SoftBank's Masayoshi Son made two major misjudgments

In 2016, when Masayoshi Son tried to convince investors that it was a smart move to buy ARM, one of the world's most successful chip companies, the SoftBank Group CEO conveyed A clear message: "In the age of the Internet of Things, I think ARM will be the champion." However, the Internet of Things, which aims to connect billions of everyday and industrial devices to the Internet, has been slower than expected.

Son's attempt to capture the chip design market for the Internet of Things was the first bet he made on an ARM that didn't pay off; the second was the failed $66 billion sales of ARM to Nvidia. In walking away from the sale, SoftBank said that despite good-faith efforts by both parties, significant regulatory challenges were preventing the deal from closing.

Bet on the Internet of Things miss the server chip market

At present, ARM still occupies a dominant position in the field of smartphone chip design. Smartphones remain the most ubiquitous form of computing, but their growth has slowed considerably in recent years. Ahead of an initial public offering (IPO) that could be launched as early as this year, ARM is gearing up to strengthen its position in new, hitherto underexploited markets, while trying to boost profits to attract a new crop of investors.

Arm’s CEO-designate Rene Haas said in an interview that Arm’s products are now “much more competitive” in data centers and automobiles than when SoftBank bought the company.

When Son launched the $31 billion acquisition of ARM, he saw it as a bet on the future of the entire tech industry. At the time, the future around IoT concepts was taking shape. He continues to push the executive team firmly toward designing chips for a machine-connected future. Five and a half years later, it is becoming clear that IoT gambling is a costly disaster. And it also distracts ARM from attacking Intel's dominance in the much larger data center market.

As Son's vision collided with reality, SoftBank quietly revised its own estimates of the market. A 2018 report predicted that by 2026, the IoT controller market will be worth $24 billion and the server market will be worth $22 billion. However, a similar report in 2020 predicts that the IoT chip market will be just $16 billion by 2029, while the server market - where ARM currently has only 5% of the market - will reach $32 billion. SoftBank also lowered its estimate of the value of the IoT market, from $7 billion in 2017 to $4 billion in 2019.

Strong opposition to IPO for sale to Nvidia

In December, ARM strongly opposed the IPO in a regulatory filing, supporting a sale of the company to Nvidia. ARM outlined in the filing that shareholder pressure from going public would curb the company's ability to invest in the data center and PC markets, which have been "difficult to conquer" and where the company has made only "limited progress." ARM added in the filing that public market investors will "demand profits and performance," which means companies need to cut costs and lack the financial resources to invest in innovative new businesses.

"We've always believed that the Nvidia acquisition would give us a great opportunity to invest and do more," Haas said. "Now that we're in the (IPO) stage, I'm very optimistic about our prospects. ."

Underestimating the cost of innovation

Son also underestimated how expensive it might be to innovate in semiconductors. Although ARM does not make its own chips, the cost of ARM soared from $716 million in 2015 to $1.6 billion in 2019, according to SoftBank. In 2019, ARM's revenue rose 20% to $1.9 billion, while profits plunged nearly 70% to $276 million.

More recently, ARM has embarked on self-correcting steps, increasing its investment in the server and PC markets that have grown over the past four years, winning allies such as Apple and Inc.'s cloud computing service AWS. AWS is developing the third-generation Graviton chip based on the ARM architecture, while Apple is transitioning the entire Mac lineup from Intel processors to the ARM architecture-based M1 self-developed chip.

"While IoT is still a very important area for us, we're very, very focused on computing," Haas admits, referring to servers and PC chips. He declined to say how much of ARM's revenue came from outside its core mobile business, citing "rigorous regulatory processes" for the Nvidia deal.

Post a Comment