The problems to be solved behind the concrete and glass facade are not merely academic. Glaxo is heavily dependent on Zantac, some analysts say too heavily. The world's best-selling drug makes up 40 per cent of the company's revenues, and it is already beginning to slip, six years before its patents start to run out. The company aims to triple the number of replacement drugs it brings to market while its revenues dwindle and research costs (£900m this year) inflate.
Sir Richard has taken two steps towards solving Glaxo's dilemma, starting with the high-profile £9.2bn bid for Wellcome. Combined, the two companies would have more weight in what has become a buyers' market. Savings from cutting duplicated overheads will also help to fund more research. And Glaxo will be able to capitalise on any likely drug candidates now in Wellcome's experimental pipeline.
The smaller, but potentially more significant step is the agreed takeover of Affymax in Palo Alto, California, for up to $550m (£348m). "Affymax is about the future,'' Sir Richard said at the weekend. "There has to be a constant flow of new medicines. Wecan't do that with traditional techniques that produce one new molecule a week.'' Until recently, pharmaceutical companies discovered new drugs, usually by sifting through naturally occurring organic chemicals. Now they want to create medicines from scratch. The first step is to make a molecule that will bind with its target, be it a cell membrane, an enzyme or the shell of a virus. Sometimes only one in a million will do this, so lots of molecules have to be built up quickly. Affymax does just that.
The US company was founded in 1989 by Dr Al Zaffaroni, the father of the oral contraceptive, who is now in his seventies. It is regarded as the leading player in a new field called combinatorial chemistry. One of its key techniques, borrowed from the computer industry, is a method of building hundreds of thousands of unique molecules from scratch on a silicon chip in just a few hours. These can then be quickly tested to see which ones will bind with the desired target.
Once a molecule that works is found by this method, thousands of variations can be tested to see which work best.
A handful will then be sent to Stevenage, Hertfordshire, for testing. The result, argues Sir Richard, will be drugs that are more accurate, more therapeutic and have fewer side-effects.