Made at Intel. Acquisitions – 2
In 1996 Oleg Syutin created the Nizhny Novgorod Software Technologies Lab (NSTL), the core of which was people from Sarov. It began to develop projects related to video and audio compression, graphics, computer vision, vector mathematics, program performance analysis, and many others. NSTL started out as a typical outsourcing company, taking orders from different firms, but stayed in this status for a very short time. In 2000, it became part of Intel. It was the so-called hiring action – NSTL employees who worked on Intel projects received “blue badges”. Immediately after that, many Americans came to us. Loading into the corporate culture has begun. Trainings were almost daily, they explained to us what we needed and what not to do, how to properly position ourselves in the company. I was always amazed by their enthusiasm, some kind of missionary work that lives inside everyone. In fact, the corporate culture of Intel is not an easy thing, and we fully immersed ourselves in it (integrated) only after 5-7 years. Played a big role here Alexey Vasilievich Odinokov is a Russian who worked for Intel in the USA.
He understood both the Western mentality and the Russian one and proved to be an important link. He deserves special credit for the fact that the Nizhny Novgorod site got on its feet and took its place in the corporation. Well, personally, he had a great influence on me at that time.
Another important factor was that NSTL employed mostly young people.
(The photo shows those who started OpenCV 25 years ago, from left to right – your humble, Sergey Molinov, Gary BradskySergey Oblomov, Vadim Pisarevsky, Stas Bratanov). We were in our 20s at the time and we were just starting our career path, so we didn’t have to retrain, but learning was natural. As explained, so we worked, without any internal resistance. He listened, tried to understand – and dug into his code, which he wrote. That is, we had no idea how the processes in a large corporation should be arranged, which means we didn’t have to break anything. In many ways, the acquisition of NSTL turned out to be quite successful. For 20 years, the Nizhny Novgorod site has grown more than 20 times from 50 people in 2000 to more than 1000 in 2021.
Intel “legitimized its relationship” with the Sarov laboratory (STL) in 2003 (before that, work was done on a contract basis). There, people were mainly engaged in compilers and mathematical libraries – MKL and IPP. The integration into Intel was much more difficult for the guys in Sarov than for the people from Nizhny Novgorod. For a number of reasons – foreign guests visited there much less frequently than in NN due to difficult logistics, although Intel moved its office to the Technopark outside the “thorn”. There was no one in Sarov The same Alexei, which would help to understand the intricacies of the American mentality. But the main thing is that the Sarov team consisted for the most part of already established engineers with a VNIIEF past behind them. And the transition from the regime “red” office to the “white” was very difficult for them. How different these worlds are and how fragile the transition from one to another is accompanied, I felt much later – when I moved from Intel to Huawei at the age of 43.
One way or another, if the population of the Nizhny Novgorod office of Intel was constantly growing, then the Sarov office, on the contrary, was steadily declining. This led to the closure of the site in 2011. And I will touch on this topic in the Cursed Days chapter on various cutoffs, downsizing and office closures. And the decline of the “cradle of Russian Intel” was accelerated by one event that radically redrawn the entire landscape and many human destinies.
If the acquisition of NSTL and STL was a typical acquisition of competencies, then the deal with Elbrus and Unipro completed in 2004 had, in my opinion, a much more serious business background.
This project was called Bear – well, how else can amers call a big (about 700 people) deal in Russia? 😊 Hiring action of Elbrus and Unipro employees is certainly a milestone event for the Russian division of Intel. However, I still don’t have a definitive answer to the question of why Intel needed this deal. There was an interesting story there. The fact is that the initiator of the purchase was Mike Fister (Captain Itanic) – the man who at that time was at the head of the Intel server business. Perhaps for Fister, buying Elbrus’ competencies in the field of compilation and binary translation for the VLIW / EPIC architecture was an attempt to breathe life into his brainchild Itanium. And perhaps it would have been so if he had become the CEO of Intel. However, preference was given to Paul Otellini, who replaced Craig Barrett. And before the deal was finalized, Fister left Intel to become CEO of Cadence Design Systems.
There is, however, another theory that also deserves attention. Back then, Intel was competing with Sun Microsystems and IBM for share of the server market. And there, Java, which was then Sun proprietary IP, was becoming more widespread. Of course, everyone did not like it terribly – not only to us, but also to IBM, HP and other big players in the IT market. It was highly desirable that Java become Open Source, but there was no leverage on Sun, and financially they were quite strong. However, there was still a weak spot. The fact is that a serious part of Java development was performed by Elbrus and Unipro under a contract with Sun. And so Intel “trip” Sun by making an offer to buy out these offices. Sun tried to compete, but as far as I understand, only for the St. Petersburg team. The lion’s share of Elbrus/Unipro’s assets goes to Intel practically without a fight. The latter initiates the Harmony project, which in theory was the same Java only in Open Source. The good news is that the same people who made Java at Sun are working on it. This project was strained, but achieved its goal: Sun posted Java in the public domain. And that was the beginning of the end. In 2009, Oracle acquired Sun Microsystems for a relatively small amount.
It was a big strategic victory for Intel, which allowed it to conquer more than 90% of the server market in subsequent years. However, after Sun abandoned the fight for Java, the need for Harmony disappeared, and more than 200 people in Moscow, St. Petersburg, and Novosibirsk lost their jobs with it. Such is the “animal grin of capitalism” ☹️ But they were by no means the first to leave Intel after the deal was completed. In the “bear project” there were all the same difficulties as in Sarov in terms of adapting to the corporate culture. The “rejection of the leaders” was superimposed on this – in less than two years after the end of the deal, Intel left – Roman Poborchiy (St. Petersburg), Ivan Golosov (Novosibirsk), Vladimir Volkonsky (Moscow) and many others. Another stressful phenomenon was the massive redistribution of work within the Russian Intel. Compiler work was curtailed in Nizhny and Sarov and consolidated in Moscow and Novosibirsk. MKL also migrated from Sarov to Novosibirsk. Sarov never recovered from this blow, but Nizhny, it must be said, shook up pretty well. Someone left for new locations, someone lost their jobs. The organizational structure changed, and not everyone liked it.
In general, Russian Intel was stormy after this deal for another 10 years. Somehow settled down. But by 2014, only 30% of those who came with the “bears” remained from the force. In many teams in Moscow and Novosibirsk there was a “change of generations”. At the helm stood young leaders better adapted to the Intel culture. And yet, from the Bear Project, I still have some kind of feeling of an “unsung song”. Yes, there were certain results, but it still seems that they could have been much more…
The next big acquisition of Intel in Russia took place 12 years later and also in Nizhny Novgorod. Unlike the brutal Bear, this project was called Violin. 😊 I don’t know if the name played any role, but in terms of efficiency, this acquisition was probably the most successful. The Itseez company was created at the end of the 2000s by people from Intel who were at the origins of the OpenCV library – Gary Bradsky, Viktor Erukhimov, Sergey Molinov, Vadim Pisarevsky, Alexander Bovyrin and some others. Also, Kirill Kornyakov and vice-president of MERA, who joined in 2013, played key roles there Alexey Myakov and Yuri Gorbachev. So when Intel seriously moved towards the Internet of Things, it was difficult to find a more “convenient” candidate for acquisition. The Itseez team happily avoided the difficulties of adaptation. The fact is that its culture was built by people from Intel and is extremely close to “light blue”. In addition, the guys were able to quickly rebuild from projects based on computer vision in the direction of accelerating neural networks. This is how the OpenVINO product was born, which has become one of the hallmarks of not only Nizhny Novgorod, but also global Intel.
This is the last (so far?) acquisition Intel is in Russia and I know relatively little about it. After leaving the post of CEO of RND Intel in Russia in 2015, I practically stopped appearing in Moscow. However, the first impression, when the deal was announced within the company, I clearly remember – “a feast during the plague.” The fact is that at that time Intel was rampant ACT – a large-scale wave of layoffs associated with the closure of small (<500 people) sites, relocations, layoffs, etc. (More on this will also be in the Cursed Days chapter.) And suddenly I find out that 150 people around the world and 60 in Russia are coming to Intel. The first impression is some kind of nonsense. It feels like the left hand doesn't know what the right hand is doing. Although, if you look at it, everything is logical - the deal with Soft Machines was agreed upon a long time ago, and AST was being prepared in the strictest confidence. But somehow it so happened that a spoonful of honey drowned in a barrel of tar. If we talk about the absorption itself, then it also belongs to the category of "extremely friendly" The founder of Soft Machines Mohammed Abdallah is a native of Intel. Yes and for Alexandra Drozdova – a key person in Russia, this was not the first “walker” in the corporation. However, he quite safely left her a second time in 2020. 😊 As for the team itself, it was waiting for a “reversal on the march” akin to the one experienced by the “bears” 12 years earlier. Soft Machines promoted the VISC architecture, thanks to which it attracted the attention of the buyer. But with the advent of Intel Raji Koduri in 2017, the GPGPU craze began. And the guys were urgently transferred to a “hot topic”. How successful they were in it is hard for me to judge. But to survive in such a turbulent sea as Intel for 6 years, while maintaining a relatively low attrition rate, deserves respect in itself.
That’s how it was. Long, difficult and not always logical. Nevertheless, Intel will remain a bright page in the history of Russian IT.
To be continued