The Data Must Come Home. The Profits May Not.

CBN’s payment-data localisation order and the NGX stocks being fitted for crowns they have not yet earned

Seven Gates Research | The Lokoja Contrarian | 25 August 2026 | 24-minute read

Independent research on companies, markets and power. The NGX Valuation Lab is where the arguments meet the arithmetic.

The market heard that Nigerian payment data must live in Nigeria and immediately promoted every company with a server room to Duke of the Cloud. By lunchtime, a systems integrator had become Digital Realty, a payment processor had become a landlord, and an announced subsidiary had already collected rent from racks that do not exist.

That is too quick.

The Central Bank of Nigeria’s June circular is consequential. From 1 January 2027, financial institutions and payment participants must ensure that payment-transaction data generated in Nigeria is stored and managed in Nigeria.1 The same circular also tightens beneficial-ownership disclosure, market-concentration limits and systemic oversight. It will bring work to Nigerian data centres, connectivity providers, cloud platforms, cyber-security firms and systems integrators. Some revenue will be recurring. Some will be the one-off wages of migration. Some firms will merely pay the bill.

The investable conclusion is less theatrical than the theme:

Regulation can compel a purchase. It cannot confer a moat.

Among the liquid listed names, MTN Nigeria is the highest-quality business and reasonably priced, but data localisation is a garnish on its earnings, not the meal. CWG is the most direct small-cap implementation play, but its cash conversion and undisclosed data-centre economics make the current price a watch, not a chase. Airtel Africa owns serious future capacity through Nxtra, but the Lagos facility is still being built and the shares have already run hard. eTranzact and Chams are not clean data-centre beneficiaries at all. At current prices, both ask investors to pay mature-franchise multiples for benefits that remain conditional, indirect or imaginary.

The verdict before the sermon

Prices and market values below are as at 24 or 25 August 2026. Multiples are approximate because interim reporting dates and share-price cut-offs differ.

These are valuation disciplines, not price targets engraved on stone. MTN deserves a premium for its network, cash generation and repaired balance sheet. CWG deserves a discount for working-capital absorption. eTranzact and Chams need earnings to catch the prices, not another policy headline.

The quality score is our 100-point assessment of business economics, balance-sheet strength, cash conversion, competitive endurance and disclosure. It is a comparative judgment, not a credit rating.

What CBN actually ordered

The circular, dated 15 June 2026, has four relevant parts.2

First, payment-transaction data generated in Nigeria must be stored and managed in Nigeria from 1 January 2027. The regulated population is broad: banks, fintechs, payment service providers and other institutions facilitating domestic payments.

Second, participants must disclose their ultimate beneficial owners. This is aimed at opacity in ownership and control, not server location.

Third, CBN introduced a paired market-concentration rule. An institution above 25 per cent of card issuance may not also exceed 15 per cent of merchant acquiring, with the reciprocal constraint applying in the other direction. Monthly market-share returns are required. This can matter to smaller processors and acquirers even if they never rent an additional rack.

Fourth, systemically important participants face greater supervisory reach, reporting and potential sanctions.

The one-sentence localisation command is clear enough to alter budgets and vague enough to occupy lawyers. What is “payment transaction data”? Does it include fraud models trained on the data, token vaults, dispute files, call recordings, device fingerprints and telemetry? Does “managed” prohibit a foreign engineer from administering a Nigerian instance? May an encrypted disaster-recovery copy sit offshore if the primary and keys remain in Nigeria? Does AWS’s Lagos Local Zone qualify? What happens to a card transaction that begins in Kano, is authorised by an international scheme and settles through several jurisdictions?

Kafka would recognise the architecture. The door is plainly marked; the corridor behind it has no floor plan.

This is not a pedantic problem. The answers decide whether a fintech needs a few local database instances, an entirely duplicated production stack, or a new operating model. Those are different invoices.

The legal tension is real, but it is not an investment thesis

Nigeria’s Data Protection Act does not impose a universal ban on cross-border transfers. It permits them where adequacy, safeguards or another lawful basis exists. The Nigeria Data Protection Commission’s General Application and Implementation Directive treats financial institutions, payment gateways, fintechs and telecoms as very-high-impact controllers, but it still contemplates governed cross-border transfers.3

CBN’s circular is therefore stricter than the general data-protection regime. Privalex is right to flag the tension, the breadth of the language and the market consequences.4 Yet a sector regulator can ordinarily impose stricter conditions on its licensees where financial stability, supervision and operational resilience are concerned. The useful investor question is not whether a clever counsel can identify friction. Counsel always can. The question is whether CBN will clarify, enforce and audit the rule in a way that forces incremental domestic expenditure.

Our answer is yes, but not uniformly and not at the scale implied by the more excited share-price stories.

The strongest case for a genuine boom

There is a serious argument against our caution.

Nigeria’s payment volumes are large and rising. KPMG cites 11.2 billion NIBSS Instant Payment transactions in 2024, with transaction volumes up 120 per cent between 2022 and 2024. Records may need to be retained for at least five years. KPMG also identifies practical deficiencies in compute, storage, security, disaster recovery, network redundancy, power, cooling, operations and specialist skills.5 Foreign-backed fintechs and offshore-first businesses have barely four and a half months from this article’s date to redesign critical systems. Capacity cannot be summoned by circular.

If CBN interprets “managed in Nigeria” strictly, requires local disaster recovery, audits the physical location of keys and backups, and refuses elastic waivers, Nigerian operators could enjoy a step-change in occupancy and pricing. The deadline itself strengthens vendors: nobody negotiates beautifully with a regulator’s clock on the wall.

That is the best version of the thesis. It is plausible. It is not yet evidenced in the reported numbers of the listed companies.

Why the windfall may be smaller than the headline

Much of the banking system is already local

The directive does not begin from zero. Rack Centre says CBN has required banks since 2015 to maintain at least one data centre in a Tier III-certified facility and that most banks already hold primary or secondary storage in such facilities.6 Uptime Institute’s Nigerian certification register includes facilities operated for or by GTBank, Galaxy Backbone, MainOne, MTN, OADC and Rack Centre, among others.7

That industry claim should not be mistaken for a complete audit of every workload. It does establish the direction of travel: large banks already own Nigerian infrastructure, rent local colocation, or both. Their incremental requirement may be remediation at the edges, not wholesale repatriation. The bigger forced buyers are likely to be offshore-first fintechs, international payment companies, smaller PSPs using foreign clouds, and cross-border platforms whose Nigerian operations were designed as a branch of a global system.

Payment records are important, not obese

Eleven billion transactions sounds like a storage bonanza until one asks what a transaction record weighs. It is not a Nollywood master file. The record itself may be measured in kilobytes. Replication, audit logs, tokenisation, fraud data, encryption overhead, backups and five-year retention make the estate larger, but raw bytes are not the principal economic prize.

The expensive services are reliable power, cooling, network paths, compute, database performance, cyber defence, key management, disaster recovery, migration, continuous monitoring and the engineers who prevent a Friday-night settlement failure. A megawatt kept alive through grid instability has more commercial meaning than a petabyte written in a brochure.

Announced megawatts are not occupied racks

Nigeria’s data-centre market is undeniably growing. One industry estimate puts revenue at about $323 million in 2025 and $374 million in 2026, reaching $783 million by 2031.8 Such estimates vary widely by what they count. They are useful as a compass, not a cadastral survey.

Capacity still has to be financed, energised, certified, connected, sold and collected. Utilisation matters. Contract length matters. Churn matters. Carrier neutrality matters. So does the uncomfortable fact that the customer may already own a facility.

Wu-Tang’s monetary theory survives contact with Nigerian cloud infrastructure: cash still rules. The rack must be occupied, invoiced and collected. Announced capacity earns no rent.

Foreign platforms do not necessarily lose

The policy says “in Nigeria”, not “owned by Nigerians”. AWS operates a Lagos Local Zone, and global operators can deploy infrastructure inside Nigerian colocation sites.9 Equinix owns MainOne and MDXi. Digital Realty owns Medallion. A strict residency rule may divert spend from South African or European regions while leaving the customer relationship, software layer and a portion of the economics with a foreign platform.

The winner may be a Nigerian building with a foreign logo on the contract. National geography and national ownership are not synonyms.

Follow the invoice

The value chain is more useful than the company names being circulated in chat rooms.

This is the analytical error at the centre of the theme: server adjacency is not server ownership, and server ownership is not automatically a profitable data-centre business.

MTN Nigeria: the best company, not the purest policy trade

At ₦779, MTN Nigeria is worth roughly ₦16.36 trillion. Its trailing earnings of about ₦67 a share put the stock on 11.6 times earnings.10 In a market where nominal growth can flatter everything and the risk-free rate is not shy, that is not bargain-basement. It is reasonable for a company whose balance sheet, margins and cash generation have been repaired.

H1 2026 service revenue rose 25.9 per cent to ₦2.98 trillion. EBITDA increased 39.2 per cent to ₦1.67 trillion, with a 55.9 per cent margin. Profit after tax rose 70.6 per cent to ₦707.5 billion. Free cash flow reached ₦712.7 billion, net cash was ₦116.3 billion and the company had eliminated foreign-currency loans. Subscribers reached 92.2 million, including 55.7 million active data users.11

Those numbers describe a strong telecom and fintech-distribution franchise. They do not describe a data-centre stock.

The capacity is real

MTN’s Dabengwa facility is no artist’s impression. The group has committed more than $240 million to a two-phase, 9MW project. The first 4.5MW phase entered commercial service in 2025, with disclosures around 700-plus processing racks and almost 49 petabytes of storage.12 MTN brings fibre, mobile distribution, enterprise relationships, engineering depth, power-management experience, brand and a balance sheet large enough to fund expansion. Those are actual advantages.

The directive can improve customer acquisition and shorten the debate over whether regulated workloads should be local. MTN can bundle colocation, connectivity, cloud, security and managed services. A smaller vendor has to assemble that orchestra one instrument at a time.

The arithmetic refuses to become excited

In H1 2026, the “other” line containing cloud, infrastructure and ICT services produced about ₦23.9 billion, less than 1 per cent of group revenue. Enterprise revenue did not show the inflection that a localisation boom would require.13 Even the whole $240 million facility is a small asset beside a ₦16.36 trillion equity value. A spectacular ramp can be operationally important and still barely disturb near-term earnings per share.

There is also a customer-perception problem. Large banks and fintechs often prefer carrier-neutral facilities where they can connect to several networks without making a strategic competitor their landlord. Equinix, Digital Realty, Rack Centre and OADC can sell neutrality. MTN sells integration. Both propositions have customers, but they are not the same moat.

Moat, price and action

MTN’s moat is the combined system: spectrum, towers and radio network, fibre, customer distribution, enterprise sales, MoMo, operational know-how and a local-currency balance sheet. The data centre strengthens that system. It does not create it.

Using trailing EPS of about ₦67 and a justified range of 12.25 to 13.5 times gives a rough value of ₦820 to ₦905. We would accumulate below ₦720, where the multiple falls towards 10.7 times and the return does not require heroic cloud assumptions. At ₦779, hold or accumulate selectively. Buy it for the network and cash flow. Accept localisation as optional upside. Anyone buying chiefly for CBN’s circular has mistaken a side dish for a restaurant.

What would make us more bullish: disclosed data-centre occupancy, contracted backlog, cloud gross margin and a separate enterprise infrastructure revenue line growing materially faster than the group.

What would make us retreat: renewed leverage, weaker cash conversion, price regulation, carrier-neutral competitors winning anchor banks, or capex rising faster than monetisation.

CWG: the closest listed shovel, with mud on the cash-flow statement

CWG is the most intellectually defensible small-cap beneficiary. It provides IT infrastructure, software, platform services, systems integration, managed services and colocation. It has longstanding banking relationships and Finacle expertise. When a financial institution must map data, buy hardware, redesign architecture, migrate databases, test disaster recovery and document compliance, CWG can plausibly appear on several invoices.14

This is better positioning than simply owning a payment licence. It still needs underwriting.

The reported business is growing

FY2025 revenue rose to ₦65.55 billion. Gross profit was ₦15.94 billion, profit before tax ₦7.87 billion and profit after tax ₦4.97 billion, equivalent to EPS of ₦1.97.15 At ₦21.40, the market capitalisation is about ₦54 billion and the stock trades at roughly 10 to 11 times recent earnings. On the face of it, that is sensible.

The face is not the whole person.

CWG’s mix is weighted towards infrastructure and software supply, where vendor costs and imported equipment can dominate revenue. Its own colocation offer is real, but the company does not publish MW, rack count, certification, occupancy, average contract life, segment revenue or return on invested capital for the facility. Uptime Institute’s public Nigerian list does not identify a CWG-certified commercial facility.16 That does not prove the facility is poor; it proves the investor lacks evidence for treating CWG as a scaled data-centre landlord.

Profit arrived before cash

FY2025 operating cash flow was negative by about ₦2.17 billion. Contract assets rose to roughly ₦11.6 billion, an increase of more than six times, while inventories and receivables also absorbed capital.17 By Q1 2026, inventories were about ₦11.7 billion, receivables ₦27.2 billion and cash ₦3.7 billion.18

This is the fulcrum. Large implementation contracts can lift accounting revenue while cash sits with customers, vendors demand payment and imported equipment waits in inventory. Localisation could magnify the opportunity and the financing burden at the same time. A company can grow itself into a rights issue if collection discipline fails.

CWG’s P&L makes a persuasive speech. The cash-flow statement keeps asking for receipts.

Moat, price and action

CWG’s moat is moderate: embedded relationships, certifications, vendor partnerships, banking-domain knowledge and engineers who understand legacy systems. Switching costs are meaningful during a migration. They are weaker after delivery. Global vendors control important products, customers have bargaining power and rival integrators can hire talent.

On normalised EPS of ₦2.00 to ₦2.20 and a cautious 9 to 10.5 times multiple, fair value sits around ₦18 to ₦23. Current price is within that range, not below it. We would want ₦18 or less, or proof that cash conversion has improved, before taking a speculative position.

CWG is the best expression of the implementation trade, not the best company in the set. Position sizing should respect liquidity, receivable risk and the possibility that migration revenue peaks before the market notices.

What would make us more bullish: named localisation contracts, a disclosed recurring managed-services backlog, positive operating cash flow, lower contract assets and credible data-centre unit economics.

What would make us retreat: another year of negative operating cash flow, rising inventory, customer concentration, material FX losses or capex without occupancy disclosure.

Airtel Africa: 38MW of optionality, presently under construction

Airtel Africa belongs on the list even though it is often omitted. Its Nxtra unit is building a 38MW hyperscale facility at Eko Atlantic and aspires to a broader African data-centre network.19 If completed, energised and leased, that is far more capacity than MTN’s disclosed Lagos project.

The important verb is building.

Nxtra’s Lagos campus is not yet an operating earnings stream. CBN’s directive can help pre-leasing discussions with banks, fintechs and foreign platforms. Airtel can also combine connectivity, mobile distribution, enterprise sales and regional reach. Its balance sheet and brand confer credibility that a new independent operator must buy expensively.

The objections are substantial. Construction, power connection, certification, tenant acquisition and commissioning remain ahead. A telecom-owned facility faces the same carrier-neutrality question as MTN. The listed group spans 14 African markets; Nigeria data-centre rent must become very large before it moves consolidated earnings. Airtel Money’s proposed listing, currency movements and the core mobile business will dominate the share price.

At about £3.36 in London, the shares trade around 24 times trailing earnings but roughly 12.3 times forward consensus earnings after a powerful rerating.20 That is not absurd if delivery is excellent. It is not a fresh policy bargain. We would become interested around £2.70 to £2.90, where execution risk is better paid.

Verdict: hold or watch. Nxtra is credible optionality; do not pay today as though 38MW is occupied tomorrow.

eTranzact: a payment company cannot invoice itself into prosperity

The instinct to exclude eTranzact from the data-centre beneficiaries is correct. CBN’s own register identifies eTranzact as a mobile-money operator and switching/processing participant.21 It moves payments. It does not own a disclosed commercial data-centre platform of consequence.

For the localisation limb, eTranzact is closer to the regulated customer than the landlord. It may have to audit workloads, retain more local capacity, strengthen disaster recovery and absorb higher security or hosting costs. Existing local integrations and regulatory experience can make it relatively better prepared than an offshore-first rival, but relative readiness is not new revenue.

The circular’s market-concentration rules offer a more interesting route. If dominant issuers or acquirers must reduce overlap, smaller processors could win mandates. Yet the circular does not name eTranzact, guarantee migration or remove competitors. NIBSS, Interswitch, Paystack, Flutterwave, Moniepoint and bank-owned capabilities do not disappear because a ratio changed.

The valuation assumes a much better business

FY2025 revenue was ₦30.61 billion and profit after tax ₦2.47 billion, with EPS around ₦0.27.22 H1 2026 revenue rose 23 per cent to ₦16.28 billion, but cost of sales grew 39 per cent and profit after tax fell 19 per cent to ₦1.22 billion.23

At ₦14.50 and a market value of ₦133.4 billion, the stock trades at roughly 54 times FY2025 earnings and about 61 times our trailing estimate. The market is charging a network-effect multiple while gross economics are moving in the wrong direction.

The company’s real moat consists of licences, integrations, settlement relationships and a long operating history. That is worth something. It has not yet produced the margins, cash compounding or dominant network that would justify this price. Payment switching can be sticky, but customers are concentrated, technology changes quickly and regulation can convert an advantage into a cost centre.

Normalised EPS of ₦0.24 to ₦0.30 at 18 to 20 times suggests ₦4.30 to ₦6.00. We would require ₦5 or below, or a decisive earnings acceleration with stable margins. At ₦14.50, avoid.

What would make us wrong: disclosed share gains caused by the concentration rule, rising transaction take rates, sustained margin expansion and free cash flow that grows with profit.

What would confirm our caution: compliance costs, another profit decline, continued multiple expansion without cash growth, or market-share gains that require uneconomic pricing.

Chams HoldCo: an option wearing a hard hat

Chams is frequently put into the basket because it has technology, payments and now ChamsCorp, a subsidiary launched in February 2026 to pursue AI, cloud and data-centre opportunities through partnerships.24 The language is ambitious. The capacity is not disclosed.

ChamsCorp is not yet a data centre. It is a corporate intention wearing a hard hat.

The existing group comprises card personalisation, identity and access services, ChamsSwitch and ChamsMobile. These businesses may benefit from wider digitisation, card issuance, identity programmes and payment growth. ChamsSwitch, however, is itself a payment participant subject to the localisation obligation. Its compliance expenditure should not be counted as group revenue merely because another subsidiary hopes to enter infrastructure.

Better profit, weaker operation

FY2025 revenue rose 18 per cent to ₦17.50 billion and profit after tax increased 55 per cent to about ₦606 million.25 The group also raised fresh equity, taking the share count to 9 billion and strengthening liquidity.

H1 2026 revenue grew only 2.5 per cent to ₦10.34 billion. Gross profit improved 31 per cent, but administrative costs rose 71 per cent and operating profit fell 44 per cent to ₦529 million. Finance income of ₦539 million, against ₦18 million a year earlier, helped profit after tax rise 13 per cent to ₦473 million.26

That is not the earnings composition on which to capitalise a new infrastructure franchise. Finance income is welcome cash, not proof that the core operation has acquired a moat.

At ₦4.05, Chams is worth ₦36.45 billion. Trailing earnings are roughly ₦660 million by our estimate, producing a P/E near 55 times and a return on equity in the mid-single digits. The market is paying for several years of successful execution before the company has disclosed site, MW, financing, certification, customers, occupancy or economics for the data-centre plan.

The existing moat is modest: licences, public-sector and enterprise relationships, card-personalisation capability, identity expertise and installed systems. The new infrastructure moat is zero until capital and customers turn it into one.

At normalised EPS of ₦0.07 to ₦0.09 and 16 to 18 times, value is roughly ₦1.10 to ₦1.60. Our preferred entry is ₦1.50 or below. At ₦4.05, avoid.

What would make us wrong: a financed, certified facility with a credible operating partner, anchor tenants, disclosed contracted MW, and recurring cloud or colocation revenue visible in the accounts.

What would confirm our caution: continued reliance on finance income, further dilution, rising central costs, or another announcement without physical and financial milestones.

The firms most likely to benefit are mostly not on the NGX

The cleanest operational winners are owners of certified, connected and already energised capacity.

Equinix/MainOne/MDXi. Equinix owns MainOne and operates Lagos facilities with an interconnection ecosystem that is difficult to reproduce quickly.27 The moat is not concrete. It is the dense community of carriers, banks, clouds and enterprises connected inside the building. Each additional participant makes the site more useful to the others. That is a genuine network effect. Nigeria is still too small to make the CBN circular a primary reason to buy global Equinix shares.

Digital Realty/Medallion. Digital Realty operates several Lagos facilities through Medallion and has added capacity.28 It brings global customers, design standards, capital and carrier-neutral positioning. Again, a Nigerian demand bump is desirable and immaterial to the valuation of the global REIT.

Rack Centre. The Actis/Jagal-backed operator has certified capacity, industry relationships and expansion plans.29 It is a plausible direct beneficiary and privately held.

OADC. Open Access Data Centres markets a 24MW Lagos site load with Tier III design, PCI-DSS and ISO credentials.30 It has the right product for regulated workloads and is private.

Africa Data Centres and specialist local cloud, cyber and integration firms. Several may win meaningful contracts. Public investors cannot buy most of them directly. This is why the listed basket feels awkward: the purer assets sit behind private ownership or inside global companies where Nigeria is a rounding error.

The Medici understood that control of ledgers was control of power. CBN has made the ledger’s address part of that control. The immediate rent, however, will go to whoever already has a secure, powered room and a contract, not whoever gives the most Renaissance speech about digital sovereignty.

What is the moat, really?

Data-centre economics can look commoditised from outside: land, walls, chillers, generators and racks. The strongest businesses build several reinforcing barriers.

1. Power and operational reliability. In Nigeria, power is not a utility assumption. It is a manufactured service. Fuel procurement, grid strategy, batteries, generators, maintenance and uptime records matter.
2. Certification and trust. Banks do not move settlement systems into a facility because the reception is attractive. Tier certification, PCI-DSS, ISO controls, audits and incident history reduce perceived risk.
3. Connectivity density. Multiple carriers, cloud on-ramps, internet exchanges and cross-connects create network effects. This favours carrier-neutral operators.
4. Customer switching cost. Once racks, links, controls and disaster-recovery procedures are installed, moving is expensive and risky. Contracts can be sticky.
5. Scale and cost of capital. New capacity is capital-intensive and imported equipment creates FX exposure. Cheap capital and procurement scale matter.
6. Managed-service depth. Operators that sell monitoring, security, databases and migration can earn more per customer, but also assume more liability.

On this test, Equinix/MainOne, Digital Realty/Medallion, Rack Centre and OADC have clearer data-centre moats than the local listed names. MTN has the strongest overall corporate moat but a less neutral facility. CWG has customer and implementation know-how without disclosed scale. Chams has ambition. eTranzact has a payment moat of uncertain width and is on the buying side of the rule.

What the market may be missing

The circular’s competition rule may matter more to listed payment companies than the localisation rule. If CBN forces separation between concentrated issuance and acquiring positions, mandates can move. Smaller switches, processors and acquirers could gain volumes without owning a single rack. This is the legitimate eTranzact and ChamsSwitch upside case.

It is also the least underwritten part of the circular. Investors need current market shares, the identities of institutions caught by the paired caps, the remediation timetable and evidence that mandates will go to independent processors rather than bank affiliates or larger private fintechs. Until then, it is an option, not an earnings estimate.

The second overlooked point is that systems integration may have a sharper 2026 revenue pulse than colocation. Migration must occur before recurring rent begins. CWG and private integrators may book hardware and project revenue first. The danger is extrapolation: a one-off compliance sprint can produce a fine year and a poor terminal multiple.

The third is that clarification can move value between layers. If CBN permits a Nigerian primary with offshore disaster recovery, fewer local racks are required. If remote foreign administration is allowed, global managed-service firms keep more work. If encryption keys and all replicas must remain in Nigeria, local security and secondary-site demand rise. A definition in an FAQ can be worth more than ten launch events.

Risks that can spoil the party

Policy and enforcement

• CBN may delay the deadline, phase compliance or grant waivers because the migration window is short.
• Guidance may define “managed” less strictly than the market expects.
• Enforcement may be uneven, turning compulsory demand into prolonged negotiation.
• The rule may create legal challenge or conflict with cross-border scheme requirements, though outright invalidation is not our base case.

Industry economics

• Banks may use owned capacity, leaving less incremental colocation demand.
• New builds may create excess capacity after the initial migration wave.
• Diesel, power, imported equipment and FX can consume nominal revenue growth.
• Customers are sophisticated and concentrated; anchor tenants can negotiate hard.
• A cyber incident or outage can erase years of trust and expose the operator to claims.
• Technology density can improve, allowing more compute per rack and reducing physical-space growth.

Company-specific execution

• MTN and Airtel may spend heavily before utilisation matures, while carrier-neutral rivals win financial institutions.
• CWG may convert revenue into receivables and inventory rather than cash.
• eTranzact may bear compliance cost without winning concentration-rule mandates.
• Chams may dilute shareholders or fund infrastructure before establishing demand.

Market and valuation

• Small NGX technology stocks can rerate on thin liquidity and reverse without ceremony.
• Naira devaluation makes nominal growth look healthier while imported capex and shareholder purchasing power deteriorate.
• A policy label can keep a stock expensive longer than valuation discipline feels comfortable. It does not make the eventual arithmetic optional.

The evidence we require by January 2027

The thesis should be updated when facts arrive, not when enthusiasm gets a new font.

Valuation in three currencies

For scale, we translate Nigerian market values at approximately ₦1,347/$, ₦1,572/€ and ₦1,836/£, using 24 August official and market cross-rates.31 These are indicative, not trading conversions.

Airtel Africa’s primary London quotation implies a market value around £11.8 billion, with cross-listing and timing differences explaining why a mechanical conversion of the NGX line can differ.32

The contrast is instructive. A single institutional contract could move CWG’s or Chams’s reported revenue. It cannot move MTN or Airtel to the same degree. Conversely, the large telcos can fund infrastructure without staking the company. The small firms have more sensitivity and less resilience. That is not automatically a better bet.

Final judgment

The directive is real. The demand response will be real. The neat listed-stock basket is not.

MTN Nigeria is the only name here we would be comfortable owning as a core-quality position at a reasonable valuation. The circular improves an existing enterprise proposition, but the investment case stands without it. Accumulate below ₦720; at ₦779, buy selectively rather than chase.

CWG is the most direct listed implementation exposure. At ₦21.40, valuation is fair and cash conversion is not. Wait for ₦18 or evidence that contracts become cash. It can be a good trade and eventually a good investment, but only if working capital stops eating the celebration.

Airtel Africa has the largest announced listed capacity optionality. The facility is unfinished and the shares have already rerated. Watch below £2.90.

eTranzact should not be purchased as a data-localisation beneficiary. It is a processor facing compliance cost, with a separate and speculative opportunity from competition rules. At more than 50 times earnings, avoid.

Chams owns useful technology businesses and an infrastructure ambition. The market price capitalises the ambition before the asset, tenants or cash flow. Avoid above ₦1.50.

The likely direct winners are operating, certified private facilities and the foreign groups that own Nigerian capacity. Their edge is neither patriotism nor vocabulary. It is power, connectivity, trust, occupancy and contracts.

CBN can order the data home. It cannot order returns on capital. That part remains the investor’s work.

Research notes and sources

This report distinguishes reported fact from Seven Gates estimates and opinion. “TTM est.” combines the latest full year with subsequent interim results where comparable figures permit. Fair-value ranges use normalised EPS and explicitly stated multiples rather than a false-precision DCF. This is independent research, not personal financial advice.

1. Central Bank of Nigeria, Circular on market structure requirements, data localisation, beneficial ownership and systemic oversight, 15 June 2026.↩︎

2. Central Bank of Nigeria, Circular on market structure requirements, data localisation, beneficial ownership and systemic oversight, 15 June 2026.↩︎

3. Nigeria Data Protection Commission, General Application and Implementation Directive 2025.↩︎

4. Privalex Advisory, The CBN Payment Data Localisation Directive: legal tensions, market consequences and the road to 1 January 2027.↩︎

5. KPMG Nigeria, CBN’s Data Localisation Directive, August 2026.↩︎

6. Rack Centre, Meeting the rising data-centre demand in Africa. The 2015 requirement is the operator’s description and should be read as industry evidence, not our independent legal verification.↩︎

7. Uptime Institute, Nigeria awards and certifications register.↩︎

8. Mordor Intelligence, Nigeria Data Center Market. Market-research estimates differ by scope and methodology.↩︎

9. Amazon Web Services, AWS Local Zones locations.↩︎

10. NGX, MTN Nigeria company profile; Financial Times, MTNN summary and trailing EPS.↩︎

11. MTN Nigeria, H1 2026 earnings release.↩︎

12. MTN Nigeria, Dabengwa Data Centre investment announcement.↩︎

13. Techpoint Africa, MTN Nigeria’s enterprise revenue and cloud/infrastructure disclosure, 21 August 2026. Segment descriptions differ slightly from the headline enterprise channel in MTN’s release; we use the disclosure only to establish scale.↩︎

14. CWG, Colocation and IT infrastructure services.↩︎

15. CWG, Investor relations and FY2025 results; NGX Pulse, CWG market data.↩︎

16. Uptime Institute, Nigeria awards and certifications register.↩︎

17. Technext, CWG’s ₦65.6bn revenue and negative operating cash flow.↩︎

18. Simply Wall St community analysis based on CWG’s Q1 filing, Q1 2026 working-capital figures. Secondary source; figures should be reconciled to the exchange filing before trading.↩︎

19. Airtel Africa, Nxtra data centres.↩︎

20. Yahoo Finance, Airtel Africa valuation statistics; NGX Pulse, Airtel Africa market data.↩︎

21. Central Bank of Nigeria, Licensed payment service providers.↩︎

22. AfricanFinancials, eTranzact 2025 annual report; NGX Pulse, eTranzact market data.↩︎

23. BusinessDay, eTranzact H1 2026 profit falls 19 per cent.↩︎

24. ChamsCorp, Corporate site; TechCabal, Chams expands beyond payments with AI and data-centre play.↩︎

25. AfricanFinancials, Chams HoldCo 2025 annual report; NGX Pulse, Chams market data.↩︎

26. Chams HoldCo, H1 2026 performance release.↩︎

27. Equinix, Lagos data centres.↩︎

28. Digital Realty, Lagos data centres; Data Center Dynamics, Digital Realty launches additional 2MW facility.↩︎

29. Actis, Rack Centre portfolio.↩︎

30. OADC, Lagos facility.↩︎

31. Central Bank of Nigeria, official exchange rates; Reuters, 24 August 2026 sterling/dollar market context.↩︎

32. Yahoo Finance, Airtel Africa valuation statistics; NGX Pulse, Airtel Africa market data.↩︎