AMD quer comprar a Xilinx por 30 mil milhões

[Torak]

AMD reportedly in talks to acquire Xilinx for $30 billion

The Wall Street Journal notes in its report that discussions between the two companies regarding a takeover have stalled in the past, so there’s no guarantee we’ll see AMD acquiring Xilinx. If it does happen, though, it would be another massive deal in the consolidating chip industry—Nvidia is in the process of buying chip designer ARM for $40 billion.


Xilinx has been focusing heavily on the data center in recent years. A buyout by AMD would give team red an arsenal of new products to compete against Nvidia and its data center GPUs. Xilinx’s Versla Premium line—its next-gen 7nm Adaptive Compute Acceleration Platform (ACAPs) products—is designed for high bandwidth networks and boasts features including 123TB/s bandwidth across its network-on-chip (NOC) and PCIe 5.0 support.

Xilinx products are found beyond the data center. They’re also used in wired & wireless communications, AI, Machine Learning, automotive, industrial, consumer, aerospace, and defense industries.

A Xilinx acquisition would help AMD in its battle against Intel. Back in 2015, Chipzilla acquired Xilinx rival Altera for $16.7 billion, integrating the firm into its Programmable Solutions Group (PSG) and acquiring its FPGA line. Intel will have been watching the Ryzen 5000 reveal last night, which included “the fastest gaming CPU,” and AMD could soon be giving its rival new headaches in other markets.

Wallstreet Journal - AMD Is in Advanced Talks to Buy Xilinx

Isto é muito dinheiro para a AMD.
Mas seria um passo enorme para a sua presença em servidores.

 

[Nemesis11]

Sendo o WSJ, este rumor é bem credível.

Link para o artigo completo: https://www.wsj.com/articles/amd-is-in-advanced-talks-to-buy-xilinx-11602205553?mod=djemalertNEWS

Pelo que encontro, a Xilinx tem +/- 50% do mercado de FPGAs, enquanto a Intel, que comprou a Altera por 16,7 mil milhões há uns anos atrás, tem +/- 35% de market share.

30 mil milhões é mesmo muito dinheiro, mas o market cap actual da AMD já é superior a 100 mil milhões. Em menos de 5 anos, o market cap da AMD disparou a pique.

Além disso, o market cap da Xilinx é de 26 mil milhões. 30 mil milhões parece ser razoável.

O mercado tem-se vindo a consolidar cada vez mais em grandes grupos e parece-me que pode ser bom negócio tanto para a AMD como para a Xilinx.

 

[miguelbazil]

Estava referido no tópico de Zen 3 que as acções da AMD poderiam estar a cair por causa desta potencial compra. Qual é a razão para algo esta compra provocar uma queda de valor?

 

[Torak]

Talvez o preço muito elevado.
Quando a AMD comprou a ATI, por um preço acima daquilo que ela valia, trouxe muitos problemas financeiros para a AMD.

 

[miguelbazil]

Hmmmm e o facto da Xilinx ter um cap tão elevado nos FPGA's, na ordem dos 26 mil milhões como referido antes, não deveria dar solidez a este negócio? A expectativa é que seja como a ATI foi? Estou apenas curioso de que tipo de análise se faz para isto acontecer

 

[Torak]

A AMD não tem bolsos muito fundos e uma aquisição deste tamanho deixa sempre os investidores nervosos.
Creio que quando a nVidia comprou a ARM, as ações também baixaram um pouco.

 

[Dark Kaeser]

AMD Reported to Be Negotiating Purchase of Xilinx

The AMD-Xilinx combination is much more reminiscent of Intel’s purchase of Altera in 2015 for approximately $16.7 billion, however. It would be another instance of a leading provider of CPUs buying a specialist in FPGAs, and for similar reasons — to gain control of a complementary chip technology important in data centers in general, and with artificial intelligence (AI) workloads specifically. FPGAs are often used for AI acceleration.

FPGAs are used for AI acceleration, but so are GPUs — which brings us back to GPU specialist Nvidia. Nvidia has been explicit that it expects to become a stronger competitor in the AI and data center markets by combining its technology with Arm’s.

AMD taking control of Xilinx appears to be a way of keeping up with companies that are rivals in so many market segments.
Intel, AMD, Nvidia, and Arm are also among the leading vendors in the high performance computing (HPC) market as well.

https://www.eetimes.com/amd-reported-to-be-negotiating-purchase-of-xilinx/

Há ainda a curiosidade de o actual CEO da Xilinx, Victor Peng, ser um ex-ATI

Peng previously served as Corporate Vice President of the graphics products group (GPG) silicon engineering with AMD and was a leader for AMD's central silicon engineering team supporting graphics, console game products, CPU chipset and consumer business units. Prior to that, Peng held executive and engineering leadership roles at TZero Technologies, MIPS Technologies, SGI and Digital Equipment Corp.

https://www.xilinx.com/about/management-team/victor-peng.html

Mas a aquisição da XIlinx pela AMD faz mais sentido que a compra da Altera pela Intel, na altura.


A AMD já tinha uma estratégia (Fusion/HSA) para integrar "cenas" no die e desenvolver software para integrar isso tudo (ROCm e HIP), a Intel só há cerca de 2 anos é que fez o mesmo com a OneAPI.


‎10-02-2018

At today’s Xilinx Developer Forum in San Jose, Calif., our CEO, Victor Peng was joined by the AMD CTO Mark Papermaster

AMD and Xilinx have shared a common vision around the evolution of computing to heterogeneous system architecture and have a long history of technical collaboration. Both companies have optimized drivers and tuned the performance for interoperability between AMD EPYC CPUs with Xilinx FPGAs. We are also collaborating with others in the industry on cache coherent interconnect for accelerators (the CCIX Consortium – pronounced “see-six”), focused on enabling cache coherency and shared memory across multiple processors.

Xilinx and AMD see a bright future in their technology collaboration. There is strong alignment in our roadmaps that align the high-performance AMD EPYC server and graphics processors with Xilinx acceleration platforms across its Alveo accelerator cards, as well as its forthcoming Versal portfolio.

https://forums.xilinx.com/t5/Xilinx...ilinx-announce-a-new-world-record/ba-p/895034

Acho que era este sistema
https://www.amd.com/system/files/documents/xilinx-deep-learning-solution-on-amd-epyc-processors.pdf

Mas até havia parcerias entre ambos em 2009, desenvolver "Cache Coherency" para o HyperTrasnport, na altura em que isso ainda nem era suportado.
http://ra.ziti.uni-heidelberg.de/coeht/pages/events/20090211/parag_beeraka.pdf

Com a ajuda da Universidade de Heidelberg, a mesma que agora tem uma parceria com a Intel para o OneAPI.

Ambos faziam também parte do consórcio CCIX, junto com a ARM e a Mellanox
https://www.ccixconsortium.com/about/members/

 

[akgis]

Acho que é uma reação a futura obsolescência do x86 la para o final desta decada, e as empresas do ramo anda a fazer casais.

Intel+Altera já casados
Nvidia+ARM já pediu a mão em casamento
AMD esta a fazer olhinhos á Xilinx

 

[Dark Kaeser]

HPC In 2020: Acquisitions And Mergers As The New Normal
January 16, 2020

Since Intel acquired Altera in 2015, Xilinx is the only standalone company of any size that makes reconfigurable logic devices. Give that, the natural buyer for Xilinx would be AMD, Intel’s arch-nemesis. AMD, of course, already has a highly competitive lineup of CPUs and GPUs to challenge its much larger rival, and the addition of an FPGA portfolio would open a third front. It would also provide AMD entry into a whole array of new application markets where FPGAs operate: ASIC prototyping, IoT, embedded aerospace/automotive, 5G communications, AI inference, database acceleration, and computational storage, to name a few.

https://www.nextplatform.com/2020/01/16/hpc-in-2020-acquisitions-and-mergers-as-the-new-normal/

The only problem is that Xilinx’s current market cap of around $25 billion, or about half the current market cap of AMD. And if you’re wondering about AMD’s piggy bank, the chipmaker has $1.2 billion cash on hand as of September 2019.

este último quote reflecte a situação na altura. Entretanto alterada (as reservas em dinehiro são agora 1.8B$ a dívida total cerca de 0.8B$)

Pondering That Rumored $30 Billion AMD Acquisition Of Xilinx

You can have a strategy, but you can’t buy one and you can’t sell one, either. But you can mimic one, perhaps, and we wonder what AMD is seeing in the FPGA market that Intel clearly saw when it shelled out $16.7 billion in June 2015

There is no reason to question why AMD is considering any acquisition at this moment. By gaining around 10 percent market share in server CPUs with its Epyc processor line, by having a very credible lineup against Intel in PCs, by owning the game console chip business for second generation, and by running the table along with Cray (now a unit of Hewlett Packard Enterprise) on exascale systems in the United States for its CPU-GPU hybrid architecture, AMD is definitely at a local maxima in terms of its perceptions in the market and therefore its market capitalization is huge, even after taking a 4 percent or so hit from rumors of the Xilinx deal. Even with that hit, AMD’s market capitalization is $101.6 billion, a huge increase compared to the low of $1.45 billion set in August 2015, when AMD’s strategy was in question and president and chief executive officer Lisa Su took control and got AMD back on track.

But if AMD is taking a run at Xilinx, and putting $30 billion on the table to scare off potential rivals, then AMD wants something else more than an FPGA business. It wants an ecosystem and a partner channel and a customer base, something that upstarts like Achronix and Innovium have yet to fully develop.

While all of this sounds interesting what AMD needs is a coherent software development environment, something that is broader than the Radeon Open Compute (ROCm) hybrid CPU-GPU environment that AMD created several years ago and improved radically in recent years. ROCm now offers a CUDA runtime so applications compiled to offload to Nvidia GPU accelerators can run on hybrid systems using AMD GPUs, and improving this software is a key aspect of the “Frontier” exascale system that AMD is working with HPE/Cray to build for Oak Ridge National Laboratory. But what AMD probably needs if it is indeed going to add Xilinx to its fold is something that looks and smells like Intel’s oneAPI environment, which allows data parallel C++ code to be written once and compiled for CPUs, GPUs, or FPGAs as the system dictates. The Xilinx Vitis environment compiles C and C++ code with OpenCL hooks to communicate to the FPGA accelerator and offload parallel routines to it.

https://www.nextplatform.com/2020/10/09/pondering-that-rumored-30-billion-amd-acquisition-of-xilinx/

 

[Dark Kaeser]

Mais 2 artigos de 2 autores diferentes

- AMD-Xilinx Deal: Bringing The Fight To The Data Center

Is AMD trying to build a complete data center computing platform, similar to what Nvidia is trying to do with Arm? Nvidia GPU + Arm CPU + Nvidia DPU is a compelling combination for building a complete data centre offering. AMD’s platform would be AMD CPU + AMD GPU + Xilinx FPGA + Xilinx SmartNIC. Will this recipe be tasty enough to beat Nvidia’s, given Nvidia’s head start?

https://www.eetimes.com/amd-xilinx-deal-bringing-the-fight-to-the-data-center/

- AMD and Xilinx: A Match Made in Silicon Valley?

First, it was not known AMD was in acquisition mode and certainly not for a company as large as Xilinx and not for so much money. The deal to buy Xilinx is rumored to cost AMD $30 billion, which is almost a third of its market capitalization.

On top of that, Xilinx was not known to be ready to be acquired, as their new CEO was in the midst of transforming the business with the Versal products and extending their drive into the data center, AI, and telecom markets.

AMD-Xilinx connections
It should be noted that the two companies do not have much in the way of overlapping product lines so there is not a significant chance that it would trigger any anti-competitive investigations. And for those who are interested in a little history, it is interesting to note that back in 1988 AMD became a second source for Xilinx FPGAs after it bought MMI (Monolithic Memories, Inc.). Later AMD discontinued the FPGA product line after coming to an agreement with Xilinx which gave AMD stock in Xilinx, that it later sold. There is another noteworthy connection between the two companies – Xilinx CEO Victor Peng Is a former AMD executive.

What does Xilinx bring to AMD?
While Xilinx is making steady progress in the data center market, it still represents only 12% of the company’s revenue, as seen in its latest financial report (below). The company’s Aerospace & Defense, Industrial and Test & Measurement (AIT) division represents 45% of revenue, which includes aerospace, defense, industrial, scientific, medical, test, measurement, and emulation. AMD has a weak or non-existent market presence in these markets and would be complementary if the companies were to merge. The wired and wireless group (WWG) represents 32% of Xilinx revenue and could also be a great opportunity for AMD to move its CPUs into this market to compete with Intel, Qualcomm and Nvidia.

Probably the strongest case that can be made for the AMD- Xilinx combination is that Lisa Su is known to be a conservative and smart executive, and if this deal makes sense to her, many in the industry would give her the benefit of the doubt. The two companies are also major customers of Taiwan Semiconductor Manufacturing Company (TSMC) and this might give the combination a stronger negotiating position with the foundry.

https://www.eetimes.com/amd-and-xilinx-a-match-made-in-silicon-valley/

 

[muddymind]

Primeiro a intel compra a Altera e agora a AMD a Xilinx. Fica alguma empresa de peso em FPGAs independente neste mercado? Sempre tive a ideia que o mercado de FPGAs era um pouco duopólio entre essas duas empresas.

 

[Dark Kaeser]

Tens uma tabela acima, postada pelo Nemesis11, ha pelo menos 3 (na realidade mais uma ou outra), mas a XIlinx é claramente a nº 1, daí o preço a pagar face aos 17B$ que a Intel pagou pela Altera.

- Microsemi (ou Microchip)
- Lattice Semi
- Achronix

De notar que este negócio a concretizar-se traria não apenas os FPGA mas também a parte de Network & Wireless.
https://www.xilinx.com/applications/data-center/network-acceleration.html
https://www.xilinx.com/applications/wired-wireless/wireless.html

 

[muddymind]

Não tinha visto a tabela efectivamente. Mas assim confirma-se que essas duas tinham praticamente o mercado todo e as restantes apenas "migalhas".

 

[Dark Kaeser]

Mas até é algo relativamente recente, acho que no caso da Altera "apostaram" no cavalo errado quando escolheram a Intel para foundry (14nm).
Em 2010 o cenário era este:

com a curiosidade da Vantis (7% do mercado) ser uma spin-off da... AMD

Vantis – A spin-out of the AMD product division, this was another early entrant into the market. In April, 1999 Lattice acquired Vantis from parent company AMD for $500 Million.

https://sourcetech411.com/top-fpga-companies-for-2013/

 

[miguelbazil]

Em 2010, ninguém conseguia prever que a Intel acabaria com estes problemas que tem tido, e a AMD a subir. A aposta fez todo o sentido na altura, sinceramente.

 

[JPgod]

Ehh, vamos ver se isso não é uma nova ATI, que mandou a AMD quase pro buraco negro.

 

[nullus]

Agora é mesmo oficial, press release da AMD:


https://www.amd.com/en/press-releas...ing-the-industry-s-high-performance-computing

 

[Nemesis11]

35 mil mihões e mais 5 mil empregados, quase metade do que a AMD tem actualmente.

Interessante esta parte:

AMD reported last quarter that it had broken above the 10% market share in Enterprise with its EPYC product lines, and today’s earnings call is also expected to see growth. AMD is already reporting revenue up +56% year on year company-wide, with +116% in the Enterprise, Embedded, and Semi-Custom markets.

É hoje que a AMD revela os resultados trimestrais. Aliás, não sei se já foi se ainda irá ser. O crescimento no mercado empresarial é importante, porque é lá que estão as margens maiores.

 

[Torak]

35 mil mihões e mais 5 mil empregados, quase metade do que a AMD tem actualmente.

Interessante esta parte:


É hoje que a AMD revela os resultados trimestrais. Aliás, não sei se já foi se ainda irá ser. O crescimento no mercado empresarial é importante, porque é lá que estão as margens maiores.

Já revelou e parecem ser bons.

Q3 2020 Results

• Revenue was $2.80 billion, up 56 percent year-over-year and 45 percent quarter-over-quarter driven by higher revenue in both the Enterprise, Embedded and Semi-Custom and Computing and Graphics segments.
• Gross margin was 44 percent, up 1 percentage point year-over-year and flat quarter-over-quarter. The year-over-year increase was primarily driven by EPYC™ and Ryzen™ processor sales. Gross margin was flat quarter-over-quarter as an increase of Ryzen and EPYC processor sales was offset by a higher percentage of semi-custom revenue.
• Operating income was $449 million compared to $186 million a year ago and $173 million in the prior quarter. Non-GAAP operating income was $525 million compared to $240 million a year ago and $233 million in the prior quarter. Operating income improvements were primarily driven by revenue growth, including an increase in Ryzen and EPYC processor sales and semi-custom product sales.
• Net income was $390 million compared to $120 million a year ago and $157 million in the prior quarter. Non-GAAP net income was $501 million compared to $219 million a year ago and $216 million in the prior quarter.
• Diluted earnings per share was $0.32 compared to $0.11 a year ago and $0.13 in the prior quarter. Non-GAAP diluted earnings per share was $0.41 compared to $0.18 a year ago and $0.18 in the prior quarter.
• Cash, cash equivalents and short-term investments were $1.77 billion at the end of the quarter.

Quarterly Financial Segment Summary

• Computing and Graphics segment revenue was $1.67 billion, up 31 percent year-over-year and 22 percent quarter-over-quarter. Revenue was higher year-over-year driven by a significant increase in Ryzen processor sales partially offset by lower graphics revenue. Revenue was higher quarter-over-quarter driven by a significant increase in Ryzen processor sales and an increase in graphics revenue.
  • Client processor average selling price (ASP) was lower year-over-year due to a higher mix of mobile processor sales. Client processor ASP was higher quarter-over-quarter driven by higher desktop and mobile processor ASPs.
  • GPU ASP was lower year-over-year due to product cycle timing and higher quarter-over-quarter due to product mix.
  • Operating income was $384 million compared to $179 million a year ago and $200 million in the prior quarter. The year-over-year and quarter-over-quarter increases were driven by higher Ryzen revenue.
• Enterprise, Embedded and Semi-Custom segment revenue was $1.13 billion, up 116 percent year-over-year and 101 percent quarter-over-quarter. Revenue was higher year-over-year and quarter-over-quarter due to higher semi-custom product sales and increased EPYC processor sales.
  • Operating income was $141 million compared to $61 million a year ago and $33 million in the prior quarter. The year-over-year and quarter-over-quarter increases were primarily driven by higher revenue.
• All Other operating loss was $76 million compared to operating losses of $54 million a year ago and $60 million in the prior quarter.

 

[Nemesis11]

Interessante as declarações deles sobre o Epyc Zen 3. Shipping este quarto para clientes Cloud e HPC, shipping para OEMs no primeiro quarto de 2021 e não falam de quando estará disponível em geral, nem deram datas para o anúncio e/ou lançamento.

Os resultados também são curiosos. O net income disparou, não só em relação ao quarto igual do ano passado, como em relação ao quarto anterior deste ano. No entanto, as margens só cresceram 1% e a AMD culpa as margens muito baixas dos SOCs das consolas. Vamos ver se a AMD não se vai querer afastar deste mercado no futuro.

Quanto á compra da Xilinx, espero que a AMD não tenha dado um passo maior que a perna. Pagou um preço mais elevado que o esperado e vai integrar um empresa muito grande relativamente ao seu tamanho.
Apesar dos bons resultados, as ações da AMD estão a cair mais de 5%. Acho que o mercado não está a ver com bons olhos esta compra.