Decades-old market research cooperations are breaking down. A few of the big guys don't want to play nice any more, perhaps because they think things have come to the point where they are just giving their rivals a helping hand. Peter Clarke says they mess with the system at everyone's peril.
What's going on?
I feel like the nerdy schoolkid who has just had one of the lenses in his spectacles broken. I can still see but everthing is a bit blurred and fractured.
First it was the dropping of the chip market's book-to-bill ratio for the global chip industry. OK that was some time ago in January 1997. But it still represented the passing of useful market indicator that shows whether the market was going up, down or was on the turn.
But more recently has been the withdrawal of foundries Taiwan Semiconductor Manufacturing Co. Ltd. and United Microelectronics Corp. from the SICAS project (Semiconductor International Capacity Statistics) to measure semiconductor wafer production capacity. And now it is the similar withdrawal of Advanced Micro Devices Inc. and Intel Corp. from the World Semiconductor Trade Statistics (WSTS) organization.
The latest we have heard is that the Semiconductor Industry Association has pulled the plug on the whole SICAS report.
Both the SICAS and WSTS programs rely on being able to capture and aggregate data from multiple companies to give the larger market picture. In the case of WSTS the raw numbers are then multiplied up to give a global figure based on a factor that is cross-correlated with other market research data.
Some people think that the semiconductor industry is maturing and becoming more like the automotive industry – a market dominated by a few large players and showing single-digit annual percentage growth. Others think there is plenty of life in a youthful chip business that has just been going through some growing pains recently.
One thing's for sure is that the chip business continues to change.
This is a very well argued opinion piece. If anyone on our staff who has followed WSTS, SIA numbers closely over years, it's Peter.
We heard from some of you -- but silence on this topic from the industry is deafening.
I don't think Peter is alone who feels like his own glasses being smashed.
Well said, Peter. But I think it's important to re-state that Intel left the WSTS program only after AMD did. With AMD out, Intel almost had to follow suit. Otherwise, the actual MPU data presented would be just Intel, and Intel may well be understandably uncomfortable putting out this much naked data.
I do hope that both companies will reconsider their position on participation. Otherwise, I agree that it's a real shame.
Are these services are still useful to any companies to make future decisions or investing in right technologies. I believe the technology is matured to a level where there is not much help due to sharing of information.
Information is power and companies have every right to withhold it if they can gain a competitive advantage as a result. Of course, it's not good news for us, but times are tough and I can understand the temptation. I guess some form of information sharing will re-emerge at some stage in the future when the market stabilises.
AMD & Intel pull data?
AMD losing share blowing antitrust case. Knows Intel market rigging and model for determining supply into future time.
Intel’s gaining share. Does not want to attract attention surpassing 81% including embedded.
Last time I observed such silencing 1989. Computer Supplier’s ad spend went dark in CRN. Right before Intel Inside morphed into metered price discrimination.
Intel expansion toward foundry seems another reason likely urgent.
Discontinuous memory innovation is a concern. Micron is a more efficient fabricator. Yet this leading memory, materials, fabrication aspect at inflection of physics goes beyond Micron.
Some picked up WSTS to project Intel revenue; since 1997. SEC is investigating use of analytics.
Intel supply data is used for projecting procurement of product routes for margin values into future time. So at inflection point in physics why doesn’t Intel want to report data?
Masking data makes Intel hard to see. Except for those who own a pair of special glasses which is the quantitative model. Not having WSTS data will not stop savvy QUANTA players on their investment in Intel analytics.
Intel foundry targets a physical space others
have been competing at for a long time. For Intel to leap frog on 20 years monopolization speaks poorly for industrial social values and democratic capitalism.
I'm for Intel expanding their business. So long as Intel does not anoint, step on, infringe, limit, restrain, shift revenue among those who invest organically on best practices to compete there.
So the big challenge is not AMD and Intel report to WSTS, but defining how to govern, regulate command, control, monitor INTEL as process saturates to new competitive potentials.
Intel strategy is to push through molecular at process saturation so they own quantum on long time monopoly gaming. That can't happen. It would destroy what's meant to come naturally.
To @Mike Bruzzone, what is wrong with Intel "Intel strategy is to push through molecular at process saturation so they own quantum on long time monopoly gaming"? Many start-up and large companies are looking at quantum computing, if I had a great idea in this space I would had started my own company and hope to be acquired by Intel one day(Unfortunately I don't)...this seems natural to me, Kris
P.s Quantum computing is probably 20-30 years from being commercially useful
Kris - Intel took lead from IBM @ .35, pushed through nodes since & took on 300 mm. So what’s wrong with Intel taking major R&D burden? PC MPU where I hail Intel also takes majority profit. By artificially throttling supply into broker mkt relieving foundry from surplus cost greater than real time PC demand. Intel produces roughly 2x MPU demand & 1/3 are priced less than Intel cost. This strategy takes development, sales & jobs from my prior employer’s; Cyrix, NexGen, AMD, IDT, ARM & licensees in future? This is one way Intel monopolizes. Lacking this tactic Moore’s axiom is 4 years. Industry w/longer cycles; all have opportunity to make $ today to pay for R&D tomorrow. Reason Moore’s law sustains 18-24 month cycles is surplus production relieved from inventory on tactics of structured market rigging. Otherwise Moore’s axiom is not economically achievable. Concentration effects with w/Moore have devastated domestic semi & compute industries. On Rocks Law process economic opportunity of trickle down from Intel RSIK development is a myth.
Too respond to query. Ok, you’re a band of developers & hit it big in molecular moving to quantum; rare but possible. Outside new materials for memories where there are start ups, most are research institutions or large enterprise labs. In these environments IP is everything and Intel will steal it. There are numerous examples; starts up, mid caps, universities. In this environment never hire from Intel; relations remain. Recruiters will fill you up. BOD issues too. For BizDev & mkting, media channels, ad & PR agents are IP transfer bridges. They’re paid by Intel in business opportunities. Have people you trust. For large design houses working with Intel there is an exchange system; trade your leading IP for Intel sustaining your current business. Many have played this game and most have lost. If you can commercialize perhaps Intel will buy. Most Intel purchases are beyond design expertise. Good luck.