The semiconductor equipment segment has seen quite a bit of consolidation such that today there are just a handful of companies worth looking at. But this has led to some really big players such that further consolidation may be difficult due to anti-trust hurdles.
Earlier, Applied Material’s (AMAT) proposed acquisition of Tokyo Electron fell through on anti-competitive concerns and now there is a possibility that similar issues in China, Japan and Korea will work against the Lam Research.
Here is a brief analysis of the two market leaders in the backdrop of shrinking process nodes, 3D NAND production ramp up and an emerging Chinese semiconductor sector.
Applied Materials remains the market leader in wafer level manufacturing equipment according to Gartner with a 19.1% share. The company has been the market leader for as far back as I can remember and it continues to take share (growing 1.3% in a market that shrank 1.0% in 2015).
But Lam Research has seen more remarkable growth of 24.7% in 2015 to end the year with a 14.3% share of the market. Lam has grown through acquisitions, most notably that of Novellus Systems in Jun 2012.
Unlike other markets, semi equipment includes lighting sources, process control equipment and other things with companies specializing in specific areas, so Gartner’s top 10 aren’t always directly comparable and share gains aren’t necessarily indicative of competitive strength.
The two major technology inflections driving semiconductor engineering at the moment are shrinking process nodes and 3D NAND. Major design breakthroughs and innovations are very positive for equipment makers in particular because they involve increased complexities that can only be taken care of with new, specialized equipment.
Shrinking process nodes: Very simplistically, the CMOS manufacturing process for semiconductors involves the deposition of several layers of conductive and semiconductive materials on a silicon wafer and using laser light sources (now moving to extreme ultra violet or EUV) to implant the material in specific designs using stencil-like things called photomasks or reticles. The excess material on the surface of each layer is then cleaned or “etched” away using liquid or gaseous materials. What follows is a process of planarization, cutting, sorting and packaging the final wafers into “chips”. So when these chips become smaller in size, each wafer yields more, leading to cost efficiencies. Increasing wafer sizes also does the same thing because the number of chips per wafer increases but this can involve more capital because new capacity or significant upgrade of existing capacity may be required. Moreover, the need to shrink is leading to materials innovation IBM announced carbon nanotubes) because of the limitations of using silicon at the wafer level. The bottom line is, semiconductor manufacturers are therefore keen on shrinking chip size or increasing wafer size and every time this happens, they will buy fresh equipment first for designing, and then for volume production and to manage yields.
3D: As chip sizes shrink, there is the increased risk of electrostatic discharge, impacting yields and functionality. Semiconductor designing has therefore moved to three dimensional designs, thereby enabling both functionality and electrical stability in smaller areas. This again is leading to increased demand for new manufacturing equipment.
Applied Materials recently announced that it has started shipping its PROVision ebeam inspection tool offering down to 1nm resolution inspection for foundry, logic, DRAM and 3D NAND customers.
The company hasn’t specified the number of machines sold to date but said that it was more than 12, that two important foundry and memory manufacturers have taken delivery and that orders at existing and new customers had been received. Last year, it launched the Centura Tetra Z for etching at 10nm and below and announced earlier this year that its Selectra etching tool for design and manufacture of 3D logic and memory chips was shipping to foundry logic and memory chipmakers.
As far as Lam is concerned, the company recognizes that yield management in the more complex new processes will be linked to deposition and etch as materials used in these processes impact resistance and stress, which also affect yields. So it is trying to buy KLA-Tencor, which has historically led in process control and yield management equipment to integrate these functions in its equipment. As part of its Altus product line, the company announced that its new atomic layer deposition tool facilitating chip shrinking was already in use at several R&D sites as well as at leading 3D NAND and DRAM players. The company is particularly well positioned at memory manufacturers.
Given China’s growing importance in chip consumption and production, it is an important market for equipment suppliers. The Chinese government is aiming to domestically manufacture 40% of domestic consumption in the next five years. But while the government is doing all it can to grow the Chinese semiconductor segment, demand continues to outpace the output from these efforts. So the government is welcoming foreign players to fill the gap. For equipment suppliers this is positive because while it will try to rely on domestic companies as much as possible and even buy up U.S. technology, it will continue to buy equipment from international players as well.
Applied Materials has for long been a major player in China, so it has important relationships and regular business from the region. China revenue has continued to grow in the last few years although the rate of growth appears to be slowing down. This could be because of the Chinese government’s efforts or increased competition with Lam Research. It still accounts for around 17% of this market.
Lam Research will see increasing strength in China as both Intel’s Dalian fab and the Chinese company XMC (which has licensed 3D NAND technology from Cypress-owned Spansion) ramp up spending. Of these, Intel is in the tooling phase for 3D NAND, so its impact will be immediate. On the other hand, XMC just started construction, so equipping will likely be in 2017.
However, real gains for these leading equipment makers come from market share wrested from each other or if the market itself expands as a result of new semiconductor applications. That’s because even if manufacturing changes location (to China for instance), but chip volumes (and therefore capacity) doesn’t increase, there is no change in equipment demand. Market expansion will likely become the most important factor over time, as growth in the Chinese semiconductor sector means gradual and growing competition from newer Chinese players that the current market leaders will increasingly be pitted against. Most of the Chinese fabs are trailing edge, so upgrading to leading edge will also be a driver.