If AMD buys Xilinx, it may return to debt tension

Recently, the news that AMD hopes to acquire Xilinx has received a lot of attention. After our multi-party verification, this transaction is likely to be completed before the next earnings conference call. But the capital market did not seem to buy it, and its share price subsequently fell 4%. Xilinx is currently valued at about $30 billion, but it seems that AMD will not easily accept this price, because the former cannot ignore the huge debt scale.

However, this does not affect my long-term bullishness on AMD. A large part of the reason depends on the existence of CEO Lisa Su. Even so, this transaction is not a smart move in my opinion.

First, Xilinx isn’t cheap.

Xilinx is a chip company that makes base stations for data centers and 5G communications. The stock tumbled during 2019 after shipments fell due to trade tensions with China that limited their cooperation with Huawei. At the same time, since Huawei accounted for as high as 8% of Xilinx’s revenue in 2019, this also led to a decline in the company’s full-year revenue.

In addition, Xilinx is a company that has not experienced a lot of revenue growth in the past decade and is stuck in a development bottleneck. The company generated $2.4 billion in revenue as early as 2011, but it only recently exceeded $3 billion in revenue, and then encountered the problem of Huawei’s supply cut.

On the other hand, AMD is a star company that runs all the way at a speed of 30%. In contrast, we think AMD may have seen some potential to join forces with Xilinx, such as increasing market share in 5G base stations, or acquiring more data center sales. Potentially, AMD prompted the merger to meet customer demands for new chip technology.

Unfortunately, however, analysts are adamant that, just as Intel’s acquisition of Altera was a failed example, it was a failed move due to limited synergies.

Objectively speaking, Xilinx is indeed not expensive compared with AMD’s higher forward price-earnings ratio. The deal isn’t necessarily dilutive, and Xilinx posted an operating margin of 27.5% in fiscal 20, which ended in March, based on a gross margin of 67.7%. The company’s profit margins are much higher than AMD’s.

But this advantage will be offset by operating expenses. Xilinx’s annual SG&A spending is about $370 million, which can be partially cut in cost synergies from the transaction. But the business units don’t overlap, so the combined company can’t cut significant operating expenses.

From a technical point of view, the deal still has merit.

Just like the acquisition of Intel and altera mentioned above, AMD should also be optimistic about the absolute FPGA technology advantage that Xilinx has. The advantage of this technology is that the chip can be reconfigured and re-edited, which means that once customers choose your product, retention chances will increase considerably.

Companies in the industry are very optimistic about the prospects of FPGA technology, so I do not doubt that this deal makes strategic sense, and even I think the “leak” of the news is likely to stir up competition.

But AMD can’t ignore financial issues like Intel can.

The combined company will end up with about $20 billion+ in debt, which is a huge amount, and our guess is that AMD will use the debt to offset some of the cash/stock interactions in the acquisition.

Xilinx has about 244 million shares outstanding and is on track to generate $854 million in net income based on an EPS target of $3.50 in fiscal 2022. Even with low interest rates, the deal is unlikely to provide much additional income due to its modest valuation. A $30 billion debt deal at 3% would add $900 million in interest expense. This makes it much more difficult to generate value-added benefits through synergies.

It’s worth noting that the stock deal could be dilutive to future profits, as AMD is on track to hit its EPS target of $4.50 within a few years. At $4.50 per share, AMD will generate $5.4 billion in profit in a few years. Xilinx’s expected earnings in 18 months are only $85 million. That means the Xilinx business must generate $1.6 billion in annual profits to not dilute AMD’s future EPS potential. At a valuation of $35 billion, at the current price of $83, the deal would issue 422 million shares and would require $1.9 billion in annual revenue.

So it’s clear that the debt deal will put AMD back on the tight balance sheet it’s been in for the past five years. From a technical point of view, this is a good business combination, but from a financial point of view it may not be successful.

The Links:   NL8060BC26-28 G104SN02