Why does it take so long to make more chips?
According to US Secretary of State, Gina Raimondo, chip shortages are likely to last through 2023. Intel’s CEO Pat Gelsinger, went as far as to predict shortages will last through 2024 in April. As many in the industry know, building new fabs to ramp up production is not a quick task – it can take 3-5 years to build a new fab. Expanding capacity of existing fabs is also not a “quick fix.” To truly understand why ramping up chip production is a problem without an immediate solution, it’s helpful to understand on the day-to-day level what is involved in installing new equipment in a fabrication center – from build to running full production.
The process of ordering semiconductor equipment to run full production at the customer fab can be a long, timely process. Besides the chips themselves, semiconductor manufacturing machines can be some of the most difficult equipment that we manufacture on a large-scale production manufacturing line. To take the equipment from a custom configuration order to running full production in a fab can take years based on the type of equipment you are ordering.
There are two methods of manufacturing and testing equipment. One is Integrated Final Test (IFT) and the other is the Module Final Test(MFT). IFT is when all of the different modules like chambers and mainframes are manufactured,integrated, and tested as a whole unit completely from back to front just like it would be set up in the fab. IFT is a more efficient way to test the unit but many fabs for the sake of time and money are going to modular final test (MFT). MFT is where individual units like the chambers, the factory interface, and the mainframe are built and tested separately, then a more detailed IFT test is run in the customer fab after all the parts are integrated. This method is cost-efficient for OEMs, especially when OEMs have separate modules made in different parts of the world.
After a tool is shipped to its customer, the tool is unpacked, inspected, wiped down, and brought into the fab, and, depending on the type of tool, this could mean just rolling it in and plugging it into the wall outlet or carefully floating in the tool to the specific location on a specially designed hovercraft that prevents any bumps or shocks to the fragile internal equipment. After setting the tool or separate modules on a template with all the floor cutouts and sometimes shock absorption platforms, the process of integrating all the chambers, factory interface, mainframes, cables, gas lines, and sub-fab support equipment to the tool begins. Depending on the tool type and the type of install speed the customer ordered and paid for, the installation process of the tool can take days up to many months to complete the full integration of a tool. But once the tool is finally integrated and fully connected to all the facilities it is now ready for power-up and test. Many people would assume this means flipping on a switch and turning on the machine, but this is far from what really has to be done. Some of the first testing includes running virus checks, connecting the tool to the fab mainframe / AMHS, and making sure it is talking to all the other systems. Other important systems checks include running leak rates, training robots, verifying configurations, and other facilities connected to the tool. Other testing includes the operations checking of all the different components and support equipment, running particle count wafers to be sure there are no internal particle issues, uploading recipes, and running dummy wafers to be sure all the robot handoffs are correct. After days or months of testing and bringing the tool online, the next stage of running the full process on test wafers begins.
What is a full process test? This depends on the tool and the process. Some of the equipment used in semiconductor manufacturing uses extreme temperatures as hot as the surface of the sun and pressures that can pump down to a vacuum of 1x10-9 Torr, relatively close to the vacuum of deep space that is 1x10-6 to <3x10-17 Torr.The final stage of bringing a tool online to running full production includes the final tuning, testing recipes, and verifying this tool can run full production for every wafer identically every time it runs the process.
As we can see, the process of ordering a tool, to running full production is complex. With equipment shortages and delays leading to lead times of up to 18 months from OEMs before a tool even arrives at the fabrication center, it is easy to see how the process of expanding capacity could take not months, but years.
200mm and Below, Semiconductor Equipment & Parts
With the rise of the power semiconductor industry and the current shortage of automotive chips, the 200mm market has secured a foothold that is ALMOST guaranteed here to stay. But, If you had asked the average person in the industry 23 years ago what they thought of the 200mm equipment they would have told you that the 200mm market had peaked and the migration of 300mm was here to take its place, as the phasing out of 200mm equipment began. Some fabs had upgraded all their facilities and tools to only run 300mm and all the new fabs being built were set up to run 300mm equipment. Some OEMs had shut down 200mm manufacturing lines as they ramped up the 300mm manufacturing lines and they slowly phased out supporting the supply of parts on some of the older 200mm and below equipment.
We were on track to keep up with Moore's law and the increased chip demand only meant that we would continue migrating toward using 300mm equipment until about 2015 when we started to see a rebirth of the 200mm equipment demand.
Today, the number of 200mm fabs has not only increased from 65 fabs in 1995 to 216 in 2022 according to SEMI but also the demand for chips using the 6µm to the 110nm nodes using older Legacy technology is keeping the average 200mm fab running at 100% capacity. Today we have far exceeded the 200mm peak of the 90s in demand, but the only difference is that a lot of the 200mm equipment being sold is used due to the fact that many OEMs quit producing and supporting them.
A critical part of keeping any fab running is to be sure to verify that you have critical replacement parts and backup equipment to keep the manufacturing lines running when there are breakdowns, also known as equipment downs. Today, some of that equipment and those replacement parts are becoming more and more difficult to find depending on the equipment you are running. And, when you are running at 100% capacity, that makes it even more critical that your backup supply is readily available when that time comes. Today a common practice of some fabs is to purchase used equipment so that they can use it for spare parts. Other fabs search for spare parts on general-purpose marketplaces like eBay when OEMs no longer support the sale of replacement parts.
With the advent and realization that the 200mm market is here to stay, and grow, many OEMs and fabs are carefully evaluating how this industry will move forward as they realize that the current methods are not sustainable to support the future growth of the 200mm market. Some rumors have surfaced that many OEMs may look at the possibility of bringing back some select 200mm equipment manufacturing lines and supporting spares for that equipment. Other options are to modify the 300mm equipment to run 200mm wafers or to refurbish older equipment.
According to Carter Hall, one long-time industry expert who specializes in the 200mm and below parts industry, the parts problem could be partly resolved through 3D printing technology.
Regardless of what the resolution is, we must come together to confront and resolve the upcoming challenges so o that we can continue providing semiconductor chips to the world.
Is it a Global Chip Shortage or a Global Capital Equipment Shortage?
Unfortunately building a semiconductor manufacturing facility is not as easy as flipping a switch. despite the $52 billion investment the CHIPS Act would provide in aid to the semiconductor industry, it still does not completely solve the source of our problems.
We could definitely solve many of these chip shortages a lot faster if we could get this Chips Act passed & confront the true shortage of capital equipment that is manufactured by original equipment manufacturers (OEMs).
Today OEMs are experiencing some of the worst extended lead times on parts and some of the worst supply chain issues they have ever had to deal with as they attempt to produce the highest amount of orders they have ever seen. With all the compounding problems, OEMs are also experiencing supply chain lags, staffing issues, geopolitical tensions, and COVID shutdowns around the world. Today, most OEMs struggle to produce and fill the current backlogs of equipment orders for their customers like Intel, Samsung, & TSMC.
Although it does not seem likely they will immediately solve many of these issues at hand, it may take some out-of-box thinking and time to solve many of these bigger problems. Companies like Moov Technologies and GTI are attempting to solve some of these issues by refurbishing older equipment and reselling idle or unused equipment to companies that are either using older technology or just can not wait for the extended lead times we are seeing for purchasing equipment today.
To deal with chip shortages, we have to confront the problem at the source. Capital Equipment spending is expected to reach $98 billion in 2022. As it turns from a chip shortage to a manufacturing equipment shortage OEMs and Fabs will have to work together and in conjunction with government funding to proactively plan and deal with this global surge of capital equipment spending. As OEMs benchmark proactive companies like Intel who are investing in things like education as they train the future workforce to run the fabs of tomorrow, OEMs too will need to invest and act in a proactive method to better handle the problems at hand.
The Future of Power Semiconductors
What are power semiconductors and why are we seeing so much industrial growth in the Power Semiconductor sector? In the past, we have used (Si) also known as silicon as a substrate for semiconductor chip manufacturing. With the industry growth & demand for servers, portable electronics, electric vehicles, and EV charging stations we have also seen an increased demand for products that can handle higher voltages, higher temperatures, and of course a reduction in the size of the chips.
The industry has faced many challenges meeting this demand, due to the fact that (Si) Silicon is not able to handle these higher limits. Now, with the advanced discovery of one Power Semiconductor, Silicon carbide (SiC) is a next-generation material that aims to significantly reduce power losses and enable higher power densities and frequency switching while reducing heat dissipation, thus enabling it to use electrical energy more effectively. According to Maurizio Di Paolo Emilio, PhD, an industry expert, "the semiconductor industry is focused on wide bandgap materials for energy and EV applications in particular." As the global demand for these technologies grows, we will continue to see the advancements in technology that will bring us into the new world of safer, cleaner, and faster innovations.
The U.S. Semiconductor Industry Resurrection will not come without its challenges
Global semiconductor sales just surpassed the $600 billion mark and it is expected to become a $1 trillion per year industry by the 2030’s. With six or more large fabs going up in the US alone, and with this continued growth around the world, we will also experience many challenges.
One of the top challenges we will face is labor: according to Korn Ferry, labor shortages in the semiconductor industry are currently costing companies over $2 trillion around the globe. And that outlook is not expected to get any better in the short term within the semiconductor industry. With the skilled labor shortages, we will continue to see competition between companies – as well as opportunities for employees to advance their careers for higher-paying jobs or better work-life balance. Today the industry is seeing incredible turnover and a large number of experienced workers leaving the workforce that will be looking to retire in the near future. Currently, the US semiconductor industry is not able to handle this large turnover with such little backfill. The impact of labor shortages will continue to worsen the semiconductor shortages we are already seeing.
Chipmakers and equipment manufacturers currently have two options to fill skilled positions. They can either look for candidates who have been trained and educated through a university and obtained an academic degree, or they can hire candidates who have been trained as tradespeople through many years of hands-on experience. Both are slow processes. With technology advancing at the speed that it does, universities cannot give students hands-on experience with the latest technology since the equipment can cost over $100 million for a single machine.
The second major challenge our industry is facing is related to the supply chain. Today, semiconductor equipment manufacturers have an 18-month waiting list. Some specialized transformers required to build a fab are on a backlog for 30 months. This week, Nanya Tech announced they will delay the build of their new $10.3 billion fab by 6 months due to labor shortages and supply chain issues. The other supply chain issue we may experience further down the line is wafers, will wafer manufacturers be able to keep up with the supply needed for all these new fabs running at full production? Supply chain issues are not a temporary problem, but a larger challenge the industry will need to tackle in the future due to all the infrastructure and resources required to keep a semiconductor manufacturer running at full production.
As we continue to navigate an era of rising global instability and war, additional resources required to keep the semiconductor industry operating smoothly may prove bottlenecks. For example, recently there has been much talk focused on Neon gas. . Currently, Neon mostly comes from war-torn Ukraine, There are a lot of unknowns about whether this supply chain will be sustainable in the future.
According to the CEO of Global Foundries, the chip industry output must double in the next 10 years. In order to support that portion of growth in the US, there are several possible solutions to these issues that challenge the industry. To begin, Congress needs to pass the new Chips Act bill so that we can fund companies like Intel who are going to invest $100 million over the next decade to establish semiconductor manufacturing education and research collaborations with universities, community colleges, and technical educators across the U.S.
To further help resolve labor shortages, the US government can open the doors on immigration policies so highly skilled and educated workers from other countries can apply for open positions and university scholarships that are not being filled.
To make this industry growth in the US successful, we must collaborate with private corporations so they can build a sustainable industry together. Corporations and governments partnering and funding universities will be necessary for the future of this industry.
Why is everyone talking about 200mm fabs and equipment?
Back in the day, many industry leaders expected the 200mm market to dwindle over time as the 300mm market was more profitable because you could fit more chips on a single wafer and 300mm equipment had the newest technology going down to the smaller nodes pushing Moore's law every year. But, in 2015, there was a noticeable change in 200mm demand as we began to see many automotive and IoT manufacturers design their own chips and prefer to manufacture on 200mm. Suddenly, we started to see more production of the power chips made with gallium-nitride (GaN), and silicon carbide (SiC) wafers that are mostly made in 150mm and 200mm fabs, and we saw a total number of 200mm fabs around the globe go from 184 in 2016 to 216 in 2022. This sudden increased demand in the 200mm market altered the way that we looked at the old 200mm and 150mm equipment as most OEM manufacturers had already shut down their 200mm manufacturing lines and moved on to manufacturing the new 300mm equipment.
As manufacturers scrambled to figure out how to ramp up 200mm manufacturing lines, they found that there were parts and assemblies made by vendors that were no longer in business or they no longer manufactured those parts. As some OEMs began trying to manufacture new 200mm equipment, others started refurbishing older 200mm equipment for resale. Recent supply chain issues and chip shortages have exacerbated the issue, and used 200mm tools have skyrocketed in value as the wait times to purchase a new or refurbished 200mm tool have gotten longer and longer and the demand for these tools increases by the day.
Companies like Intel are dealing with this problem by snatching up 200mm fabs in mergers & acquisitions, like Tower Semiconductor, which Intel recently purchased for $5.4bn. As 200mm fabs around the world are running at 90% capacity or more, we will continue to see more M&A’s by big players like Intel move back into the 200mm business.
Another way companies are dealing with 200mm equipment shortage is by diversifying their supply chain strategies. Companies are increasingly turning to the secondary market, incorporating used equipment into their sourcing strategy. Indeed, at Moov, we have seen the average sale price for used equipment increase over the past year – in some cases more than double!
According to SEMI, the long-term trend to the end of 2024 is for a 17 percent increase in capacity for 200mm facilities. However, analysts believe this will not be enough to keep up with demand. The Electronics Components Industry Association recently issued a statement warning of continued shortages.
“As demand for automotive electronics has rebounded, the shortage of chips produced on 200mm wafers has become much more acute,” according to ECIA’s Joel Huskra. “The typical car requires 50 to 150 semiconductors.”
In conclusion, the 200mm market is here to stay and we can expect to see the value of pre-owned 200mm equipment hold for the foreseeable future.
Semiconductor industry facing critical staffing issues
This last year the semiconductor industry has faced major challenges from logistics delays, to supply chain shortages, to natural and man-made disasters. The latest challenge to the industry is being able to staff the open positions for thousands of skilled roles. Taiwan, a global semiconductor manufacturing hub, had thirty-four thousand industry vacancies in the month of December 2021 alone. The multi-billion-dollar industry investments to expand production and solve chip shortages have only exacerbated challenges to fill the skilled engineering workforce that is required to operate a fab or an original equipment manufacturer (OEM) facility.
Amidst this hiring frenzy to gain experienced and skilled staff, talent poaching within competing companies is becoming a bigger issue as companies are offering up to 70% more than their current salaries for skilled workers. Not only are semiconductor companies competing against each other, but they are also competing against companies like TESLA, SpaceX, and AMAZON who are creating an increasing talent war for skilled technicians and engineers.
So how does the industry plan to face these issues? Some places like Taiwan are leading the industry as they are racing to set up many specialized chip schools that are in session year-round. These academic institutions are collaborating with many of the fabs and OEMs to not only get the knowledge but also the hands-on training that one needs to work on semiconductor manufacturing equipment. Currently, some countries are evaluating and amending their immigration policies allowing people from other countries with proper education and experience to enter the country under special work visas. And lastly, some countries and manufacturers are ramping up investment in and partnerships with academic institutions as they plan for a sustainable future in the industry. As the industry looks for solutions to this talent shortage, we can look toward Taiwan to set a standard to help further sustain the talent pool for the semiconductor industry.
Moov adds $2M in funding from Mark Cuban, NFX, and Flatiron Health Founders Nat Turner, Zach Weinberg
San Francisco, CA, March 3, 2021 — Moov, a data-fueled marketplace for used manufacturing equipment, today announced a $2M strategic investment from VC firm NFX, Mark Cuban, Nat Turner, and Zach Weinberg, co-founders of Flatiron Health, and other strategic angels. This additional investment extends Moov’s seed financing total to $4.4M, after an oversubscribed seed round last year, led by NFX with investment from Tencent co-founder Jason Zeng’s Decent Capital, David Adelman’s Darco Capital, Great Oaks Venture Capital, and other notable angel investors.
James Currier, General Partner at NFX, believes the global market for used semiconductor manufacturing equipment is vastly underestimated. “After working with Moov for two years, it’s clear they are not only improving existing transactions, but also revealing latent supply and demand to actually expand their market. For pre-owned high tech manufacturing equipment, this is a $100B+ opportunity.”
Moov is addressing this opportunity with technology that makes the end-to-end process of purchasing pre-owned semiconductor equipment significantly faster and more efficient.
“We’re humbled by the votes of confidence from these notable investors and entrepreneurs,” said Steven Zhou, Co-Founder and CEO of Moov. “Now, more than ever, the secondary equipment market provides a vital solution to enable U.S. and global manufacturers to quickly scale production and respond to growing chip shortages, while creating a liquidity channel when surplus capacity is no longer needed.”
Recent consumer trends like work from home and increased streaming as well as the sheer proliferation of smart devices across end markets has created a global chip shortage. Scaling up production is no simple matter — building and equipping a fabrication center costs billions of dollars and lead times on new equipment to manufacture chips are over a year. Meanwhile, end markets, like consumer electronics, cycle quickly, meaning entire production lines — hundreds of millions, if not billions of dollars — may prove obsolete within the span of five years.
“The semiconductor manufacturing industry is notoriously difficult for small and midsize companies to navigate due to capital requirements, rigid supply chains, and fast-cycling markets,” says Mark Cuban. “While the global Covid pandemic created new challenges and subsequent opportunities for traditional chip manufacturers, we saw small and medium size businesses struggle to adapt to the new market demand. Marketplaces like Moov are changing this dynamic by creating new opportunities for SMBs to compete in a highly antiquated industry, in a cheaper, faster, and all around more efficient way. This is one of the many reasons why we believe that Moov is well positioned to be an essential part of the industry in the future.”
“Moov is bringing the power of data and automation to 5- to 7- figure transactions that still largely occur via pen and paper,” says Nat Turner, Co-Founder and CEO of Flatiron Health and General Partner at Operator Partners. “By building the infrastructure to amass and intelligently leverage data on the resale value of capital equipment, Moov is bringing transparency and reliability to the secondary semiconductor and broader manufacturing equipment markets.”