By Louise Lucas in Hong Kong
Seldom have such small chips carried such a heavy load.
China’s fledgling semiconductor industry, after a messy birth, now finds itself in the crosshairs of conflicting political goals in Beijing and Washington.
Chips are a key plank of China’s industrial plans, attracting a massive $150bn government subsidy.
This reflects national angst over reliance on overseas markets — it spends more on importing semiconductors than oil — and a fear, as one banker puts it, of being left in the dark “in case the US flips the switch”.
Washington harbours its own fears: that these subsidies will distort the market, dent its own domestic industry, jeopardise the edge it has held in the technology and threaten security.
“Chinese industrial policies in this sector, as they are unfolding in practice, pose real threats to semiconductor innovation and US national security,” wrote the President’s Council of Advisors on Science and Technology in a letter to Barack Obama earlier this month.
China consumes more than $100bn worth of semiconductors, or roughly a third of total global shipments, but produces just 6 to 7 per cent by value, according to consultancy Bain & Co.
Many of the imported chips go into PCs, smartphones and other gadgets that are then exported, but there is still a yawning gap between the semiconductors made by domestic chipmakers and the number consumed by a gadget-hungry Chinese public.
Early efforts to narrow this gap failed: a scattergun approach resulted in a highly fragmented sub-scale industry.
McKinsey, the consultancy, calculates Beijing had invested in 130 fabrication plants across more than 15 provinces at one point.
But lessons have been learned, says Mark Li, semiconductor analyst at Bernstein, pointing to Hong Kong- and New York-listed SMIC.
Initially, he says, it wanted to be as good as the top manufacturers and invested heavily in high-end equipment, racking up losses as a result.
“They adjusted their strategy to be a follower, and stay one to two steps behind Taiwan’s TSMC [the industry leader as a foundry making chips for others] and lowered investment a bit and reduced their R&D investment.”
SMIC, the fifth-biggest semiconductor foundry by revenues in 2015, according to Gartner, produces chip wafers mainly for communications and consumer devices.
Roughly half its revenues come from overseas.
This has served SMIC well.
Its shares have generated a total return of 27 per cent in the past 12 months, massively outperforming the benchmark Hang Seng index.
Of the 28 coverage analysts tracked by Bloomberg, not one recommends selling and 22 have a ‘buy’ rating.
“Normally size gives you better margins, but SMIC defies this,” says Mr Li.
“And that’s why it’s remarkable.”
He estimates SMIC will lift revenues 30 per cent this year, compared with a 2 to 3 per cent rise predicted for UMC, Taiwan’s number two player.
But at a national level, SMIC is just one of China’s champions.
More ambitious is Tsinghua Unigroup, which in July merged its memory chip operations with government-run XMC.
Tsinghua has sought to plug another gap — a lack of leading-edge technology — by trying to buy overseas companies, including Micron of the US, for some $23bn.
Yet it has been largely stymied by US regulators.
Other attempts have met a similar fate.
In February, Fairchild Semiconductor turned down a $2.6bn bid from Chinese state-backed enterprises over fears that the deal would be blocked by the US authorities.
Failure to buy-in technology is one of the biggest stumbling blocks to Beijing’s efforts to become a global power in chips, say analysts.
“Buying technology from the US will be more and more difficult,” says Mr Li of Bernstein, who alludes to a blocked attempted acquisition of Aixtron to show that even European deals fail.
“But I think China will continue trying.”
Roger Sheng, analyst at Gartner, says: “Without technology acquisition via [takeovers, joint ventures] or technology licensing, Chinese local companies still lack the capabilities to produce high performance processors and DRAM/flash [memory chips].”
Kevin Meehan, who heads Bain’s Asia technology practice, adds: “They can make progress in total production capacity. But they don’t have a clear path to obtain leading-edge process technology.”
Nor is it just acquisitions at which the US is balking, as testified by this month’s report to Obama.
In November, Penny Pritzker, US commerce secretary, attacked China’s $150bn plan, designed to ramp up Chinese-made integrated circuits at home to 70 per cent in 2025.
Despite setting a bold goal, Beijing has been less clear on how it plans to get there or how the spending will be directed.
But multinational chipmakers have spotted the momentum shift and have begun setting up manufacturing and partnerships in China.
The authors of the report to the US president, while noting that the $150bn spend is below the average $23bn spent annually on M&A by US semiconductor companies in the past five years, said acquisitions would help advance China’s aim.
Massive subsidies and intellectual property theft
They identified two further strategies: “subsidies and zero-sum tactics”.
Three examples were listed of the latter, including “forcing or encouraging” domestic buyers to only purchase from local suppliers, “forcing” technology transfer in exchange for market access and intellectual property theft.
Analysts add a further risk, one that has repercussions across the globe: pricing.
As one analyst baldly puts it, “China is good at getting into businesses,” as it did with solar panels. “They are also good at shrinking profit pools” — alluding to the consequent oversupply and collapse of prices.
President Donald Trump heightened concerns in December, when he raised the prospect of levying duties of up to 45 per cent on Chinese goods to level the playing field for US manufacturers.
Such tariffs could hurt not just Chinese companies, but also the multinational players such as Qualcomm, Intel and Samsung that have set up shop in China, through joint ventures or partnerships.
Some analysts point out that this co-operation, together with an ecosystem of suppliers on the ground, means China is closer to its goal of creating a globally competitive industry.
“They’ve done it in PCs, in high-speed rail — but haven’t done that yet in semiconductors,” says Mr Meehan.
“They are certainly closer than they were, but they are going to face obstacles.”
Regulators prove a stumbling block
The graveyard of failed semiconductor deals is littered with Chinese names that have fallen foul of US regulators.
Most recently, the Obama administration last month blocked Fujian Grand Chip Investment’s proposed €670m acquisition of German chip equipment maker Aixtron, which has a US subsidiary making products that have military applications.
The US Committee on Foreign Investment in the United States blocked the deal after flagging up security concerns to Aixtron, whose technology is being used to upgrade US and foreign-owned Patriot missile defence systems.
Berlin earlier withdrew clearance amid similar concerns that chips made with Aixtron equipment could be used in China’s nuclear programme.
Before Aixtron, there were Micron, Fairchild and Lattice Semiconductor — all of the US, and all targeted for bids or investments.
Some were thwarted by the threat of a Cfius red light, rather than a veto per se.
Take Fairchild Semiconductor, which cited the threat as reason to reject a $2.6bn Chinese bid in favour of a lower bid from ON Semiconductor of Arizona.
Chinese companies too have called a halt on proceedings ahead of Cfius reviews, the reason given by Tsinghua Holdings’ Unisplendour for pulling the plug on its $3.8bn bid for a stake in Western Digital, a US data storage company.
Not all blocked deals were obvious security threats: many bankers questioned the quashing of Philips’ planned $3.3bn sale of its US-based Lumileds lighting division to a consortium of Chinese private equity buyers.
But while China’s outbound hunt for chips may be stymied by politics and regulation in the short term, bankers — who have their own interests in robust M&A appetites — do not see mainland buyers giving up.
“This is going to be relevant for the next 10 years,” says one banker.
Aucun commentaire:
Enregistrer un commentaire