I believe there are a couple manufacturing industry likely staying in China for a longer time - semi-conductor and high-end electronic manufacturing. Some others may just go elsewhere - labor intensive manufacturing. There are at least 2 reasons to support it.
1) Labor cost has been going up in China because of various reason, primarily rising cost of living.
2) Because of 1 child policy for last couple decades, China is actually an aging nation. Although it is still young compared to a lot of developed countries, the median age of China is definitely older than that of India.
@chanj0: Does it mean the manufacturing region is shifting from China to India? If it does, how fast will it go?
That depends on what kind of manufacturing you are talking about.
The manufacturing that moved to China was labor-intensive assembly line work, and it did so in search of lower costs.
China is in the process of boot-strapping from an agrarian third-world economy to an industrialized first world one. It's doing so the same way others did it: move the peasants off the farm to the cities, to become the basis of an industrial workforce.
China was able to achieve rapid growth as a manufacturing power because it could offer lower costs. Chinese factory workers got paid a fraction of what a comparable worker might get elsewhere, but everything is relative: it costs far less to live in China. For those peasants off the farm, the factory jobs were a step up. They had better hours, better working conditions, and paid better than being a peasant on the farm, and China found itself with infrastructure and urban development issues as peasants flocked to the cities to get those jobs.
But the pool of cheap labor is drying up, in part due to demographics and China's "One child per family" policy to reduce the population. Manufacturers are having to compete for workers, with wages increasing as a result, and China is no longer the low cost producer in that sort of manufacture. (There was a report in EETimes a while back about a major Chinese manufacturer who announced a full-scale push into robotics as a consequence.)
Vietnam is one place eyed as a possible new manufacturing center. The Philipines are another. India might well be a third, as the fundamental requirement is a pool of low cost labor for whom factory work will be an improvement over what they have. In fact, I can see Chinese manufacturers looking at outsourcing to India, funding and operating the plants. They already have the experience in dealing with corrupt and inefficient bureaucracies where the first requirement is knowing what palms to grease.
As you move up the value chain, things get more complicated. Semi-conductor fabs are highly sophisticated and capital intensive operations, and you have the question of whether the supporting infrastructure is there to support them once built. Is there transport to get raw materials in and goods out? Is there water? If there power? Is there sanitation? Is there a local work force who can operate the fab? Are there local suppliers who can provide what thee fab and its workforce need on a day to day basis? (Indian universities are certainly producing skilled engineers, but that doesn't mean any are where the fabs will be, or will want to relocate there.)
I look at India, and I see a number of formerly independant states that have become a patchwork quilt of a nation, and I sometimes wonder if all might not be better off if India dissolved into its component parts again. Certainly how well new enterprises do is strongly influenced by where they are. Some regions are simply better at supporting new development than others. If you could take the need to deal with the Indian national government and bureaucracy out of the equation, and developers could just deal with the regional authorities, all sorts of things might be possible.
Call me Cynical... I do not believe this will ever happen... What somebody should do is checkout where the JAYPEE group has land that is yet under-developed... Then you can make a bet that they would announce the fab around there... JAYPEE along with the politicians will unload their lands etc at astronomical prices , make the money and run... Checkout Fab city in Hyderabad... If you take a look at who made money - it is the politicians and the few cronies... Who holds all the land near the fab city now - mostly NRI's ( Non resident indians or No return on Investment). The JAYPEE group has one of the highest debt to equity ratios among India Inc..and they have been trying to unload all their cement factorues etc to pay down the debt. What makes somebody think that they would carry thru on a fab plant, that will run into about 200-300 million loses the first few years before ever turning a profit......pipe dream....My 2 cents.....
@GSMD, thanks for sharing this history. You offer fasicnating perspectives on the industrial development in Inida -- with so many twists and turns, then, unintended consequences. We shall see how two fabs in India will change India's electronics landscape.
I was talking to one of the the guys from the India Electronics and Semiconductor association, the trade body that represents the chip and electronics industry. He was saying that 40% of the worldwide production was in the 180 and 130 nm nodes. If that is indeed the case, there should be enough opportunities for these fabs.
Thanks Rick for bringing that saga of the Indian technology history. It was a master stroke by the govt. which opened up the industry. It actually did not mandate India made systems.
The order simply said all servers should be POSIX and UNIX SVR3 compliant. This turned out rather well for the Indian server industry (yes, there used to be one ! And world class too.) since UNIX was not that popular in the mid 80s. There were very few US made UNIX boxes in those days and even those OEMs who had UNIX boxes preferred to push their prop. OS boxes.
I was in the middle of this transition since I was a rookie engineer doing performance engineering on HCL's quad CPU SVR3 boxes (SMP). We built our own buses (VME for I/O and a futurebus derivativefor cache coherent CPU interconnect ) and of course all HW was custom made. Processors were 68020/30/40. We also made an ethernet cluster with four of these boxes for a 16 CPU design complete with auto-parallelising Fortran compiler. A local lab wanted this since the US of A refused to supply it Crays !
Wipro a major competitor eschewed SMP and went for an i386 based shared nothing architecture. A few others like ORG also made local HW.
A whole bunch of local ISVs mushroomed supplying all kinds of enterprise SW on these boxes.
The server industry blew its lead since it could not really step up to selling these boxes outside Inida.Design was great but quality was shoddy, support could not scale up and do not even ask about documentation. HCL actually came to the US not to sell its services but to sell these boxes. Once that was deemed a futile attempt, the R&D was completely shutdown and the engineers deployed as tech guns for hire. the supply of the battle hardened engineers finally dried up (no R&D to churn them out) the industry fell into the pattern that you see today.
But the primary reasons for the demise of the industry was commercial. It was more lucrative being a distributor than an OEM.
The Indian SW industry that you see today actually had its genesis in those ISVs who had to close shop because the Indian server industry went belly up. Having no local market they reinvented themselves as service companies.
A sad tale of commercial greed killing a vibrant technology industry. Nobody beleives me now if I say that 1988/89 I worked on a UNIX box more scalable than a Sun box - all designed out of India !
And may the unsung govt officer who in a fit of absentminded genius madated SVR3 compliance be nominated to the technology hall of fame ! He of course had no clue as to what he was doing.
I think that the independent development of hardware ( and Software incl. OS & Appl. that went with it ) put India in a position to specialize and boom in Software after the home grown hardware part became unviable. Contrary to the expectations of uninformed "experts", no simple-minded linearity there !
Blog Doing Math in FPGAs Tom Burke 3 comments For a recent project, I explored doing "real" (that is, non-integer) math on a Spartan 3 FPGA. FPGAs, by their nature, do integer math. That is, there's no floating-point ...