SoftBank, ARM, Nvidia (and Vision Fund) - not as good as it looks
Updated: Nov 2, 2020
SoftBank Group (“SFTBY”) and Vision Fund announced yesterday that had reached agreement for SoftBank to sell ARM to Nvidia (“NVDA”) for (up to) $40 billion, as follows:
$2 billion cash paid at signing,
$10 billion cash and circa $21.5 billion (44,366,544 shares) paid at closing,
Up to $5 billion based on contingent payments (cash or NVDA shares, option of SoftBank/Vision Fund) based on ARM meeting certain performance targets post-closing, and
$1.5 billion of NVDA stock awards payable to ARM employees.
You can find the details of the transactions here, in SoftBank’s press release from yesterday.
So, what was SoftBank’s return on its original investment in ARM? I will provide some more detail below based on the way I see it, but it appears SoftBank and Vision Fund realised a gain of around $10 billion on their investment over a four year period, or a total return of 33.3% over the period. This equates to an average annual return of 10.1%/annum over a four-year period. That sort of return might sound OK on the surface, but let’s see how this $30 billion would have done in some other investments over this period.
Interestingly, NVDA – the buyer of ARM – was once an important investment for of SoftBank (and subsequently Vision Fund), which held shares for around two years from early 2017 to early 2019. Had SoftBank/Vision Fund just held on to their publicly-listed NVDA shares purchased in early 2017, its $30 billion would be worth over $150 billion based on yesterday’s closing price. This means that SoftBank would have increased its original investment by over five times over a shorter period, generating an average annual return of 54.8%/annum. Perhaps even more telling is that fact that had SoftBank simply invested in index funds for the S&P 500, the NASDAQ or even (domestically) the Nikkei 225, it would have also generated far superior returns than it did on its investment in ARM, with substantially less risk and fees paid out (to third parties or internally to Vision Fund), as you can see in the table below.
It is true that SoftBank (and the Vision Fund) benefit from raising liquidity related to the sale of ARM, although keep in mind that the proceeds flow in over a period of time extending to 2022 as outlined in the press release. The harsh reality though is that the transaction end-to-end does not result in a particularly good outcome for the company or its VC arm, Vision Fund, either when compared to normal VC investor expectations or even to alternative investments available during this period. For me, it also flags some of the issues with internal transfers between SoftBank and Vision Fund, raising issues of governance and the ties between SoftBank and Vision Fund – see final section below for more detail. You can read more about my general Vision Fund concerns in this post in emorningcoffee.com from July 16thhere. To be fair to SoftBank, I will add a caveat that the auditor of the financial results of SoftBank during the periods that these transfers occurred gave the company a clean bill of health. Nonetheless, I would not like these sorts of transactions under any circumstances were I a third-party investor in Vision Fund.
And if you want to know more…..
Let’s look back for more context. Softbank closed on its acquisition of ARM in September 2016 for around $30 billion, as you can see from the extract below from the company’s
2017 Annual Report. For reference at the time, SoftBank reported in its interim financial statements that ARM had six-month revenues for the period ended September 30th, 2016 of ¥14.356 trillion ($1.45 bln) and EBITDA of ¥7.236 trillion ($733 million).
Subsequently, according to SoftBank’s interim financial results for the 6-month period ended September 30th, 2017, SoftBank pledged to commit $8.1 billion of ARM shares (24.99% of its holdings) to Vision Fund as part of its $32.5 billion commitment to the Vision Fund (in interim financial report for period ended September 30th, 2017).
As far as NVDA, SoftBank bought 4.9% of NVDA in early 2017 (reported May 2017), then transferred it to Vision Fund early in the fund’s inception, reported in the interim financials for the period ended June 30th, 2017 (1Q2018, see p. 19). Based on back-of-the-envelope calculations, it would appear that SoftBank invested around $3.1 bln to buy its interest in NVDA. Subsequently, at the time it agreed the transfer of the shares to Vision Fund, SoftBank booked a gain on the sale of ¥106.9 trillion ($950 million), implying that the sale was for around $4 billion.
Vision Fund subsequently exited the NVDA investment in early 2019. Over their holding period (and certainly since the transfer of the shares to Vision Fund), the stock appeared to be more or less flattish albeit volatile. The graph to the left depicts the performance of NVDA over this period, certainly a disappointing chapter for NVDA investors, and especially in retrospect given the amazing performance of NVDA shares since then as they have more than tripled.
Shareholders of both SoftBank (shares up 7.9% ) and NVDA (shares up 5.8% yesterday) like this transaction for different reasons. However, what should not be lost in the shuffle is the less-than-stellar outcome for SoftBank and Vision Fund as far as demonstrating its investment acumen, at least on this particular transaction.