A wireless start-up claiming radical results has achieved a $1.66 billion valuation on the London Stock Exchange, amid mounting scepticism from the wireless world. Eighteen months from its first showing, we are still waiting for real proof that it works, either in the form of products or open demonstrations - and criticism of the company is starting to mount.

xG Technologies claims it can send high speed data using very low power and almost no spectrum, using a technique called "xMax". Eighteen months ago, we visited XT Technologies in Florida saw a demonstration of xMax. The demonstration was inconclusive,

The company explains its technology in various ways, but the one we find easiest to grasp is the idea of "single cycle modulation". As xG's site puts it, "Single cycle modulation is implemented when individual sinusoidal cycles of RF energy are modulated to represent one or more bits of data. This proprietary modulation method differs from conventional approaches where tens to hundreds of thousands of RF cycles are required to convey a single bit of information."

This sounds clear enough. Unfortunately, according to one critic of xG, Qualcomm's Phil Karn, such a statement could only have been written by someone who "doesn't have a clue about digital radio communications."

Any new player can expect scepticism from established vendors, of course. xG has promised to deliver its technology to service providers, in the form of handsets and base stations that will deliver wireless voice so cheaply that anyone can set up as a service provider.

It has floated on the London Stock Exchange, where it currently has a valuation of around $1.66 billion.

The actual delivery of product has been slow however. It is promising a wireless voice service, which so far has one customer and one joint-venture partner.

Does the technology add up?

xG's most obvious weakness is its refusal to give full explanations or public, repeatable demonstrations of its technology. The inventor Joe Bobier hasn't published peer-reviewed papers, though he has had backing from Professor Stuart Schwartz of Princeton University.

Phil Karn, a Qualcomm employee and long-time radio expert, argues that xG's explanation of its technology is a simple contravention of Shannon's Law. a mathematically proven limit for any signal channel. Put simply, there needs to be a certain energy to carry the signal, and at these frequencies, one cycle per bit doesn't provide enough energy.

Earlier descriptions of xMax didn't appear to have this fundamental objection - or at least we thought not - because they described a system where the energy is carried in broad, low-power sidebands.

xG has provided some more detail, in a US patent no 7003047, but Karn's analysis is that this describes an existing technique called continuous phase frequency shift keying (CPFSK)

Karn is scathing about Bobier's apparent misunderstandings of frequency spread in signals. We're not qualified to judge between the two - and it's worth pointing out that Karn's employer, Qualcomm would have a huge interest in any big change to our understanding of radio communications.

Nevertheless, if you look at xG's output, you have to wonder. "It just occurred to me that the least amount of radio energy I could have is one photon — one wavelength — and the least amount of data I could have is one bit,” says Bobier, in an MRT article. Most of us should spot pretty easily, that one wavelength is not the same as one photon.

Why wireless VoIP?

While the technology has scientists scratching their heads, the roll-out has business-people wondering. For something this revolutionary, xG has only managed to convince small players to take low risks, and it has done so in a surprising field.

Why is xG offering a mobile VoIP service? It's jumping into one of the most competitive markets of all, where its new RF technology is only a fraction of the puzzle.

"There might be something clever in the air interface bit (although most people I've spoken to have dismissed it as snake-oil)," says analyst Dean Bubley of Disruptive Analysis. "But even so, there was no way that a small RF company was going to be able to put together the whole value-chain including handsets, software, backhaul, authentication, application layer, network security and all the other 2000 "moving parts" that make up a working system."

xG has a handset, the TX100, a dual-mode Wi-Fi/xMAx handset promised by Swedish Wi-Fi handset maker Lund IP products. The handset and base station hasn't been produced at a public show, though it says it has shipped commercial base stations to a neighbouring service provider, Far Reach. "I realise all startups have to begin somewhere, but this strikes me as a pretty odd choice of reference site if your ultimate aim is to sell to the world's major service providers," says Bubley, who has been unable to trace Far Reach's claimed backers - a group called the "Phoenix Foundation"

xG also has an agreement for Telefonica Mexico to trial the technology, with a view to a potential joint venture.

Is there anything there?

Investors in xMax have taken a pretty big risk - but they knew that already. The rest of us are, as ever, keeping a watching brief on this kind of thing, in case there is some new technology out there ready to undermine and transform our way of wireless life.

In this case xG's secrecy, its choice of market, the lack of verifiable demonstrations, and the absence of testable products all add up to cause for doubt.