The Chain Nobody Follows
Telecommunications costs are reported at the top of the chain and experienced at the bottom, and almost nothing connects the two in public.
A spectrum valuation is announced and covered as a lobbying story. A carrier result is covered as a share price story. An NBN wholesale change is covered as a consumer broadband story. Meanwhile the person actually affected — someone signing a three-year agreement for twenty-two extensions and a business internet service — gets no help at all working out whether any of it matters to them.
The chain is real. It is also five steps long, with years of lag, and the final step is set by competition rather than by cost. That combination is why the obvious conclusion is usually the wrong one.
The argument of this article in one paragraph
So this walks the chain properly, from the spectrum price to the invoice, and then names the three things that actually change what an Australian business pays — none of which appears in any of the three headlines.
Seven Point Three Two Billion Dollars
Mobile spectrum in Australia is licensed rather than owned. A substantial tranche of those licences expires between 2028 and 2032, and instead of re-auctioning them the ACMA set a renewal price. A preliminary figure of $7.34 billion was published in December 2025; the final valuation settled at $7.32 billion.
| Element | Detail |
|---|---|
| Bands covered | 700MHz, 850MHz, 1800MHz, 2GHz, 2.3GHz, 2.5GHz and 3.4GHz |
| What they carry | 4G and 5G mobile, plus fixed wireless broadband — underpinning more than 30 million mobile services in Australia |
| Who pays | Telstra, Optus, TPG and NBN Co |
| How it was priced | Per megahertz per head of population, band by band. Lower bands revised slightly down, mid-band prices revised up |
| Timetable | Renewal applications opened 18 June 2026, starting with 850MHz and 1800MHz, with nine-month application windows |
| When money moves | As each licence expires from 2028 onward. Nothing is being paid today |
Why it matters now even though it is paid later
Capital planning runs years ahead of cash. A carrier that knows it owes billions in 2028 makes different decisions in 2026 about what to build, where to build it, what to retire and what to charge. That is the real transmission mechanism, and it is much more consequential than the headline figure — which is precisely why the industry fought the valuation so hard.
The Fight Over Whether It Is Too High
The industry objection deserves stating fairly rather than waved through, because it might well be correct and a reader is entitled to weigh it.
Telstra argued the ACMA had overvalued the spectrum by $3.3 billion, and urged Treasury to cap the total at $3.9 billion — roughly half what the regulator settled on. The Australian Telecommunications Alliance warned that higher spectrum costs would mean “less investment, or higher prices or both”. The telcos were unable to overturn the valuation.
The case that it is too high
Spectrum is an input cost. Inflated input costs come out of either customers or network investment, and regional coverage is the first thing cut when capital tightens — which is exactly the area everyone already agrees is the weakest.
The case that it is not
Spectrum is a scarce public resource. Underpricing it transfers value from the public to shareholders, and there is no mechanism guaranteeing the saving would reach customers rather than dividends.
The methodological objection
A price set administratively rather than by auction is an estimate, and this is a large estimate. The counter is that auctioning expiring incumbent-held licences has well-documented problems of its own, which is why the renewal mechanism exists at all.
What a business should conclude
Nothing about who is right. The number is settled, it is large, it lands from 2028, and every carrier now has it in their planning. That last fact is the usable one.
The Year Telstra Just Had
Telstra published its full-year results for the year ended 30 June 2026 on 13 August 2026.
$23.405b
Total income, down 0.9%
$2.406b
Net profit, up 2.7%
+11.6%
Cash earnings, to $2.9b
270k+
Mobile services added
Operating profit rose 1.7% to $4 billion, earnings per share rose 5.3% to 19.9 cents, cash earnings per share rose 13.8% to 25.5 cents, and EBITDAaL came in at $8.2 billion. A total dividend of 21.0 cents was declared, with a final dividend of 10.5 cents, alongside a $1 billion share buyback. Mobile service revenue grew 3.2%. It is the first year of the Connected Future 30 strategy.
Do not read a carrier result as a pricing forecast
Income fell while profit rose, which is a cost-discipline story rather than a pricing story. Carrier results and retail prices are only loosely connected in a market this concentrated, and a strong result is entirely compatible with prices going up, down or nowhere. The genuinely useful signal is subtler: adding more than 270,000 mobile services in a year is not the profile of a market under heavy price pressure, which is a fair input into how hard you should expect to negotiate.
Where NBN Co Moved the Line
Two changes, and one of them removes an obstacle that has irritated a lot of Australian businesses for years.
- The fibre-to-the-curb condition is gone. From July 2026, NBN Co removed the requirement to order a high-speed tier in order to be eligible to upgrade from FTTC to full fibre. Roughly 600,000 single dwelling premises become eligible without being made to buy a bigger plan than they want. If your site is FTTC, check eligibility now rather than assuming the old rule still applies.
- Wholesale speeds were lifted. Under Accelerate Great, eligible FTTP and HFC services moved: 100/20 and 100/40 to 500/50, 250/25 to 750/50, and 1000/50 to 1000/100. On the business side, Business 100 at 100/40 was uplifted to 250/100.
- Consumption keeps rising underneath both. Average monthly data per premises went from 460GB in June 2024 to 508GB in June 2025 — more than ten per cent in a year. That is the pressure driving the uplifts and it shows no sign of easing.
For voice specifically, only one of those numbers matters
Download speed is largely decorative where calls are concerned. Upload headroom determines what happens when eight people are on calls and someone starts a video meeting. The business tier moving from 40 to 100 upload is the meaningful line in that table, and a surprising number of businesses are entitled to an uplift nobody has told them about. Ask.
So Who Actually Pays?
The honest answer has three parts, and only the third one is the one people assume.
| Who | How much of it | How you would know |
|---|---|---|
| Shareholders, first | A meaningful share, absorbed through margin and cost programmes before anything reaches a customer | Falling income with rising profit is what absorption looks like on a results slide |
| Network investment, second | The part the industry warns about — build slows, or shifts to where returns are highest, which is not regional Australia | Coverage improvements that were promised and then quietly rescheduled. This is the cost that never appears as a price |
| Customers, third and last | Some of it, slowly, through published price adjustments over years rather than a step change | A rise applied across a published price list rather than an individual letter about your account |
The part worth holding onto
The second row is the real risk and it is invisible on an invoice. If higher input costs slow regional network investment, the cost to a business outside a capital city is not a higher bill — it is a coverage gap that does not close. That is a much more expensive outcome than a few dollars a seat, and it never shows up as a line item anywhere.
The Lag Nobody Prices In
Five steps, with realistic timing at each. This is the bit that gets left out.
| Step | What happens | When |
|---|---|---|
| 1 | Spectrum renewal price settled at $7.32 billion | Done, 2026 |
| 2 | Carrier capital plans adjust to a known future obligation | Immediately, and invisibly to you |
| 3 | Licence payments actually made as each licence expires | 2028 onward |
| 4 | Input costs feed into wholesale rates for resellers and access seekers | Slow, and partially absorbed at each layer |
| 5 | Retail pricing moves | Slowest and least predictable, driven more by competitive pressure than by cost |
Which makes one sales line worth recognising
“Lock in a longer term now, before prices go up.” Prices linked to a cost landing in 2028, transmitted through four intermediate steps, are not a reason to sign a five-year agreement in 2026. Your contract will come up for renewal at least once before any of it lands. In that trade the provider is not the party taking the risk, and it is worth saying so out loud when it comes up.
The Three Lines You Control
Having spent most of an article on things outside your control, here is the part that is not. Three items, all inside your own building, all bigger than anything above over a twelve-month horizon.
- Duplication. The largest single item in almost every audit. Conferencing, SMS tools, answering services, call recording, all bought separately from a platform that already includes them. Find it by auditing twelve months of card statements rather than the phone bill, and write the job next to each line rather than the product name — duplication is invisible in brand names and obvious in job descriptions.
- Seats nobody uses. Frequently ten to twenty per cent of a per-user bill: licences for departed staff, duplicate accounts, extensions on desks nobody sits at. Nobody audits this, and no provider has ever volunteered it. Count actual humans, then count licences, then compare the two numbers.
- The price review clause. Almost every agreement has one and almost nobody has read theirs. Find out what your provider may raise, by how much, and with what notice. That clause is where the money is genuinely decided, and it determines whether a long term is a good deal or a trap with a discount attached.
The test for any price rise you are handed
Ask one question: is this applied to every customer on published rates, or only to mine? A genuine input-cost increase appears in the published price list because it affects everyone and there is no reason to conceal it. An opportunistic one arrives as an individual letter citing industry conditions in general terms, and it is negotiable. That question settles most cases, and asking it politely costs nothing.
Four Signals Worth Watching
If you want to track whether any of this is turning into something real, these are the four things to watch over the next two years — and none requires a subscription.
Published price lists, not letters
When carriers move published business pricing rather than sending individual notices, cost pass-through has genuinely started. Until then it has not.
Regional build announcements
The industry warned that higher spectrum costs mean less investment. Watch whether promised regional coverage programmes proceed on schedule. This is the warning being tested in public.
Satellite direct-to-device
The ACMA is reallocating 2GHz spectrum for mobile satellite services with an auction expected later in 2026, and the ACCC is examining the role of low earth orbit services. Early, capacity-constrained, and the most likely thing to change the picture within five years.
The roaming inquiry
The ACCC inquiry launched 5 August 2026 runs roughly twelve months. If it recommends declaring domestic roaming, regional competition — and therefore regional pricing — changes materially. If it does not, the status quo hardens.
The short version
A $7.32 billion spectrum bill lands from 2028 and is being planned for now. Telstra had a solid year on falling revenue. NBN Co removed a real obstacle for about 600,000 premises. None of it changes your invoice this year, so treat any long contract sold on future price rises with appropriate suspicion. Then go and look at your duplication, your unused seats and your price review clause, because that is where the money actually is.
Related: what a business phone system costs for the pricing structure, the cost of a fragmented stack for driver one worked through, and full fibre upgrades for the NBN change in detail.
Frequently Asked Questions
What is the $7.32 billion Australian spectrum renewal?
Mobile spectrum in Australia is licensed rather than owned, and a substantial tranche of licences expires between 2028 and 2032. Rather than re-auction them, the ACMA set a renewal price: a preliminary figure of $7.34 billion was published in December 2025 and the final valuation settled at $7.32 billion. It covers seven bands — 700MHz, 850MHz, 1800MHz, 2GHz, 2.3GHz, 2.5GHz and 3.4GHz — which together carry 4G and 5G mobile plus fixed wireless broadband, underpinning more than 30 million mobile services. The four licence holders paying it are Telstra, Optus, TPG and NBN Co. Pricing was set per megahertz per head of population, band by band, with lower bands revised slightly down and mid-band prices revised up. Renewal applications opened on 18 June 2026 beginning with the 850MHz and 1800MHz bands, with nine-month application windows, and money moves as each licence expires from 2028 rather than now. The reason it matters today despite being paid later is that capital planning runs years ahead of cash: a carrier that knows it owes billions in 2028 makes different decisions in 2026 about what to build, where, and what to retire.
Will Australian business phone prices rise because of the spectrum bill?
Not this year, and the reasoning is about the chain rather than about optimism. There are five steps between the spectrum price and your invoice. The price is settled now. Carrier capital plans adjust immediately and invisibly. Actual licence payments occur from 2028 as each licence expires. Those input costs then feed slowly and partially into wholesale rates for resellers and access seekers. Finally retail pricing moves, and that last step is the slowest and least predictable because it is driven more by competitive pressure than by cost. The industry position deserves fair statement: Telstra argued the ACMA overvalued the spectrum by $3.3 billion and urged Treasury to cap it at $3.9 billion, and the Australian Telecommunications Alliance warned it would mean less investment, or higher prices, or both. The counter-argument is that spectrum is a scarce public resource and underpricing it transfers value to shareholders with no mechanism guaranteeing customers see the saving. Either way, a five-year agreement signed in 2026 to hedge a cost landing in 2028 puts the risk on you rather than on the provider, and your contract will renew before then anyway.
What did Telstra's FY26 results show and what does it mean for customers?
Telstra reported for the year ended 30 June 2026 on 13 August 2026. Total income was $23.405 billion, down 0.9%, while total profit rose 2.7% to $2.406 billion. Operating profit rose 1.7% to $4 billion, earnings per share rose 5.3% to 19.9 cents, cash earnings grew 11.6% to $2.9 billion with cash earnings per share up 13.8% to 25.5 cents, and EBITDAaL was $8.2 billion. A total dividend of 21.0 cents per share was declared including a 10.5 cent final dividend, alongside a $1 billion share buyback. Mobile service revenue grew 3.2% and more than 270,000 mobile services were added, all in the first year of the Connected Future 30 strategy. For customers the important caution is that a carrier result is a weak predictor of retail pricing. Income falling while profit rises is a cost-discipline story, not a pricing story, and in a market this concentrated the two are only loosely connected — a strong result is entirely compatible with prices going up, down or nowhere. The more useful signal is that adding 270,000 mobile services in a year is not the profile of a market under heavy price pressure.
What changed with NBN fibre upgrades and business speeds in 2026?
Two things. From July 2026 NBN Co removed the requirement to order a high-speed tier in order to become eligible to upgrade from fibre to the curb to fibre to the premises. Previously an FTTC premises had to commit to a high-speed plan to qualify, which deterred a lot of businesses because they were being asked to buy more speed than they needed in order to get better technology. NBN Co estimated roughly 600,000 single dwelling premises become eligible for full fibre without that condition. Separately, under the Accelerate Great programme wholesale speeds were lifted on eligible FTTP and HFC services: 100/20 and 100/40 uplifted to 500/50, 250/25 to 750/50, and 1000/50 to 1000/100, while Business 100 at 100/40 went to 250/100. Average monthly data per premises rose from 460GB in June 2024 to 508GB in June 2025, more than ten per cent in a year, which is the pressure behind all of it. For voice, download speed is largely decorative and upload headroom is what determines call quality when several people are on calls and someone starts a video meeting, so the business tier moving from 40 to 100 upload is the line that matters.
Who really absorbs higher telco input costs?
Three parties, and only the third is the one people assume. Shareholders absorb a meaningful share first, through margin and cost programmes, before anything reaches a customer — falling income with rising profit is exactly what that absorption looks like on a results slide. Network investment absorbs the second share, which is the part the industry warns about: build slows, or shifts toward where returns are highest, and that is not regional Australia. Customers come third and last, absorbing some of it slowly through published price adjustments over years rather than as a step change. The second of those three is the real risk to a business and it is completely invisible on an invoice. If higher input costs slow regional network investment, the cost to a business outside a capital city is not a higher bill but a coverage gap that never closes, which is far more expensive than a few dollars per seat and appears as a line item nowhere. That is also the claim to watch being tested in public over the next two years: whether promised regional coverage programmes proceed on schedule.
How do I tell a genuine price rise from an opportunistic one?
Ask a single question: is this being applied to every customer on published rates, or only to my account? A genuine input-cost increase shows up in the provider's published price list, because it affects everybody and there is no reason to conceal it. An opportunistic increase arrives as an individual letter about your account, cites industry conditions in general terms, and is negotiable. That one question resolves most cases and costs nothing to ask politely. Two habits reinforce it. Prefer providers with published pricing, because a public list constrains them and gives you a reference point at renewal — bespoke pricing feels like a win at signing and removes precisely the benchmark you need two years later when the discount quietly lapses. And read the price review clause in your existing agreement, which almost every contract has and almost nobody has looked at: establish what the provider is permitted to raise, by how much, and with what notice, then decide whether the term length still looks attractive. Be especially careful with long terms sold on fear of future price rises, since the party bearing the risk in that arrangement is you.
What should I actually do about telco costs right now?
Focus entirely on the three things inside your own building, because over a twelve-month horizon they are all larger than anything happening with spectrum or carrier results. First, find your duplication, which is the biggest item in almost every audit: conferencing, SMS tools, answering services and call recording bought separately from a platform that already includes them. Audit twelve months of card statements rather than the phone bill, and write the job beside each line instead of the product name, because duplication is invisible in brand names and obvious in job descriptions. Second, count the seats nobody uses, frequently ten to twenty per cent of a per-user bill in the form of licences for departed staff, duplicate accounts and extensions on desks nobody sits at — count actual humans, count licences, compare. Third, read your price review clause and establish what your provider may raise, by how much and with what notice, since that clause determines whether a long term is a good deal or a trap with a discount attached. Separately, if your site is on fibre to the curb, check your full fibre eligibility now that the high-speed tier condition has been removed.
