Spark New Zealand: How a Telecom Monopoly Became a Shrinking Telco

Spark New Zealand is what a national telecom champion looks like after the monopoly, the network and most of the growth options have gone. This article explains how a company sold by the state for NZ$4.25bn in 1990 lost its fixed-line network, tried to reinvent itself as a digital services group, and ended up selling its mobile towers and most of its data centres to pay down debt and protect a much smaller dividend. It is part of the New Zealand Company Stories hub.

How did Spark New Zealand begin?

Spark began as the telecommunications arm of the New Zealand Post Office. It was corporatised as Telecom Corporation of New Zealand on 1 April 1987 and privatised in 1990, when two American regional phone companies, Bell Atlantic and Ameritech, bought it for NZ$4.25bn.

The sale was one of the largest of the reforming Labour government’s privatisations, and it came with an unusual condition known as the Kiwi Share. The Crown kept a single special share that obliged Telecom to keep free local calling for residential customers and to limit line rental increases. In return the buyers received an integrated national monopoly over the copper network, with very light regulation. New Zealand in the early 1990s had no specialist telecom regulator at all; competitors such as Clear Communications had to rely on general competition law and the courts to get interconnection.

For shareholders the result was superb. Through the 1990s Telecom was one of the most profitable phone companies in the developed world and the dominant stock on the local exchange, a period covered in the article on the NZX and its shrinking market. Under Roderick Deane and then Theresa Gattung, who ran the company from 1999 to 2007, it paid out most of its earnings as dividends, launched the Xtra internet service in 1996 and bought the Australian carrier AAPT, an expansion that was eventually sold in 2013 for A$450m after years of poor returns.

Why did Telecom lose its network?

Telecom lost its network because politicians concluded that a vertically integrated monopoly was holding back broadband. The government ordered local loop unbundling in May 2006, forced operational separation in 2008, and made full structural separation the price of taking part in the state-backed fibre build in 2011.

By the mid-2000s New Zealand ranked poorly in OECD broadband tables, and the blame fell on Telecom’s control of the copper lines. The unbundling announcement on 3 May 2006 wiped billions off the share price in days. Operational separation followed on 31 March 2008, dividing the company into a network unit called Chorus, a wholesale unit and a retail unit, each at arm’s length from the others.

The decisive step came with the Ultra-Fast Broadband programme. The government would co-fund a national fibre network, but only with companies that did not also sell retail services. Telecom’s board chose to split. On 1 December 2011 shareholders received one Chorus share for every five Telecom shares, and the copper and fibre access network left the group for good. That story continues in the article on Chorus and structural separation. What remained was a retailer and mobile network operator, renamed Spark on 8 August 2014 to mark the break with its past.

How does Spark make money today?

Spark makes money mainly from mobile. In FY26 mobile revenue was NZ$1,517m out of adjusted revenue of NZ$3,700m, with broadband contributing NZ$596m and the remainder coming from cloud, IT services, business connectivity, procurement and a small and fading legacy voice line.

The economics of the segments differ sharply. Mobile is the only part of the business where Spark owns the critical asset, its radio spectrum and active network equipment, and so keeps most of the margin. In broadband it is largely a reseller: it buys wholesale fibre from Chorus and the other local fibre companies and competes on price and service with dozens of retailers, which is why broadband revenue fell another 2% in FY26. Spark has tried to improve that margin by moving customers to fixed wireless access carried over its own mobile network.

The IT and cloud businesses, assembled through acquisitions such as Gen-i, Revera, CCL and Digital Island, were meant to be the growth engine. Instead they exposed Spark to the same squeeze on government and corporate technology budgets that has hurt other providers, a market described in the profile of Datacom. Legacy voice, once the core of the monopoly, is now only about 3% of revenue. Management says core connectivity produces roughly 80% of gross margin, which is the arithmetic behind the current strategy.

What went wrong in 2024 and 2025?

Spark’s enterprise and government division deteriorated much faster than management expected. A run of guidance cuts through 2024 and early 2025 ended with the shares falling more than 20% in a single day in February 2025 and the market value dropping to about NZ$4.2bn.

The half-year result in February 2025 was the low point. Net profit fell 78% to NZ$35m, EBITDAI dropped almost 21% to NZ$419m, and full-year guidance was cut from NZ$1.12bn-1.18bn to NZ$1.04bn-1.10bn. Management blamed public-sector spending cuts, businesses trimming their mobile fleets, aggressive price competition and households trading down to cheaper broadband plans. Mobile service revenue, the measure investors watch most closely, fell 3.7% in that half.

The full year to June 2025 confirmed the damage: adjusted revenue down 4% to NZ$3.7bn, EBITDAI down 9% to NZ$1,060m and net profit down 34% to NZ$227m. Free cash flow held at NZ$330m only because capital expenditure was cut by 17%. The company had for years paid dividends that exceeded the cash it generated, covering the gap with debt and asset sales. That policy could not survive a falling earnings base, and the board said so when it launched a new strategy in 2025.

Why is Spark selling towers and data centres?

Spark is selling infrastructure because specialist investors will pay far more for those assets than the share market gave Spark credit for, and because the proceeds were needed to cut debt. Towers went first, in 2022; data centres followed in a deal completed in February 2026.

The tower sale set the pattern. In 2022 Spark sold 70% of its passive mobile tower business to Ontario Teachers’ Pension Plan for NZ$900m; the vehicle, renamed Connexa, later absorbed the 2degrees towers as well, and Spark has since sold its remaining minority stake for a little over NZ$300m. Spark kept the radio equipment and now rents space on masts it used to own under a long-term contract.

The data centre transaction is larger in strategic terms. Australian private equity firm Pacific Equity Partners bought 75% of a business valued at NZ$705m, paying NZ$486m in cash up front, with up to NZ$98m more if performance targets are met by the end of 2027. Spark reported NZ$462m of net proceeds and a gain of about NZ$278m. The business, renamed TenPeaks Data Centres, runs 11 facilities with 23MW of built capacity and a development pipeline of more than 130MW that could absorb around NZ$3bn over time. Spark could not fund that pipeline itself while also paying its dividend; keeping 25% gives it exposure without the capital call. The same hunger for digital infrastructure explains why Infratil’s stake in CDC Data Centres has become that investor’s most valuable asset.

What is the SPK-30 strategy?

SPK-30 is Spark’s five-year plan, running to 2030, to narrow the company to core connectivity. It concentrates investment on mobile, takes large amounts of cost out, simplifies the portfolio through divestment and ties the dividend to free cash flow rather than to a fixed number.

The plan was unveiled alongside the weak FY25 result. Its targets are modest by the standards of earlier Spark strategies: a return to revenue growth and a return on invested capital of 11-13% by FY30. The dividend is now set at 90-100% of free cash flow. In practice that meant a payout of 16 cents a share for FY26, fully covering free cash flow of NZ$308m, against 25 cents in FY25 and 27.5 cents the year before.

Cost reduction is the most visible element. Total full-time-equivalent staff fell from 3,847 to 3,317 during FY26, a cut of nearly 14%, and the company’s own filings showed more than 5,000 employees only two years earlier. Spark booked NZ$40m of productivity benefits in the year, completed the nationwide shutdown of its 3G network and sold Digital Island. In April 2026 it reorganised into two divisions, one for connectivity and one for digital services, a structure that also makes it easier to separate or sell the services arm later if the board chooses.

Who owns and governs Spark?

Spark is a widely held public company listed on the NZX and the ASX, with no controlling shareholder. Its register is dominated by New Zealand and Australian institutions, KiwiSaver funds and a large base of retail investors who bought the shares for income.

The board is chaired by Justine Smyth, and the chief executive is Jolie Hodson, who was previously the company’s finance director and took the top job in 2019. Both have faced criticism from investors over the sequence of guidance downgrades in 2024 and 2025, and over a dividend policy that was held too high for too long. The Kiwi Share obligations still apply in modified form through telecommunications service obligations, although they now matter far less than they did in the copper era.

The shareholder base matters for strategy. Income investors valued Spark as a bond substitute, so each dividend cut produced selling from exactly the holders who had been most loyal. The savings schemes described in the article on KiwiSaver fund managers are significant owners, which makes the company’s decline a mainstream savings issue as well as a corporate one.

Who are Spark’s competitors?

Spark competes with One NZ and 2degrees in mobile and with dozens of retailers in broadband. Industry tracking in early 2026 put Spark at about 32% of mobile connections plus 6% for its Skinny brand, One NZ at 36% and a fast-growing 2degrees at 27%.

The mobile market is a three-network oligopoly, and for most of the past decade that structure allowed steady price increases. The balance has shifted. 2degrees, merged with the Orcon and Slingshot broadband business in 2022, has gained share for seven consecutive quarters, and One NZ under Infratil has used a satellite-to-mobile partnership with SpaceX as a marketing weapon. Spark has responded with its own satellite texting service and a programme of more than 100 cell site builds and upgrades. The competitive dynamics are examined in the article on One NZ and 2degrees.

In broadband the threat is structural. Fibre is a commodity input available to any retailer at regulated wholesale prices, and electricity companies now bundle broadband with power. In enterprise IT, Spark faces Datacom, global systems integrators and the hyperscale cloud providers directly. Its public cloud revenue grew about 20% in FY26, but that was offset by falling private cloud revenue, leaving the cloud line flat.

What do the latest numbers show?

The FY26 result shows a business that has stabilised but is not yet growing. Adjusted revenue was flat at NZ$3,700m, adjusted EBITDAI fell 2.4% to NZ$1,035m and adjusted profit slipped 0.9% to NZ$225m, all within guidance.

The encouraging line is mobile. Mobile service revenue rose 1.1% to NZ$998m, the first annual increase after a period of decline, and total mobile revenue rose 4.4% with help from handset sales. The discouraging lines are everywhere else: broadband down 2%, business connectivity down almost 10% and cloud flat. Capital expenditure on a business-as-usual basis was NZ$401m and is guided down to NZ$350m-380m in FY27.

The guidance range for FY27 brackets the FY26 outcome, which tells investors that management itself is not promising growth yet. The data centre proceeds took debt back to the board’s target level, removing the most immediate balance-sheet concern. The free cash flow comparison is on the company’s current definition, which is why the 18.5% increase it reported does not match the NZ$330m figure published a year earlier.

What can founders and CFOs learn from Spark New Zealand?

Spark’s lesson is that a dividend is a promise the market prices as permanent, and a promise funded by debt and disposals rather than cash flow eventually breaks. Founders and finance chiefs should match payouts to cash generated, and treat asset sales as one-off events.

• Tie distributions to cash, not to history. Spark held its dividend at or above 25 cents while free cash flow covered far less. Moving to a payout ratio earlier would have meant a gentler adjustment and less damage to credibility.
• Know which assets you are the best owner of. Towers and data centres are worth more to long-duration infrastructure funds with cheaper capital. Selling them was rational; the error would be to treat the proceeds as recurring income.
• Diversification into adjacent services is harder than it looks. Sport streaming, video on demand, IT services and health technology all promised growth beyond connectivity. Most were sold, closed or written down.
• Guide conservatively when a segment is deteriorating. Three downgrades in roughly a year cost more in market value than one honest reset would have done.
• Concentration on government customers is a cyclical risk. A change of government and a public-sector spending freeze removed a large slice of enterprise demand within months.

There is also a lesson about regulation. Telecom’s extraordinary returns in the 1990s invited the political response that broke the company up. Businesses that earn monopoly profits in a small country should expect the rules to change, a theme that recurs across the Founders Hub case studies.

What happens next for Spark New Zealand?

Spark’s next two years depend on whether mobile growth can outrun the decline in everything else. Guidance for FY27 implies flat earnings, a dividend of 16-18 cents and lower capital spending, with the TenPeaks earn-out and further portfolio simplification as possible upside.

Three questions will decide the outcome. The first is pricing discipline in mobile: if 2degrees keeps taking share and One NZ responds, Spark’s modest service revenue growth could reverse. The second is the future of the digital services division created in the April 2026 reorganisation. A sale, a partnership or a slower run-down are all plausible, and each would shrink the company further while raising its margin. The third is ownership. A company with stable cash flows, a depressed share price and a simplified structure is a natural target for infrastructure or private equity investors, the kind of buyer that has already taken its towers and data centres.

What is not in prospect is a return to the old Telecom. The network belongs to Chorus, the towers to Connexa and most of the data centres to Pacific Equity Partners. Spark’s task is to prove that the customer relationship and the spectrum, the parts it kept, are worth owning on their own.

Frequently Asked Questions