Moore's 1965 paper counted sixty components on the most complex chip then in production and forecast a tenfold increase each year for a decade; in 1975 he revised the growth to a doubling about every two years, which is the sentence the industry has been measured against since. Fifty years is twenty-five of those doublings — 50/2, the third link — and 2²⁵ is 3.3554432⋅10+7 (first link). The Intel 4004 of 1971 had 2 300 transistors, so twenty-five doublings of it is 2 300 × 2²⁵ = 7.71751936⋅10+10 (second link): seventy-seven billion, the company of the largest chips shipping now.
The law was never about physics, and the arithmetic shows why it must end. It is a statement about the cheapest number of components per function, which holds while each doubling makes everything smaller and nothing more expensive; it stops holding when the cost per transistor stops falling, which is roughly where it stopped holding. Two other numbers explain the rest of the modern picture. One is 2^10 = 1024 (fourth link), the approximation that lets a billion transistors be called a thousand million and a megabyte be 1 048 576 bytes. The other is the fact that the doubling has moved: after around 2005 the clock frequency stopped rising because the power density could not take it, so the extra transistors went into cores, caches and memory stacks instead of into faster single threads. The count still doubles; the speed of one program does not.
Answer
3.3554432⋅10+7
How to read it
2^25 on the workbench, twenty-five doublings. The second link multiplies it by the 4004's 2 300 transistors; the third is the fifty years in doublings.
Source
Gordon Moore, 'Cramming More Components onto Integrated Circuits', Electronics, 19 April 1965, revised in 1975 to a doubling about every two years; the Intel 4004 of 1971 contained 2 300 transistors
Work it in the calculator
These links open the right calculator with this problem's numbers already typed in. The line above is what the screen should show — check it yourself.