with Yapay, the 2019 acquisition of KingHost and Delivery Direto, the 2020 exchange listing and acquisitions including Social Miner, Etus, Ideris, Melhor Envio and Vindi, and the 2021 acquisitions of ConectPlug, Dooca, Credisfera, Samurai, Bling, Pagcerto, Bagy and Octadesk (https://ri.lwsa.tech/en/company/history/). The acquisitions page lists many of those targets as LWSA acquisition assets (https://ri.lwsa.tech/en/company/acquisitions/). The public story is clear: Locaweb used hosting credibility and capital-market access to build a merchant operating system.

The margin risk is integration. A portfolio built by acquisition can create more sales paths, but it can also create overlapping systems, brands, support queues, contract types and customer journeys. LWSA's 4Q25 release says corporate simplification included mergers of Tray, Bling and Etus in August 2024 and KingHost, Melhor Envio and Ideris in February 2025, which management tied to income-tax and social-contribution benefits (https://api.mziq.com/mzfilemanager/v2/d/a8a11432-9651-4cc3-9064-ff466658c119/402e6a52-0b7b-db4d-268e-3233ed509720?origin=2). That kind of simplification is not only a tax item. It is also how a group tries to turn acquired products into one operating base.

The cross-sell thesis will be strongest when the company shows rising ARPU, rising payment and logistics penetration, steady support quality, lower churn and lower cost to serve. It will weaken if product breadth turns into a confusing bundle, if customers buy only one low-ticket service, or if support costs rise because each product adds a new failure path. The 1Q26 numbers are encouraging on ARPU and platform subscription growth. They do not yet disclose enough to prove that every attachment is profitable.

Payments add revenue, financing cost and regulatory weight at the same time

Payments look attractive because they attach revenue to merchant volume. They are also a different kind of margin. A hosting plan is subscription revenue with support and infrastructure cost. A payments account adds take rate, fraud risk, settlement timing, receivables financing, regulation and working capital. Vindi's public pricing page makes the payment economics tangible: card receipt in 30 days starts at 2.75%; card receipt in 14 days starts at 3.25%; Pix is 0.95% with a minimum of R$1.60; boleto is R$2.99; and Bolepix has separate QR-code and barcode pricing (https://vindi.com.br/precos/). The merchant sees convenience. LWSA sees spread, volume and funding.

The 4Q25 financial statements show the cost side. Financial expenses in 4Q25 included R$12.4 million related to receivables prepayments in the payments operation, described as the structure chosen to finance working capital required for payment activities. That cost represented 0.49% of TPV in 4Q25, compared with 0.48% in 4Q24 (https://api.mziq.com/mzfilemanager/v2/d/a8a11432-9651-4cc3-9064-ff466658c119/402e6a52-0b7b-db4d-268e-3233ed509720?origin=2). The number is not large enough to break the model, but it shows why payments margin cannot be valued like pure SaaS.

The regulatory environment raises the operating standard. Banco Central do Brasil maintains public Pix statistics because Pix is national payment infrastructure (https://www.bcb.gov.br/en/financialstability/pixstatistics). Its fifth-anniversary note said Pix reached 63 billion transactions and R$26.4 trillion moved in 2024 (https://www.bcb.gov.br/en/pressdetail/2640/nota). BCB Resolution No. 85 sets cybersecurity and contracting rules for payment institutions that hire data processing, data storage and cloud computing services (https://www.bcb.gov.br/content/about/legislation_norms_docs/BCB_Resolution_No_85_2021.pdf). CMN Resolution No. 4,893 does similar work for financial institutions and institutions licensed by the Central Bank (https://www.bcb.gov.br/content/about/legislation_norms_docs/CMN_Resolution_No_4%2C893_2021.pdf).

Those rules do not mean Locaweb is a bank, and this article does not treat payments regulations as proof of any undisclosed exposure. They do mean that a payments-enabled commerce stack must be engineered, documented and supported differently from a simple brochure site. Buyers care about settlement, chargebacks, fraud, data storage, service availability and business continuity. A payment outage is not just a web-hosting inconvenience. It stops cash conversion.

Payments also change churn dynamics. A merchant that uses only hosting can migrate if the site is simple. A merchant that uses store software, payment processing, boleto, Pix, card settlement, subscription billing, shipping and customer support has a more complex migration path. That improves retention if the service works. It worsens reputational risk if it fails. This is why Locaweb's payments attachment is powerful but not free. It brings activity-based revenue and higher switching costs, but it also brings funding cost, fraud controls, regulatory documentation and more severe customer consequences when support is poor.

The network footprint shows a real local infrastructure role

Locaweb's local infrastructure relevance is visible in public internet records. PeeringDB lists AS27715 as Locaweb S.A., also known as Locaweb Servicos de Internet, with the website https://www.locaweb.com.br; an IRR as-set of AS-LOCAWEB-GROUP, a content network type, 1,000 IPv4 prefixes, 300 IPv6 prefixes, 20-50 Gbps traffic levels, mostly outbound traffic ratios and South America geographic scope (https://www.peeringdb.com/asn/27715). The same PeeringDB page lists two operational 40G connections at IX.br Sao Paulo with route-server peering and IPv4/IPv6 addresses. This does not make the AS number a company entity; it is evidence that the hosting and content delivery role has a real network base.

BGP.Tools gives a second public view. It describes AS27715 as Locaweb Servicos de Internet S/A, a 22-year-old BGP network peering with 78 other networks and using 8 upstream carriers, and it shows many originated Brazilian prefixes with RPKI-valid status (https://bgp.tools/as/27715). BGP.he.net likewise identifies AS27715 as Locaweb Servicos de Internet S/A with country of origin Brazil and many originated prefixes, including blocks associated with Locaweb and related names (https://bgp.he.net/AS27715). IPinfo's AS27715 page provides another public lookup surface for the autonomous system, ranges and peer/upstream/downstream context, even though detailed WHOIS fields may require login (https://ipinfo.io/AS27715).

The network record matters because local hosting economics are not purely software. Email, DNS, shared hosting, VPS, customer panels, backups, monitoring and support all depend on a real operational substrate. The customer may not ask which upstream carrier serves her website, but she notices latency, deliverability problems and outages. Locaweb's support and product pages sell ease and local reliability; the BGP and PeeringDB records show the company still has to operate a serious Brazilian internet footprint to make that promise credible.

The Sao Paulo interconnection context is also important. PeeringDB names IX.br Sao Paulo as the public peering exchange point for AS27715, and IX.br's Sao Paulo traffic page remains a public reference for local exchange traffic (https://ix.br/trafego/pix/sp). NIC.br said in March 2026 that IX.br hit a record 50 Tbit/s of aggregated internet traffic, driven by content and digital services, with 32 Tbit/s recorded in Sao Paulo and more than 2,500 directly connected networks in that metro (https://www.nic.br/noticia/releases/ix-br-hits-record-50-tbit-s-of-aggregated-internet-traffic-driven-by-content-and-digital-services/). Locaweb is not the exchange. The point is that Brazil's digital-service traffic is dense enough that local routes, peering and support affect customer experience.

This local network role is one reason a real-denominated Brazilian hosting bundle can still compete against global primitives. A small merchant may prefer a provider with local support, familiar billing, local network presence and Portuguese-language help, even if the raw compute can be found cheaper elsewhere. The risk is that developers and agencies increasingly know how to self-assemble VPS, CDN, email and checkout tools. Locaweb must therefore keep the local infrastructure advantage tied to service simplicity and merchant outcomes, not just legacy brand recognition.

Dollar-linked inputs create the real-vs-dollar wedge

The article's hardest margin question is currency. Locaweb sells much of its front-door proposition in reais, but cloud infrastructure, imported servers, network equipment, software licenses and global substitutes are often priced directly or indirectly in dollars. Locaweb's 4Q25 financial statements say 4Q25 cost of services fell 6.0% year over year to R$184.5 million, representing 48.4% of net revenue versus 53.9% in 4Q24, and management attributed the reduction to divestment of assets with more robust cost structures and initial synergy capture from cloud-cost migration (https://api.mziq.com/mzfilemanager/v2/d/a8a11432-9651-4cc3-9064-ff466658c119/402e6a52-0b7b-db4d-268e-3233ed509720?origin=2). The 1Q26 release says cost of services was R$187.3 million, 51.6% of net revenue, versus 53.4% in 1Q25, again linking lower cost intensity to asset sales and cloud-cost migration (https://api.mziq.com/mzfilemanager/v2/d/a8a11432-9651-4cc3-9064-ff466658c119/9521a0df-bf57-f304-c2f1-7690dd212003?origin=2).

This tells the reader two things. First, cloud costs are meaningful enough that management calls them out when margins improve. Second, after the sale of assets such as Squid and the Nextios portfolio, cloud is less a separately narrated business line and more a cost and architecture question inside the commerce-and-presence bundle. The margin is not created by owning every server. It is created by placing each workload on the right mix of owned infrastructure, local network assets, third-party cloud and software automation.

The global substitute keeps the pricing pressure visible. AWS Lightsail's $5 Linux bundle and larger USD bundles create a low anchor for developers who can operate their own stack (https://aws.amazon.com/lightsail/pricing/). AWS documentation says Lightsail data transfer above allowance varies by region and lists South America distribution overage at $0.11 per GB (https://docs.aws.amazon.com/lightsail/latest/userguide/amazon-lightsail-faq-data-transfer-allowance.html). AWS billing documentation says managed databases start at $15 per month for a 1 GB RAM database with 40 GB SSD and 100 GB data transfer allowance (https://docs.aws.amazon.com/lightsail/latest/userguide/amazon-lightsail-frequently-asked-questions-faq-billing-and-account-management.html). These are not Locaweb's costs, and they do not prove Locaweb uses any particular provider. They are the public reference points customers and developers can use when deciding whether a local bundle is worth the premium.

The FX record shows why real-denominated pricing can be both protection and risk. Banco Central do Brasil's exchange-rate page displayed U.S. dollar PTAX bid/offer quotes around R$5.19 on July 1, 2026 (https://www.bcb.gov.br/en). FRED's Brazilian real to U.S. dollar series showed June 2026 at 5.1241 reais per dollar, with monthly observations also visible for May, April, March and February 2026 (https://fred.stlouisfed.org/series/EXBZUS). Banco Central's December 2025 Monetary Policy Report considered average exchange rates of USD/BRL 5.36, 5.40 and 5.45 for the last quarters of 2025, 2026 and 2027 in one scenario (https://www.bcb.gov.br/content/ri/inflationreport/202512/rpm202512i.pdf). A local hosting company can bill in reais and index some contracts to IPCA, but dollar-linked infrastructure still moves.

The wedge is the opportunity. Locaweb can sell predictability to merchants who do not want to manage USD cloud bills. It can pool infrastructure knowledge, negotiate or optimize cloud usage, and hide some complexity inside a plan. The wedge is also the danger. If global cloud prices fall, if the real weakens, or if imported hardware and software costs rise faster than local plan prices, Locaweb has to absorb, reprice or optimize. The 2025 and 1Q26 cost-service improvement suggests management is actively working that equation. The article's judgment would change if future releases stopped showing cost leverage while revenue growth remained dependent on payments and commerce volume.

Competitors attack different layers of the same account

Locaweb does not face one competitor. It faces a stack of substitutes. At the entry layer, Hostinger and HostGator pressure shared hosting, WordPress and VPS prices. Hostinger's pricing page shows hosting from R$5.99 per month (https://www.hostinger.com/br/precos), while its VPS page shows KVM plans with promotional and renewal prices and technical resources such as vCPU, RAM, NVMe disk and bandwidth (https://www.hostinger.com/br/servidor-vps). HostGator's hosting page sells servers in Brazil, free migration, free email and 24-hour support from R$10.09 per month (https://www.hostgator.com.br/hospedagem-de-sites). These brands attack the same first decision as Locaweb: where does a small business put its first site?

At the cloud-primitives layer, AWS Lightsail and similar services attack developers and agencies who can trade support simplicity for control and dollar-denominated compute. A developer may decide that a $5 VPS, a managed database, an external email service and a payment SDK are enough. That substitute is powerful for technically confident users. It is weaker for a merchant who wants a Portuguese support path, local billing, domain and email packaged with commerce tools. Locaweb's defence is not to pretend raw compute is cheaper. Its defence is to reduce the total operating burden for the merchant or agency.

At the commerce-software layer, global and local store platforms attack the storefront. Shopify's Brazil pages position the product around orders, inventory, shipping, automation, checkout, payments and a large app ecosystem (https://www.shopify.com/br and https://www.shopify.com/br/precos). That is the same buyer problem Locaweb wants to solve through Tray, Wake, Bagy, Bling, Vindi, Melhor Envio and Octadesk. The difference is that Locaweb's story is more local and bundled with Brazilian hosting, payments and support; Shopify's story is global ecosystem breadth. Neither is automatically better. The choice depends on how much the buyer values local operation, local payment habits, implementation support, app breadth, global selling and developer availability.

At the payments layer, competitors are harder to isolate from public information because payment acceptance is often embedded in banks, acquirers, gateways, marketplaces and commerce platforms. Vindi's public pricing gives a visible reference, but it does not reveal LWSA's net take rate, funding spread, fraud losses, customer concentration or attach rate (https://vindi.com.br/precos/). The 1Q26 and 4Q25 releases disclose TPV and receivables-prepayment cost, which is enough to see that payments are material but not enough to underwrite payment margin with precision (https://api.mziq.com/mzfilemanager/v2/d/a8a11432-9651-4cc3-9064-ff466658c119/9521a0df-bf57-f304-c2f1-7690dd212003?origin=2 and https://api.mziq.com/mzfilemanager/v2/d/a8a11432-9651-4cc3-9064-ff466658c119/402e6a52-0b7b-db4d-268e-3233ed509720?origin=2).

The market context supports demand but not pricing immunity. The U.S. International Trade Administration's Brazil e-commerce guide cites expectations for Brazilian e-commerce revenue and online buyers in 2025, making clear that the addressable digital merchant base is large (https://www.trade.gov/country-commercial-guides/brazil-ecommerce). Large demand attracts competitors. Locaweb's advantage is that it can meet a merchant at several points in the journey. Its risk is that each layer has a specialist competitor with a sharp price point or stronger product in that layer.

The judgment changes when support or currency stops absorbing the spread

The facts that would change the Locaweb view are practical. The first is support efficiency. If public complaint surfaces worsen, if response signals deteriorate, or if management's future releases stop tying automation and onboarding to churn reduction, the small-merchant bundle becomes less attractive. Support is the mechanism that turns a low-ticket plan into a trusted operating stack. If support fails, the stack becomes a liability.

The second is net retention. LWSA discloses platform subscribers, BeOnline/SaaS clients, ARPU expansion language, GMV and TPV, but it does not disclose gross churn, net revenue retention, attach rate by cohort or support contacts per account. The 1Q26 release's 211.0 thousand platform subscribers and 381.0 thousand BeOnline/SaaS clients are useful, but the reader still needs to know whether customers are moving from hosting into commerce and payments profitably, or whether new commerce growth is replacing churn in older services (https://api.mziq.com/mzfilemanager/v2/d/a8a11432-9651-4cc3-9064-ff466658c119/9521a0df-bf57-f304-c2f1-7690dd212003?origin=2).

The third is cloud-cost leverage. The 4Q25 and 1Q26 releases both mention cloud-cost migration and improved cost intensity. That is a positive signal. It would become a concern if revenue growth required heavier third-party cloud use, if FX weakened materially, or if savings from divestitures and migration were one-time rather than repeatable. The comparison with AWS Lightsail's dollar pricing and BCB/FRED exchange-rate data is not a claim about Locaweb's supplier contracts; it is the public economic background against which local plans are priced (https://aws.amazon.com/lightsail/pricing/; https://www.bcb.gov.br/en; https://fred.stlouisfed.org/series/EXBZUS).

The fourth is payments funding cost. A 0.49% receivables-prepayment cost as a share of TPV in 4Q25 is manageable, but it is still a visible cost tied to the payments operation (https://api.mziq.com/mzfilemanager/v2/d/a8a11432-9651-4cc3-9064-ff466658c119/402e6a52-0b7b-db4d-268e-3233ed509720?origin=2). If interest rates, settlement timing, fraud or chargebacks move against the business, payments can pressure margins even as TPV grows. If payment penetration rises while funding cost stays controlled, payments become a stronger retention and monetization asset.

The fifth is network reliability and local differentiation. AS27715's public peering, prefixes and IX.br presence support Locaweb's local-infrastructure claim (https://www.peeringdb.com/asn/27715 and https://bgp.tools/as/27715). If the network footprint remains healthy and the support experience stays acceptable, local infrastructure continues to matter. If merchants and agencies increasingly see hosting as a commodity and assemble their own stacks from global providers, Locaweb must win through commerce workflow and support, not hosting heritage.

Finally, the corporate simplification story matters. LWSA's ownership page lists founding shareholders at 28.2%, General Atlantic at 15.8%, Kinea at 10.0%, other shareholders at 42.9% and treasury shares at 3.2%, with 568,561,350 total shares (https://ri.lwsa.tech/en/corporate-governance/ownership-structure/). This is a public listed company, not a founder-only hosting shop. The market will ask for margin proof, cash conversion and capital discipline. The 4Q25 release reported full-year free cash flow after capex of R$224.8 million, compared with R$33.1 million in 2024, and said 2025 adjusted net income was R$204.6 million (https://api.mziq.com/mzfilemanager/v2/d/a8a11432-9651-4cc3-9064-ff466658c119/402e6a52-0b7b-db4d-268e-3233ed509720?origin=2). Those are meaningful signs of execution, but they need to persist after portfolio cleanup and cloud-cost migration benefits mature.

The margin is earned in the joints between products

Locaweb's best public case is not that it owns a monopoly over Brazilian hosting. It does not. The best case is that it can convert a trusted Brazilian digital-presence brand into a merchant operating stack: site, email, store, ERP, payment, shipping, customer support, monitoring and local infrastructure. The company's history explains the trust; the 1Q26 and 2025 results show Commerce now carrying the growth; the product pages show the small monthly units that attract merchants; and the network records show a real local infrastructure role. The model is coherent.

The model is also demanding. A cheap hosting plan creates customer acquisition, not necessarily profit. Payments create activity revenue, but also financing cost and regulation. Cloud migration can improve gross margin, but the dollar remains the reference point for many inputs and substitutes. Support creates trust, but every extra product creates another support path. Cross-sell improves retention only if the merchant feels the bundle is easier than leaving.

That is why Locaweb should be watched through four operating questions. Are platform subscribers and ARPU rising faster than BeOnline declines? Are support and complaint signals consistent with lower churn? Are cloud and service costs falling as a share of revenue without cutting reliability? Are payments and logistics increasing wallet share without importing excessive funding or operational risk? The public filings currently lean positive: 1Q26 showed double-digit comparable revenue growth, 25.1% adjusted EBITDA margin, strong free cash flow and explicit progress on onboarding, ARPU and payments penetration. The unresolved issue is how much of that improvement is durable once the easy portfolio cleanup has passed.

For the Brazilian merchant in the opening scene, the answer is less abstract. If Locaweb lets her launch in days, accept Pix and cards, recover from support problems, ship orders and keep the bill in reais, the bundle is worth more than raw hosting. If she has to fight billing, downtime, slow support or confusing products, the cheap alternatives become rational. Locaweb's margin after software, payments and cloud stopped being separate is therefore earned not in any single product category, but in the moment the merchant decides that one accountable Brazilian stack is less risky than assembling the pieces alone.