09/03/2024 | News release | Distributed by Public on 09/03/2024 08:32
Model N hosted its annual Rainmaker conference in Austin, TX this past June 2024. The event is the only one of its kind, which focuses on revenue optimization & compliance, including the latest trends in the high-tech and life sciences industries. It gathers industry leaders, professionals, and experts to share insights, best practices, and innovative strategies for maximizing revenue and ensuring compliance. Rainmaker24 was full of expert speakers, panel discussions, workshops, and networking opportunities, offering valuable content for attendees to drive growth and efficiency in their organizations.
The keynote address for the High-Tech track was delivered by Malcolm Penn, Founder and CEO of Future Horizons. The session description was: Get a pragmatic view of the macroeconomic factors impacting market dynamics across high-tech manufacturing ecosystems. Malcolm Penn, founder of Future Horizons, offers context on what contributed to the current environment and shares his outlook on what's in store for the high-tech manufacturing industry for the rest of 2024 and beyond.
At the end of the session, the audience had the opportunity to ask a few questions of their own and gain an experts opinion. We captured this precious moment and are sharing with you now to benefit the integrity of the industry.
Question: This seems like a bit of a reversion to the mean which in turn signals a great opportunity to plan ahead. We are still going to need to catch up from a unit perspective which in turn means close coordination with whoever is selling your product and your value chain. Planning how you can make that next growth spurt is actually the most important thing to do right now because you are going to have the time rather than waiting for the unit demand to increase. What are your thoughts?
Answer: Thank you, and absolutely yes, you gain market share in a downturn and not in a boom. In a boom period, your hands are tied, you cannot do a thing. Now is exactly the time when you can plan and make some real improvements in market penetration, market share gains, for all the reasons that you have mentioned.
This is exactly the time to be planning ahead but unfortunately some firms see this as a time to cut back. That would be tantamount to shooting yourselves in the foot. This is the time that you should be doubling down to come out of this down cycle ahead of the game, ready to take advantage of the next cyclical upturn. You have got to get your ducks lined up now because then you will be all ready to go.
Question: When you look at the capacity growth in the industry, I know that you think there is a trend line in terms of investment into capacity per technology node. Do you not think that the progression to 1nm is anomalistic in terms of how much investment needed on a per node basis? It seems to kind of defy the whole trend lines.
Answer: You need to look at that question in two ways. Yes, the absolute cost of each node transition is going up, as too is the cost of design node to node but that does not matter because the market size is also increasing so the relative cost compared to the market size is staying constant. We are spending no more per node today on 1nm in relative terms than we were spending on 10-micron technology back in the early 1970s. We are still spending proportionally the same amount of money on average, namely 14 percent of our sales on CapEx and a further 14 percent on R&D. So, when you just look at the headline data, you think, oh my gosh, this is completely unaffordable but when you then look at the relative data, these costs have always been 'unaffordable'. Because the market has been growing, the revenue from which to fund that development has also been growing and it is all still affordable. The underlying fact remains, if you stop investing in technology, then we will undoubtedly fall behind. You will no longer be a leading-edge player, but a middle of the road player, and then maybe a legacy player. So, it determines your business strategy much more than it does determine your financial strategy.
One problem however is that this node and CapEx investment money is being funded in the wrong part of the value chain, with the end customers taking the lion's share of the value chain profits but paying nothing for the investment in technology, and that is systemically not right. I know it is how capitalism works… "Why should I pay for your lunch? You want to sell to me; you pay your investment costs"… but that might not be the best way to ensure longer-term security of supply.
Apple and TSMC seem to have worked out a more symbiotic relationship for the development of each new bleeding edge node, no doubt because TSMC would not be able to afford this on a stand-alone basis, and Apple gets the benefit of a clear run for the first year of production before TSMC starts wafer runs for other customers. But Apple still gets takes a disproportionally large share of the value created by the semiconductors, as too does Nvidia, for example, selling systems based on a US$ 400 TSMC processed chip for $30,000.
Nobody ever said life is fair, and semiconductors are certainly not fair, but the fact remains the industry needs to find a way for more of that value creation to trickle down to the chip firms because funding new node developments and US$ 30 billion factories is increasingly hard.
Question: So, you will consider the industry is recovered when the unit prices goes up yet on the other hand you said that the average ASP growth was zero. Doesn't that mean it can never happen? Why will it go up again?
Answer: So, there are two sides of Moore's Law. You can either take a constant chip size, shrink the geometries and keep adding more and more functionality onto it, giving more functionality per square centimeter of silicon constant, or you can take the same functionality and just simply make the chip size smaller so decreasing the cost and ASP per chip. Historically we seem to have done those in equal measures so net result evens out.
Will it be that way in the future? The balance of probabilities is yes, unless we can see a valid reason why not. I think it unwise to say categorically it will always be zero, but the balance of probability is it is more likely than not.
Ultimately it is a mix issue. What is a US$ 100 chip today will be a US$ 1 chip in five years' time and tomorrow's US$ 100 chip will be significantly more complex than today's. It is like a conveyor belt. The belt moves at a steady rate but the parts on it keep on getting more complex.
Looking at the absolute value of ICs is also quite revealing, showing that the long-term IC ASP average is a dollar. Gordon Moore first hypothesized that, it was actually his second law, but nobody liked that law, so it became consigned to history, even though it is proving quite prescient.
The average US$ 9 sales revenue per square centimeter of processed silicon is another long-term industry constant. Again, it is a mix issue, with advanced chips selling for much more, but there are a lot of other less complex parts sold at under US$ 1 per square centimeter.
Unfortunately, we had to wrap up the session to continue with the conference, but Malcolm will join us once again for our upcoming Channel Chat Series, where we will unveil our joint report "Semiconductor & Electronic Component Industry Report and Outlook on 2025." Check out the virtual event and register now for more impactful insights.
Established in April 1989, Future Horizons provides semiconductor and electronics industry consulting, market research reports, training and business support services for use in opportunity analysis, business planning and new market development. Emphasis is placed on the world-wide microelectronics industry, and the European market environment.