One company that embodies this heady mix of risk and potentially large rewards is Aromascan, which produces technology to scan smells in digital form.
Capitalised at pounds 41m, the group makes no secret of the fact that it will be showing significant losses (pounds 1.5m plus) when it reports later in June against the house stockbroker's earlier hoped-for pounds 900,000 loss. Although the balance sheet is healthy, with more than pounds 4m cash, a further fundraising will be required later in the year.
Against that background there may be little rush to buy the shares - they have recently fallen away from a peak of over 180p to 155p.
There is also a slightly comic air to a company that has invented a machine that smells - an "electronic nose" would probably be more at home in a fairground.
But investors will take the company more seriously when they understand it better. Sensors of all kinds are playing an important role in many fields including industrial processing. Other companies producing equipment based on sensors, such as Renishaw with its touch probes and Druck with its pressure sensors, are highly regarded by investors. Sensors in smell, measuring airborne chemicals, could have wide application. Aromascan is talking to a string of well-known companies in industrial controls about possible uses for its polymer-based sensing equipment. As the year unfolds there are likely to be more link-ups.
One deal announced last month is a strategic alliance with Japanese giant, Mitsubishi. The shares have weakened since, partly because Aromascan is reluctant to talk freely about what it could all mean.
Mitsubishi first began talking to Aromascan in January, 1995. It set up four pilot schemes to test the technology in the field of industrial process controls. One example related to rice. Japanese rice sells for three times the price of imports, making substitution tempting despite tough domestic controls. Aromascan's equipment offers a potential deterrent since the only thing it could not do, according to the Japanese, was identify the village from which the Japanese rice came.
No big deal, you may think, but the implications for industrial processes are enormous. One idea is that Aromascan's sensors will be incorporated in the production line and linked with neural network software that can learn from patterns of activity and deal with problems while the line is running. For example, sensors linked to a food production line might sense that tea, say, is being tainted.
Allied to other sensors based on vision and touch we have the basis of the intelligent factory of the future. Indeed Aromascan's managing director, Allan Syms, says that on the basis of pilot studies giving encouraging results, Mitsubishi believes in the growth potential in linking all types of industrial sensors, including the electronic nose.
The deal with the Japanese firm is structured in an attractive way for Aromascan. Instead of merely selling its sensors or making a royalty, it will have exclusive rights in Europe to products developed by the joint venture and share the rights with Mitsubishi in the rest of the world except Japan. Even in Japan the group will generate revenue from the sale of its sensors to Mitsubishi. Dr Syms talks of a pounds 1bn global market with potential profits of pounds 150m on normal margins.
When Aromascan was first floated, it targeted the much smaller laboratory equipment market and raised funding appropriate to those goals. Now, additional funding will be required - perhaps another pounds 10m. But shareholders should be happy to oblige, given the huge size of the markets being addressed and the relationships the company is developing.
The group is also moving from DC to AC technology. Put simply, DC means the equipment needs 37 sensors to identify aromas to the required degree of accuracy. AC makes it possible to achieve many of the same results with one sensor making miniaturisation possible and helping in the development of handheld equipment.
Initial products from Aromascan's collaboration with Mitsubishi should hit the market in the summer of 1997. Before then other such strategic alliances are likely. Aromascan will concentrate on its patented skills in using polymer technology to identify aromas and in developing the associated software while its multinational partners use their skills and financial and marketing muscle to develop products for specific applications and introduce them into world markets.
The shares look a worthwhile gamble with potential that, by the standards of many blue sky ventures, is comparatively modestly priced.