Discussion AMD Acquires Xilinx

[DisEnchantment]

AMD to Acquire Xilinx, Creating the Industry’s High Performance Computing Leader

Browse AMD’s company-wide and financial press releases.

ir.amd.com

Straight from the Horse's mouth

Strategic transaction strengthens AMD’s industry-leading technology portfolio —

• Expands AMD’s rapidly growing data center business
• Xilinx, the No. 1 provider of adaptive computing solutions, increases AMD TAM to $110 billion
• Immediately accretive to AMD margins, cash flow and EPS
• All stock transaction with combined enterprise value of approximately $135 billion

All stock. No Cash. Could not have been a better way to make use of that inflated market valuation.
Lots of people were questioning AMD why they are not capitalizing on their Market valuation. But this is it.

Management and Board of Directors
Dr. Lisa Su will lead the combined company as CEO. Xilinx President and CEO, Victor Peng, will join AMD as president responsible for the Xilinx business and strategic growth initiatives, effective upon closing of the transaction. In addition, at least two Xilinx directors will join the AMD Board of Directors upon closing.

They got decent advisors too.

Advisors
Credit Suisse and DBO Partners are acting as financial advisors to AMD and Latham & Watkins LLP is serving as its legal advisor. Morgan Stanley is acting as lead financial advisor to Xilinx. BofA Securities is also acting as a financial advisor and Skadden, Arps, Slate, Meagher & Flom LLP is serving as legal counsel.

Ex AMD rejoins the family bringing Xilinx with him

“We are excited to join the AMD family. Our shared cultures of innovation, excellence and collaboration make this an ideal combination. Together, we will lead the new era of high performance and adaptive computing,” said Victor Peng, Xilinx president and CEO. “Our leading FPGAs, Adaptive SoCs, accelerator and SmartNIC solutions enable innovation from the cloud, to the edge and end devices. We empower our customers to deploy differentiated platforms to market faster, and with optimal efficiency and performance. Joining together with AMD will help accelerate growth in our data center business and enable us to pursue a broader customer base across more markets.”

 

[Kedas]

Well, China rejects the merger and then you’re catched.

Searching for that China reject possibility guided me this article:
And there is also the fact that you do not know how AMD stock will change after the merge is final

$XLNX - Way to acquire AMD stock at 25% discount! | Utra...

TLDR: If you like free money - buy XLNX and wait. For context, AMD announced Xilinx acquisition in october last year: https://www.amd.com/en/press-releases/202...

utradea.com

 

[moinmoin]

Does that mean you can buy an AMD share for 105$ (instead of 121$) by buying an Xilinx stock now? There must be a catch otherwise everyone would be doing it I assume.

And there is also the fact that you do not know how AMD stock will change after the merge is final

You answered yourself already. Assuming the merger with Xilinx to be a done deal, buying Xilinx stock is highly preferable to buying AMD stock right now. But as you state in the second quote you don't really buy AMD's current worth at Xilinx's current price, you essentially bet on the price of both combined after the merger to be higher as well, which is not a given.

Here's a regularly updated graph showing the share price ratio of the two over time: https://public.tableau.com/app/prof...Ratio_16114624177710/XilinxAMDSharePriceRatio

 

[Doug S]

Anyone who thinks they're beating the system by buying shares in a to be acquired company should know that Wall Street is VERY aware of these sorts of things. If you see a big spread that looks attractive, what that tells you is that Wall Street thinks otherwise. Betting against the collective knowledge of Wall Street is a very good way to lose money in the long run, though it is always possible you get lucky in any one instance.

 

[moinmoin]

collective knowledge of Wall Street

lol

 

[turtile]

Anyone who thinks they're beating the system by buying shares in a to be acquired company should know that Wall Street is VERY aware of these sorts of things. If you see a big spread that looks attractive, what that tells you is that Wall Street thinks otherwise. Betting against the collective knowledge of Wall Street is a very good way to lose money in the long run, though it is always possible you get lucky in any one instance.

It doesn't matter what 'Wall Street' thinks - only reality matters. Nvidia has been pumped through the roof yet their acquisition of ARM is being investigated where Xilinx passed without issue.

What evidence do you have that the Xilinx merger will be struck down? There's generally a spread since most investment funds would rather buy AMD than take a small risk purchasing Xilinx.

 

[DrMrLordX]

It doesn't matter what 'Wall Street' thinks

It matters if what they think is going to affect immediate and near-future pricing on the stock.

fwiw the merger will probably not be struck down at all. It's just that, currently, AMD stock maybe a bit overvalued and that the market will likely correct before the stock swap takes place. If you held AMD stock valued at $120 and you knew that Xilinx owners would effectively be getting AMD sold to them at $105, would you not swap your shares for Xilinx shares first?

 

[TheELF]

Wait, so the actual price at which the stock exchanging will be done wasn't set at all?
Then how are they going to settle on the time it will take place? I mean AMD could wait for a time their stock is at the lowest or xilinx for the time AMD stock is at the highest, how will they decide?
What if one of the two stocks crashes hard?

I mean it's supposed to be a 35bil deal so that price has to be locked in somehow.

 

[DrMrLordX]

Wait, so the actual price at which the stock exchanging will be done wasn't set at all?
Then how are they going to settle on the time it will take place? I mean AMD could wait for a time their stock is at the lowest or xilinx for the time AMD stock is at the highest, how will they decide?
What if one of the two stocks crashes hard?

I mean it's supposed to be a 35bil deal so that price has to be locked in somehow.

Stock Swap Definition

A stock swap is the exchange of one equity-based asset for another.

www.investopedia.com

Example of a Stock Swap

In 2017, the Dow Chemical Company ("Dow") and E.I. du Pont de Nemours & Company ("DuPont") closed a merger where Dow shareholders received a swap ratio of 1.00 share of DowDuPont (the combined entity) for each Dow share, and DuPont shareholders received a swap ratio of 1.282 shares of DowDuPont for each DuPont share.

Note that in the case of an all-stock deal, after the swap ratio terms have been agreed upon, the stock price of the target company will fluctuate in value roughly according to the stock swap ratio.

Also, for the shareholders of the target company, the IRS does not consider the original investment as a "disposal" for tax purposes when the company is taken over. No gain or loss needs to be reported at deal closing. The cost basis for shareholders of the merged company will be the same as the original investment.

It should be obvious why share prices of one company would not collapse immediately before a stock swap.

 

[TheELF]

It should be obvious why share prices of one company would not collapse immediately before a stock swap.

I didn't mean afterwards but before, if they swap at whatever the current rate is at the moment of the swap it could change the amount of that 35 bil by a huge amount both ways.
Also the article you link to says this:

During a stock swap, each company's shares must be accurately valued in order to determine a fair swap ratio between the two shares.

Also known as a stock-for-stock deal, an acquiring company's stock is exchanged for the stock of the acquired company at a predetermined rate.

Is "during" the point at which they made the initial calculation, so they are going to swap stock not at current prices but at the prices they pre agreed upon (so 35bil) , or is "during" the moment the swapping is going to take place (so whatever it is going to be at that point) .

 

[DrMrLordX]

I didn't mean afterwards but before, if they swap at whatever the current rate is at the moment of the swap it could change the amount of that 35 bil by a huge amount both ways.

. . .

The stock swap has a ratio. In this case, it's 1.7234 shares of AMD stock per share of Xilinx stock. The price of AMD stock will normalize around Xilinx shares and vice versa according to normal trading activity, especially in the final days before the merger. In other words, if AMD is overpriced according to this ratio, AMD shareholders will likely sell to acquire Xilinx shares. Were Xilinx overpriced according to the ratio, then Xilinx holders would sell in order to acquire AMD shares. It won't be perfect but the gain/loss from holding one share or the other should be minimized by the time of the merger.

Is "during" the point at which they made the initial calculation

Yes

 

[turtile]

It matters if what they think is going to affect immediate and near-future pricing on the stock.

fwiw the merger will probably not be struck down at all. It's just that, currently, AMD stock maybe a bit overvalued and that the market will likely correct before the stock swap takes place. If you held AMD stock valued at $120 and you knew that Xilinx owners would effectively be getting AMD sold to them at $105, would you not swap your shares for Xilinx shares first?

The reason we have heard for the spread is the fears that China will not approve the merger due to US/China relations. In reality, there doesn't seem to be any evidence to suggest that they will block the merger. Wall Street isn't rational. There are many traders that have swapped to Xilinx but there are also many large investment funds purchasing AMD.

Wall Street also believed AMD was going bankrupt. They were wrong. Yes, in the short term, the stock price dropped heavily from shorting. But long term, an investor who went long was rewarded heavily. Ideally, buying in on the uptrend would have been the most rewarding.

I also believe that the merger process has a risk to be bumpy but it shouldn't worry long-term investors.

Wait, so the actual price at which the stock exchanging will be done wasn't set at all?
Then how are they going to settle on the time it will take place? I mean AMD could wait for a time their stock is at the lowest or xilinx for the time AMD stock is at the highest, how will they decide?
What if one of the two stocks crashes hard?

I mean it's supposed to be a 35bil deal so that price has to be locked in somehow.

The timing isn't up to AMD. It's whenever China approves the merger. They also must announce it immediately so there is not any insider trading.

 

[DrMrLordX]

I also believe that the merger process has a risk to be bumpy but it shouldn't worry long-term investors.

It's worth it to remember that "Wall Street" is many different entities. The net effect of their trading activity ultimately determines what Wall Street "likes" or "dislikes". And some of the acquisitions of AMD shares in the past few months may have had nothing to do with the Xilinx merger. If the merger fails, that could easily drive down prices just as easily as sellers dropping AMD in favor of Xilinx to exploit the swap ratio.

Those stubbornly buying up AMD shares (rather than Xilinx shares) offer the rest of the market an excellent opportunity to squeeze out some free money.

 

[amrnuke]

Wait, so the actual price at which the stock exchanging will be done wasn't set at all?
Then how are they going to settle on the time it will take place? I mean AMD could wait for a time their stock is at the lowest or xilinx for the time AMD stock is at the highest, how will they decide?
What if one of the two stocks crashes hard?

I mean it's supposed to be a 35bil deal so that price has to be locked in somehow.

AMD isn't "paying" any amount of money. The deal was the XLNX shareholders agreed to sell each share of XLNX for 1.7234 AMD shares. That's it. The actual dollar value of said issued stocks is interesting but irrelevant to the deal. Yes, in the books, the "amount paid" will be written somewhere for accounting purposes. But AMD isn't "paying" $35 billion or $51.3 billion in formal cash or debt in this acquisition. They're issuing ~425-430M AMD shares to current Xilinx owners in return for Xilinx the company.

Assuming earnings and market cap remain the same, the new company share price would presumptively be $116 or so, meaning each current AMD shareholder would lose $4 per share with this acquisition (I'm sure the price will be adjusted by the market accordingly prior to the deal closing), and the EPS would drop from $3.28 with AMD now to ~$2.90 with the post-merge company, due to the issuance of more shares (because the new company will have ~180M more outstanding shares than AMD+XLNX have now, dilutional effect).

 

[DisEnchantment]

Approved by China, probably some smaller refiling perhaps, but AMD+Xilinx will be a reality

市场监管总局关于附加限制性条件批准超威半导体公司收购赛灵思公司股权案反垄断审查决定的公告

市场监管总局关于附加限制性条件批准超威半导体公司收购赛灵思公司股权案反垄断审查决定的公告

www.samr.gov.cn

Machine translated

After review, the State Administration for Market Regulation decided to approve the concentration of undertakings with additional restrictive conditions.

China conditionally approves AMD's $35 bln deal for Xilinx

China's market regulator has conditionally approved semiconductor group Advanced Micro Devices Inc's $35 billion all-stock deal for peer Xilinx , it said on Thursday.

www.reuters.com

Update:
Refiling was already done, expected closure date 11:59 pm, 9th Feb 2022

https://ir.amd.com/sec-filings/content/0000002488-22-000009/0000002488-22-000009.pdf

 

[Kedas]

AMD (Xilinx) vs Intel (Altera)
Let the games begin

 

[scineram]

AMD (Xilinx) vs Intel (Altera)
Let the games begin

So far it’s ownage. Xilinx record quarter, first above a billion.

 

[positivedoppler]

Approved by China, probably some smaller refiling perhaps, but AMD+Xilinx will be a reality

市场监管总局关于附加限制性条件批准超威半导体公司收购赛灵思公司股权案反垄断审查决定的公告

市场监管总局关于附加限制性条件批准超威半导体公司收购赛灵思公司股权案反垄断审查决定的公告

www.samr.gov.cn

Machine translated

China conditionally approves AMD's $35 bln deal for Xilinx

China's market regulator has conditionally approved semiconductor group Advanced Micro Devices Inc's $35 billion all-stock deal for peer Xilinx , it said on Thursday.

www.reuters.com

Update:
Refiling was already done, expected closure date 11:59 pm, 9th Feb 2022

https://ir.amd.com/sec-filings/content/0000002488-22-000009/0000002488-22-000009.pdf

What I don't understand is why two US companies need permission from the Chinese

 

[scannall]

What I don't understand is why two US companies need permission from the Chinese

To do business in China. Simple as that.

 

[Hitman928]

What I don't understand is why two US companies need permission from the Chinese

They technically don't, but most likely China would impose bans/restrictions on the purchase/sell/manufacturing of their products in China which would be pretty painful to the U.S. companies.

 

[Vattila]

If you all allow me to toot my horn on this happy occasion for me as an AMD shareholder, I predicted the AMD+Xilinx merger back in 2019:

Looking ahead, could Xilinx be Lisa Su's first big acquisition?

AMD, ARMH, INTC, NVDA Message Board - Msg: 32246620 (siliconinvestor.com)

As I noted in my linked post, AMD CEO Lisa Su had started including FPGAs on her slides, showing her vision for High Performance Computing, and I had noticed that Xilinx CEO Victor Peng was an ex-AMD employee. They are a pretty enthusiastic tech couple. It will be really interesting to see what they have planned and what they can achieve together.

Here is a taste of AMD+Xilinx co-development on the software side:



Xilinx + AMD Converged Runtime Technology Demonstration (@SuperComputing 2020)

[AMD's framework] ROCm is based on the ROCm runtime environment, which can be targeted with code written in HIP C++, and of course many libraries are supported. This allows uniform code to be written to the platform itself, which can be run by either the CPU or the GPU, possibly even simultaneously. [Xilinx's framework] Vitis is similar in principle, there the base is an XRT runtime environment, which can be accessed through the XRT API, but an important factor is to write the appropriate device-level code, which must be done separately. The latter is required for the program to run on not only CPUs but also FPGAs.

The two companies are changing that, and support for Xilinx accelerators is merging into the ROCm software package. This will be done in two parts. First of all, HIP C++ interoperability is used, so the ROCm runtime environment is supplemented with enough to be able to interpret HLS C or AIE C code written for Xilinx devices. With this, the rest of the Vitis environment disappears, as ROCm can replace them and even offer more than them. With this new implementation, Xilinx accelerators will be programmable through the ROCm package, but they will not have unified virtual memory or various other unification capabilities, so on the development side you will definitely need HIP C++ code that can run on the CPU and on the GPU, and also to an alternative HLS C or AIE C code base for Xilinx hardware. By the way, this software package is already in full swing, larger customers can even test it.

The next step will be full unification. This eliminates the need for device-specific code for Xilinx accelerators, meaning that they will also be programmable via HIP C++, just like CPUs and GPUs. At the same time, support for unified virtual memory will be introduced, meaning that all components will see and manage the available memory areas coherently, and the same libraries and frameworks will be available for each piece of hardware.

tekdeeps.com

PS. Note that AMD's HIP C++ is effectively and intentionally a dialect of Nvidia's CUDA C++.

 

[DisEnchantment]

As I noted in my linked post, AMD CEO Lisa Su had started including FPGAs on her slides, showing her vision for High Performance Computing, and I had noted that Xilinx CEO Victor Peng was an ex-AMD employee. They are a pretty enthusiastic tech couple. It will be really interesting to see what they have planned and what they can achieve together.

This one got a lot of mentions by AMD.

In one interview with Emily Chang she is hinting about growth through further acquisitions.
Smaller software specialists, specialty AI, networking, low power SoCs etc operating only in US or EU would be better for quick integration, provided they have good patent pool.
Xilinx has a small patent pool which is quite miniscule compared to AMD's.

 

[Vattila]

Smaller software specialists, specialty AI, networking, low power SoCs etc operating only in US or EU would be better for quick integration, provided they have good patent pool.
Xilinx has a small patent pool which is quite miniscule compared to AMD's.

Although there were other potential smaller takeover targets for gaining FPGA capability — such as Lattice Semiconductor (where, by the way, another ex-AMD executive Jim Anderson is currently serving, and doing very well, as CEO) — I guess the primary lure of Xilinx is their leadership and dominant presence in the FPGA field. Also, Xilinx is very profitable, making the deal accretive to AMD from the get go. Apparently, as disclosed after announcing the deal, Lisa Su and Victor Peng have had discussions going back to 2017, I think (correct me if I am wrong). At the time of my prediction in 2019, AMD and Xilinx were about the same size in terms of market capitalisation, around $30B. Hence I wondered about how a merger would be achieved. AMD has of course grown tremendously since then, with a current valuation of $127B, and is thus able to pay for the takeover in shares.

AMD and Xilinx have a long history of co-development on the software side. An example, of my particular interest, is the triSYCL research project, which was started at AMD and currently is lead by Xilinx, and which hopefully is indicative of future work on support for the SYCL standard — see my discussion thread "SYCL will replace CUDA".

PS. On Victor Peng, here is a nice article on his achievements since becoming Xilinx CEO in 2018:

A company transformed

Appointed CEO back in 2018 Victor Peng has taken Xilinx through a transformative period, the next stage of which is its acquisition by AMD. Neil Tyler gets an update.

www.newelectronics.co.uk

 

[DisEnchantment]

Although there were other potential smaller takeover targets for gaining FPGA capability — such as Lattice Semiconductor (where, by the way, another ex-AMD executive Jim Anderson is currently serving, and doing very well, as CEO) — I guess the primary lure of Xilinx is their leadership and dominant presence in the FPGA field. Also, Xilinx is very profitable, making the deal accretive to AMD from the get go. Apparently, as disclosed after announcing the deal, Lisa Su and Victor Peng have had discussions going back to 2017, I think (correct me if I am wrong). At the time of my prediction in 2019, AMD and Xilinx were about the same size in terms of market capitalisation, around $30B. Hence I wondered about how a merger would be achieved. AMD has of course grown tremendously since then, with a current valuation of $127B, and is thus able to pay for the takeover in shares.

AMD and Xilinx have a long history of co-development on the software side. An example, of my particular interest, is the triSYCL research project, which was started at AMD and currently is lead by Xilinx, and which hopefully is indicative of future work on support for the SYCL standard — see my discussion thread "SYCL will replace CUDA".

PS. On Victor Peng, here is a nice article on his achievements since becoming Xilinx CEO in 2018:

A company transformed

Appointed CEO back in 2018 Victor Peng has taken Xilinx through a transformative period, the next stage of which is its acquisition by AMD. Neil Tyler gets an update.

www.newelectronics.co.uk

You missed one sentence when you quoted me which alters the context. I was talking about future acquisitions.
Xilinx is a good acquisition (SmartNICs , programmable ML HW etc,. are big ticket items) albeit with a small patent pool.
What I was trying to say is that Lisa hinted at future acquisitions to grow in newer markets, but another Xilinx like consolidation is going to be increasingly difficult due to the current political developments.
On the other hand, smaller companies with good IP pools should be easier to absorb and quick to integrate as long as they operate in EU/US but AMD should do their homework/research and avoid buying snake oil peddlers.

 

[Vattila]

Great news! The merger is stipulated for Valentine's Day:

"AMD (NASDAQ: AMD) today announced that it has received approval from all necessary authorities to proceed with the acquisition of Xilinx, Inc. (NASDAQ: XLNX). With the exception of the remaining customary closing conditions, all conditions to the transaction closing have been satisfied and the company expects the transaction to close on or about February 14, 2022."

AMD Receives All Necessary Approvals for Proposed Acquisition of Xilinx :: Advanced Micro Devices, Inc. (AMD)

PS. Hopefully, we'll see rapid acceleration now on common software effort, in particular support for heterogenous programming models and SYCL in particular.

Speculation: SYCL will replace CUDA

I've had my eye on SYCL for some time now. I don't think it is widely recognised that this is going to be the standard for heterogeneous programming models going forward (eventually to be merged into the ISO C++ standard, it seems). If you read any discussion on programming, tech and investor...

forums.anandtech.com

 

[ThatBuzzkiller]

If anything AMD acquiring Xilinx means death for SYCL. Xilinx will likely drop all of their future plans on working with SYCL and will start doubling down by working on their proprietary ROCm stack. There's no corporate interest in making SYCL succeed by offering interop with FPGAs when corporate strategy dictates that it should be ROCm that succeeds so all of their FPGA interop work will likely move to be in line with that vision ...

No big players are interested in offering a working interoperable industry standard and that includes AMD as well ...