News
As the resident NBN and home-internet expert for TechRadar Australia, I spend much of my time looking out for the best internet deals, and there are almost invariably a number of compelling options at every different speed tier — and these can be a great way to sense-check if you’re overpaying for your current NBN service.
When choosing deals, much of my decision-making revolves around two easily identifiable metrics: price and typical evening speed (TES). Typical evening speed is something all internet service providers (ISPs) have to disclose before you sign up for a plan. In essence, it tells you the download speed you can realistically expect to achieve during the busier evening hours of 7pm to 11pm. Being able to view the TES also means you can easily identify which ISPs ultimately offer better value — i.e. the ones advertising fast evening speeds at competitively low prices.
With almost all of the best NBN 500 plans promising to deliver their maximum 500Mbps speeds even during the peak evening period, it’s a tier where direct price comparisons make sense. However, things get a bit wobblier when you’re shopping for faster speeds, such as NBN 750 and NBN 1000.
Most plans on these tiers make more conservative promises when it comes to typical evening speed — for example, rather than the 1,000Mbps you might expect from NBN 1000, on these gigabit plans, most providers only promise evening speeds in the 820-860Mbps range. With that TES figure closer to 750Mbps than 1,000Mbps, it prompted me to ask a simple question: Is NBN 1000 actually worth it, or should you save yourself some moolah every month and opt for NBN 750 instead?
Cue a whole load of number crunching!
So averageFirst, I calculated the current average monthly cost of NBN 750 and NBN 1000 plans, alongside the average TES advertised by the providers I monitor. After crunching the numbers on NBN 750 plans from 14 providers and NBN 1000 plans from 20, the final results were:
Measure
NBN 750
NBN 1000
Average price
AU$105.98
AU$114.95
Average TES
730.2 Mbps
835.4 Mbps
As expected, the results show that NBN 1000 plans do not, on average, get as close to their nominal 1,000 Mbps tier speed as NBN 750 plans get to 750 Mbps. The gigabit gap was quite substantial, falling 164.6Mbps shy of the tier’s nominal 1,000 Mbps tier, compared to a shortfall of just 19.8 Mbps for NBN 750.
With the average monthly cost less than AU$10 difference between the two tiers — specifically AU$8.97, or 8.5% — the question then became whether the NBN 1000’s typical evening speeds are worth it from a value perspective. For that, we of course needed to do some more maths:
Measure
NBN 750
NBN 1000
Shortfall from theoretical maximum
19.8 Mbps
164.6 Mbps
TES as share of maximum
97.36%
83.54%
Cost per advertised TES Mbps
14.5 cents
13.8 cents
Next stepsWith NBN 1000 plans falling far short of 1,000Mbps in their typical evening speed promises (averaging just 83.5%), intuitively they seemed like they’d come out second fiddle to NBN 750. Surprisingly, however, NBN 1000 is slightly ahead when comparing cost per megabit for typical evening speeds: 13.8 cents for NBN 1000 versus 14.5 cents for NBN 750. There are a couple of NBN 750 plans that come in below that average at 13 cents per megabit — Kogan Internet and Spintel — which I've listed below.
In other words, NBN 1000 might cost more in absolute terms, but it does offer slightly better value per advertised megabit. If you need some recommendations as to which NBN 1000 plans are worth checking out, look no further than Dodo and Superloop. The cost per megabit on these two works out to be 13 cents, placing them among the cheapest for the tier.
Spintel Home Turbo Fast | 750Mbps | AU$69p/m (first 6 months, then AU$94.95p/m)
If it's the most affordable NBN 750 plan you seek, you simply can't ignore this option from Spintel. It has gone slightly lower in price before, yet it continues to be the cheapest I can currently find in the first year. Spintel is also only one of five providers on this tier to claim maximum plan speeds.
• AU$69 total minimum cost
• AU$983.70 first year cost
• AU$1,139.40 ongoing yearly costView Deal
Kogan Internet Platinum | 740Mbps TES | AU$84.90p/m (first 12 months, then AU$94.90p/m)
The introductory cost of this Kogan Internet plan is evidently much higher than Spintel's, but the telco is currently offering a 12-month discount that results in this being the second-cheapest in the first year. Plus, with the ongoing cost being practically the same as Spintel's, you'll be paying the same 13 cents per megabit.
• AU$84.90 total minimum cost
• AU$1,018.80 first year cost
• AU$1,138.80 ongoing yearly costView Deal
Superloop | 860Mbps TES | AU$84p/m (first 6 months, then AU$114p/m)
Superloop may have increased the price of this gigabit plan as of July 1, but it continues to represent great value. Not only is the TES claim among the highest for the tier, but the telco receives overwhelmingly positive reviews from customers.
• AU$84 minimum cost
• AU$1,188 first year cost
• AU$1,368 ongoing yearly costView Deal
Dodo nbn1000 | 850Mbps TES | AU$77.99p/m (first 6 months, then AU$109.99p/m)
If you want to try NBN 1000 for as little money as possible, this plan from Dodo is tough to beat. You'll pay just AU$77.99p/m for the first six months, which represents the most affordable intro cost for this tier. What's more, with an 850Mbps TES claim, you'll pay the same 13 cents per megabit as the Superloop plan above.
• AU$77.99 minimum cost
• AU$1,127.88 first year cost
• AU$1,319.88 ongoing yearly costView Deal
The NBN 750 and 1000 plans don’t exist in a vacuum, however, and NBN Co has been pretty clear that it considers NBN 500 — introduced last September — to be the current sweet spot for most Australian households. So how does the golden child stack up, you ask? I’ve also made it no secret that I consider it the best option for most Australian households, but admittedly, that’s been without these recent calculations.
As we alluded to above, unlike NBN 750 and 1000 plans, most providers do advertise maximum plan speeds for NBN 500, with an average typical evening speed of 499.3 Mbps across 22 providers — equivalent to 99.9% of its potential 500 Mbps maximum.
On a cost-per-megabit basis, however, things don’t look as pretty for my personal fave. With an average monthly price of AU$95.75, the average megabit cost works out to around 19.2 cents, making it substantially more expensive than both NBN 750 and NBN 1000.
With that being said, it still remains the cheapest option in outright monthly terms, with the average NBN 500 plan costing AU$95.75 — AU$10.23 less than NBN 750 and AU$19.20 less than NBN 1000.
As for my top recommendations for an NBN 500 plan, look no further.
Exetel One | 500Mbps TES | AU$80p/m
This Exetel plan has been my favourite ever since it launched in July last year. It offers fuss-free billing, claims maximum plan speeds — the telco has even been found to overdeliver on speeds in ACCC reporting — and it comes with extra perks, giving you the to boost your speed even further for AU$1 a day. The cost per megabit is also the cheapest at 16 cents. It's simply the best.
• AU$80 minimum cost
• AU$960 yearly costView Deal
Kogan Internet Gold Plus | 500Mbps | AU$69.90p/m (first 12 months, then AU$85.90p/m)
If you want to pay as little as possible for an NBN 500 plan in the first year of service, then Kogan Internet's plan is the one for you, thanks to a generous 12-month discount. It also comes in at 17 cents per megabit, well under the current average. Plus, the telco receives generally favourable reviews from customers.
• AU$69.90 minimum cost
• AU$838.80 first year cost
• AU$1,030.80 ongoing yearly costView Deal
So, where to go from here? What these numbers helpfully confirm is that if you do opt for a faster plan, you are getting good value from a purely cost-per-megabit metric — so if you were looking to justify jumping up to NBN 750 or NBN 1000, you can probably stop reading here.
These facts and figures can’t take into account all the subjective-but-important real-world factors that come into play when choosing an internet plan — they can’t, for example, help you determine how much speed your household actually needs.
While it’s now been proven that NBN 1000 plans offer the cheapest cost-per-megabit, that’s only really useful if you actually need gigabit speeds regularly. What you’re ultimately paying extra for, most of the time, is time — specifically, less time waiting for large files (like video games or OS updates) to download.
And with NBN Co recently claiming that average monthly data usage increased to 574 GB per premises — up 13% year-on-year — having a plan that will download all those gigs faster is going to become more and more essential.
For me personally, I’ve decided I can stand to wait a little longer on the one or two occasions a month when that extra speed would be noticeable. It’s hard to justify paying more if the extra megabits are going unused 95% or more of the time.
Across the UK, "will AI take my job?" has become one of the nation's most searched fears, a concern echoed in a recent open letter signed by more than 200 economists and technology leaders, including former Google CEO Eric Schmidt.
The letter warned that without the right guardrails, increasingly powerful AI could drive large-scale job displacement.
This anxiety around job stability has, unsurprisingly, increased in line with some of the world's largest employers - including Microsoft and Meta - announcing cuts to thousands of roles over the last year.
Left to interpret what that means for their own futures, employees globally have become accustomed to fear and uncertainty.
The impact is far-reaching, with recent research indicating that widespread anxiety is now impacting career decisions. In fact, as many as one in three workers say they are planning to retire earlier than planned, while 25% of workers are considering moving into industries less exposed to AI.
Although concerning, this potential “white-collar exodus” isn't a rejection of AI itself; most workers say the technology makes them productive when introduced with proper training and clear goals. However, it’s a response to how AI is being rolled out: often with little clarity, minimal support and not enough acknowledgement of the potential impact on employees’ confidence in their own expertise.
Left unaddressed, the impact will be felt in retention, engagement and institutional knowledge. Designing and prioritizing ways of working that incorporate transparency and training, while prioritizing job satisfaction and performance, has become a board-level priority.
The cost of the exodusMany workers feel that AI could reduce the need for their role within a few years, and that fear is not misplaced. Finance, professional services, and IT jobs have already felt this pressure, with PwC cutting 200 entry-level roles this year.
But we must not fall into the trap of reading this as solely a graduate or entry-level problem. In reality, according to a Gartner survey of C-level executives, 56% are extremely likely to quit due to the impact of AI and burnout. This is in some ways more harmful, presenting a clear risk of knowledge attrition across organizations as expertise born of human experience is lost.
Job security writes only the first chapter. Knowledge workers are increasingly frustrated that AI allows tasks once requiring years of expertise to be completed by almost anyone. The concern is about whether experience, judgement and specialist knowledge still carry value. And when organizations fail to redefine what human expertise contributes, disengagement festers.
Compounding this is the growing issue of AI fatigue. More than half of all content online is now AI-generated. So workers clocking off from AI at work are simply wading into more of the same at home, extending far beyond the 9-5 that businesses can measure or manage. Harvard Business Review has linked AI fatigue to decision fatigue, errors, and higher intent to quit.
In response, employees are actively cutting back their use of AI tools. However, investment shows little sign of slowing. Enterprise AI spending reached £1.2 million last year, up 108% on the year before, highlighting the growing gap between organizational investment and employee adoption and engagement.
This combined effect results in a workforce that's disengaging, under-using costly tools and potentially planning their exit. This demands urgent action from businesses, who need to address how AI is implemented and communicated, and act before disengagement turns into departure.
The path forward: Building trust through transparencyPositively, some organizations are already demonstrating what successful AI adoption looks like, grounded in an understanding that communication and culture should be taken as seriously as the technology itself.
IBM, for example, is both expanding and redefining entry-level hiring for an AI-first workplace, detailing where the organization is headed. Since workers often don't understand why they're expected to use AI in their role, moves like IBM’s can help to close a critical communication gap.
Transparency, clear guardrails and genuine dialogue are what turn AI spend into engagement. When people understand where they fit into the future of work, they're far more likely to embrace the technology shaping it.
Reducing digital noiseThe instinct in most organizations is to let AI creep into everything at once. In practice, AI delivers the most value - and does the least damage to worker purpose and confidence - when it targets administrative burden and repetitive tasks.
AI tools shouldn’t undermine the judgement, experience and specialist knowledge that people bring to their roles. Careless implementation of AI is the fastest route to the fear of expertise erosion that drives people toward the exit.
Investing in reskilling and upskillingAdaptavist’s recent research into the human cost of AI reveals that 74% of workers are already taking the initiative to build AI skills, showing that willingness to adapt isn't the problem. But organizations must match that commitment. Rather than treating AI training as a one-off exercise during rollout, businesses should implement continuous learning and support that evolves alongside the technology.
The value of this approach is already being recognized widely. The UK Government has introduced its AI skills program with the ambition of training 10 million workers, backed by partners including the NHS. The recent “skills compact” plan also puts the onus on employers to address skills gaps in AI across financial services.
Both initiatives reflect a growing understanding that successful AI adoption depends as much on investing in people as it does on technology and tools. Effective training should help employees understand how their expertise continues to add value, giving them confidence that AI is there to enhance their role, not to replace it.
Building psychological safety around experimentationAn important reality is that none of the above works if people are afraid to use the tools they're being trained on. Infosys has acknowledged the importance of psychological safety by framing AI adoption as an opportunity to experiment and improve without fear that mistakes will damage confidence or career prospects.
A study published in Occupational Medicine found employees who had 1:1s five times or more within two months reported significantly higher psychological safety than those with fewer meetings. Direct conversations to understand concerns, discuss how AI fits into their career development, and address uncertainty also show a marked improvement.
None of these approaches requires organizations to pare back their AI ambitions. Ultimately, the businesses that retain their best employees will be those that treat AI anxiety as a change management problem, measured by whether their most experienced people still feel essential to the work.
We've listed the best job sites.
This article was produced as part of TechRadar Pro Perspectives, our channel to feature the best and brightest minds in the technology industry today.
The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/pro/perspectives-how-to-submit