Lessons from Intel's First Foundry

Lessons from Intel's First Foundry

Show Video

Today, Intel seeks to remake itself into a  full-service foundry with Intel Foundry Services. But did you know that this is not  their first foundry? Some 15 years ago,   Intel had Intel Custom Foundry, or ICF. Most people today hardly remember  ICF. Intel was so low-key about it.   But at its peak it had a billion  dollars in revenue contracts,   strong support by the ecosystem, over  a thousand multi-national employees ... And ten plus customers - many won over from TSMC. In today's video, the rise and  fall of Intel’s first foundry.  

And the lessons to be learned for its second. ## The Best Factories Intel in the 2000s believed that it had two Crown Jewels. The first of which was its  leading edge manufacturing tech. For two decades, Intel under its CEOs Andy Grove,   Craig Barrett and Paul Otellini  developed and scaled a brand new   technology to billions of chips. And did this  every two years to hit the Moore's Law cadence. Such technologies spanned disciplines  from lithography like the excimer lasers   to new deposition methods like Atomic Layer  Deposition to new materials like Hafnium Oxide.

Several times, the company's  engineers made astute,   bold technical choices that yielded  where others did not. Like for example,   the choice of "Gate Last" over the "Gate First"  method, backed by the tech giant IBM and its   allies. Even TSMC pursued "Gate First" before  abandoning it halfway through development. Such choices granted the company a strong  1-2 generation lead over its competitors.   And Intel's people certainly liked to  hammer that point in their presentations.

But Intel's top management also knew  that fabs were getting increasingly more   expensive. And at some point, Intel's PC market  revenues alone cannot justify the next node. And in those years, the semiconductor industry was  still hoping to make the jump to 450 millimeter   wafers. Larger wafers reduce per-die cost  but with the drawback of massive investment. The proper strategic response  then would be to roll up demand   from outside fabless customers and  use that to help fund the next node.

## Intel Architecture What Intel saw as its other crown jewel was the  x86 chip architecture, or Intel Architecture. Before Otellini took over the CEO spot from Craig  Barrett, Intel offered an Arm-based product called   XScale. Otellini sold XScale to Marvell in  2006, seeking to refocus the company on x86. In an August 2007 interview  with the Financial Times,   he explained why he was throwing the  company's whole weight behind x86: > And rather than rethink or rewrite the internet,   which today runs on X86 machines,  between us and our competition. I mean,   the entire internet is coded for that base. All  the browsers, all the Flash applications ... You catch that? I admit it is  a bit garbled. But Otellini is  

asserting that the whole internet was  written on x86 machines, referring to   Intel's then growing strength in Linux and  Xeon-powered internet servers. He continues. > Rather than rewrite the internet to a new  architecture, it’s a lot easier to – from my   perspective, to do derivatives of our architecture  to meet new form factor and power requirements The iPhone proved Otellini’s sentiments incorrect,   and Intel sensed the  immediate shift away from x86. In response, they rushed out the Intel Atom  brand of low-power processors. These later   powered small, cheap netbooks, but  that wasn’t the original intention. ## Custom Intel Architecture Foundry Intel's first foundry began as a simple initiative by Paul Otellini to monetize  these two crown jewel assets.

A small team was formed to provide  "turnkey custom foundry services"   for internal and external customers to  produce custom or semi-custom chips. The foundry's original name was "Custom  Intel Architecture Foundry" or CIAF. The name hints at the goal: Monetize Intel’s  manufacturing factories, and also defend the   threat to the Intel Architecture franchise  from the trend of low-power systems-on-chip. The effort essentially kicked off in 2008  with an Intel Technology & Manufacturing   VP discretely approaching existing customers,   and touting the benefits of customizing Intel  products using Intel manufacturing processes. This VP shortly left this role and was replaced  by Sunit Rikhi, a vice president from Intel's   Technology and Manufacturing division.  Rikhi saw the foundry as an opportunity   to build a new business and would serve as  its general manager for the next seven years.

## Building a Foundry Yes, Intel could be called the leader  in silicon manufacturing technology. But that technology had been tightly optimized  by Intel’s Technology and Product groups to   make fast x86 CPUs and not much else than  that. It was less a question of ability,   than rather those processes being in no  shape to present to external customers. Think of Intel’s factories as  like a pizza kitchen. For years,   that pizza kitchen - its equipment, ingredients,   processes and cooks - produced pepperoni pizzas  per a recipe made by another internal Intel team. The two groups know each other well  and how they worked. And together they  

made the best pepperoni pizzas in the industry. But now we must reveal the kitchen’s capabilities  to external customers, who then might use it to   make cheese pizzas, Hawaiian pizzas, Chicago  deep-dish pizzas, or New York style pizzas. Might even get the occasional  weirdo asking for a calzone. A foundry like TSMC, UMC or GlobalFoundries  will have done the extensive work needed to   lay out its kitchen’s capabilities  for a customer to pick and choose   their pizza the way they want it  to be. Intel needed to match that. Moreover, the Intel foundry team had to  also equip itself with complementary IPs   that its customers can use. Think  toppings like olives, sausages,  

and onions. Building this whole  ecosystem required a lot of work. In the case of TSMC, they had been  doing this for over three decades.   And those years of cumulative  knowledge have yielded a vast,   widely varied menu as well as a kitchen capable  of producing precisely what was ordered off   it. A different mindset, and approach,  from what Intel had always done before. ## Ditching Intel Architecture Custom Intel Architecture Foundry was the  second prong of a two-pronged strategy by   Otellini to bring a low-power x86 Intel  Architecture, Atom, to mobile products. The first prong had been an alliance  with TSMC. In this partnership,   fabless customers can license Intel Atom IP  and build systems-on-chips on top of them.

Less than a year later, however, this  collaboration was put on hold due to low customer   demand. The team at Custom Intel Architecture  Foundry experienced the same low demand issues. This low demand stemmed from the Atom CPU's  immaturity as an ecosystem or platform for   building products. And even if there had  been customer interest, it would have been an   immense technical challenge to turn the CPU into  usable IPs akin to those already offered by Arm. So some time in 2009, Otellini was convinced to  drop the Intel Architecture part of the name. They would move forward with  just Intel Custom Foundry. He reportedly did this reluctantly, and  held it in his heart until his retirement   that Intel Custom Foundry should be about  Intel Architecture designs rather than that   of others like Arm. Regardless, he gave ICF  support and should be credited for as such.

## Competing With Customers There remained one major thing. Intel going into   the foundry business raises the question  of competing against your own customers. This issue dates back to the days of the Japanese  semiconductor-makers, who gladly rented out spare   capacity on their lines in exchange for  a pound of flesh, your stock or your IP. TSMC was founded on the basis of never doing this,   and it remains enshrined  in their corporate charter. Intel would have been very much  competing with many of its foundry’s   largest potential customers. Companies like  AMD, Nvidia, and increasingly, Qualcomm. Even if Intel was not in that business  right then, there lingered the worry that   that might change. The company’s future  aspirations spread quite far and wide.

So the Intel management chose to do a slow  and discreet rollout. They restricted the   young foundry to only engaging with  customers in areas that Intel does   not compete in and had no aspirations for.  Which was not a very large playing field. In the end, the Intel Custom Foundry team  chose to compete in a specific type of chip   called a Field-Programmable Gate Array,  or FPGA. That was a large enough space   that Intel then had no intention of getting into. Later. In the first two quarters after  Intel went public with the news that it was  

getting into foundry, analysts asked twice  about Intel foundry’s impact on capacity. Then-CFO Stacy Smith, who is now on the board,   replied both times that Intel  was not focused on that business,   was only talking to very specialized companies,  and was not building a broad-based foundry. ## Big Customers That turned out to be alright,   because Intel Custom Foundry was nowhere  near ready to accept big customers. Shortly after news first emerged of Intel  offering foundry services in late October 2010,   Apple emerged as a potential foundry client. There was a brewing conflict between  themselves and their current SOC foundry,   Samsung. Later in 2011, Apple  sued Samsung for infringing   on Apple's intellectual property.  It was time to go somewhere else.

Morris Chang wrote in his autobiography - which   for now is only available in  Chinese - that in November 2010,   Terry Guo of Foxconn rung him to say that  he would bring an Apple executive to dinner. That executive was Jeff Williams, today  Apple's COO. Williams showed interest in   engaging TSMC to produce iPhone SOCs and  the two sides started working on a deal. But just as this work ramped up,  it was suddenly put on pause. Jeff   Williams called Morris to tell him that Paul  Otellini was negotiating with CEO Tim Cook   on a potential foundry deal. The TSMC-Apple  collaboration would be paused for two months.

Morris then wrote about visiting Apple  in April 2011 asking about things. Cook   simply tells him that "Intel is not good  at 'contract manufacturing'" (他們不擅做代工). Let's drill into that a bit. In late 2010,   Intel Custom Foundry was only about  two years old. They were profoundly  

unprepared to take on what is perhaps the most  demanding tech company client in the world. I want to remind you. The most important  decision that any fabless company can make   is their choice of foundry partner. Get  that wrong, and nothing else matters. For   such a company-defining decision,  you don't want to take any risk. Intel Custom Foundry, after two or so years  of work, was not yet 90% close to TSMC.   Even if they were, that last 10% would be a big  enough deal that Apple will still go with TSMC.

Apple's people took the meeting because their top  management asked for it. But in terms of actual   KPIs and milestones to get the deal, Intel was  nowhere near ready yet. It was dead on arrival. There was another opportunity - maybe -  for an Intel-Apple deal three years later,   but I will get to that in a bit. ## Startup So Intel Custom Foundry was not  ready for the major leagues.

But there were a few customers amenable to  what Intel was offering. Those guys were   the startups. They were tired  of being so beholden to TSMC. They were also more tolerant of the  competitive risks of working with Intel,   thanks to intensely structured contracts that  held Intel to severe punishments if they did   not protect customer IP from Intel's Product  divisions or provide the contracted capacity. The thing most important to these  startups was Intel's league-leading   silicon manufacturing. Intel Custom  Foundry's first Process Design Kits   were for the 32 nanometer node, but  they quickly switched to 22 nanometers.

Because customers wanted to bring out  a league-leading product and leapfrog   their competitors. Considering how long it  would take to bring a chip to the market,   the 32 nanometer process node  would already be lagging. Intel Custom Foundry's first reported launch  customer was a small fabless FPGA startup called   Achronix. Achronix's FPGAs were for specific  tasks like network traffic or data encryption. The Wall Street Journal reported in late  October 2010 that Intel agreed to give the   startup access to about 1% capacity of its  leading-edge 22-nanometer FinFET process. In an interview, Achronix's chief executive  acknowledged that the Intel wafers cost more,   but said, "This is a historic development. Intel   is far ahead of anyone else  in new process generations".

## 22 nanometers Intel's process lead on its competitors seemed the greatest when it came to the  critical transition to 3D transistors. In late 2011, Intel announced  its 22 nanometer process node   equipped with its "Tri-Gate" variant of FinFETs.  The FinFET is a type of transistor which covers   the channel on three sides, offering  superior power consumption and speed. It took over a decade for Intel to master  the various bits of making the FinFET,   which included controlling the variations you  tend to get in the manufacture step. This work  

remains one of the company's crowning  achievements in semiconductor history. Intel originally scheduled to ship its  "Ivy Bridge" 22-nanometer CPUs in April   2012. But they encountered some problems,  and the ship date slipped to June 2012. A few weeks delay, but no big deal at the  time. Intel was so far ahead of everyone   else - at least two generations  - that nobody thought more of it. For Intel Custom Foundry, Achronix’s Speedster22i  FPGAs started to ship in volume in early 2013.

And those chips were good - ICF always  did make good silicon once it was all   said and done. These FPGAs were notable  for consuming half the power of other   high-end chips from other FPGA  makers like Altera and Xilinx. The best working process node those other  two FPGA companies had access to in those   days was TSMC's 28-nanometer node  - which was then still somewhat   volume-constrained. So 22 nanometer  was maybe one or two generations ahead. In February 2012, Intel added a second  foundry customer, Tabula. Like Achronix,   Tabula was another FPGA-maker startup, and they  had been using TSMC's 40-nanometer process node. And then in April 2012, another start-up  customer called Netronome. They previously  

used TSMC's 65-nanometer process node.  The news makes it clear. From the start,   ICF was stealing customers from TSMC. ## Design-Manufacturing Integration Intel's external spokespeople regularly touted their design-manufacturing  integration as the future of foundry. In October 2012, Intel scientist  and senior fellow Mark Bohr said: > The traditional foundry model  is running into problems. In order   to survive, the foundries will have to become  more like an integrated device manufacturer.   Even some of the chief spokespeople for  the foundries have said something similar.

> The foundry model worked well when traditional  scaling was being followed and everybody knew   where we were headed. In this era, where you  continually have to invent new materials and   new structures, it’s a lot tougher being a  separate foundry and maskless design house. > Being an IDM, we have design  and process development under   one roof. That’s really a significant advantage. A fully-focused Intel foundry can provide  more than just wafer fabrication. They can   also leverage the organization's  full line of design and advanced   packaging services too. They can do  the whole course from soup to nuts. Some of the packaging stuff was special.  A bit later, Intel's teams invented a  

new type of advanced packaging called  Embedded Multi-die Interconnect Bridge. Per the name, this format embeds a small silicon  chip into the package to connect different dies   without the need for a more expensive silicon  interposer. None of the Intel product teams   could use it, but Foundry had customers who did. A  perfect example of Intel's integration strengths. ## Behind the Scenes So theoretically, this all made sense. But GlobalFoundries sang the same  tune of having IDM-like capabilities   when they first spun off from AMD.  Only to discover that in practice,   the two sides - locked in their individual  kingdoms - would struggle to come together.

Intel was no different. Many within Intel thought  that Intel Custom Foundry was a waste of money,   a side project distraction from  what was really generating revenue. They were also worried what it  might mean for their own KPIs.   Intel’s Product people worried  that having a foundry might raise   outside concerns regarding Intel's own  internal evaluation of its products.

Imagine if Intel Custom Foundry taped out  another CPU and it turned out to be faster   than what Intel Product could do. Imagine  the egg on their faces. So they cried foul. And then Intel's Technology group. They  sponsored and funded Intel Custom Foundry   in part because they thought it would showcase  the company's superior technology ... and maybe   put those pesky Intel Product guys back  in their place for complaining so much. But when that did not immediately turn  out to be the case - when customers   had feedback on things they wanted changed -  the Technology Group's support fell out too. This left the Intel Custom Foundry team  struggling to get their peers to change in   order to better serve the outside world. The  Technology people took a request from Intel  

Product at least somewhat seriously because  it impacted revenue. But Foundry? No chance. For example. Most larger fabless customers -  from their experiences working with TSMC and   other foundries - are used to getting some  level of insight into a node. But the Intel   Technology R&D team was extremely secretive  and often refused to divulge such details. They also resisted making transistor changes  along the power-performance curve to accommodate   customers - seeing such things as distracting from  their main goal of advancing the leading edge.

To manage customer frustration, the  Intel Custom Foundry team regularly   brought in senior management like  the aforementioned Mark Bohr to   meet with such customers and use his  leverage to get R&D to cooperate. ## Altera The FPGA market's two largest  players were Altera and Xilinx. The two were fierce rivals. Altera had been a  long-time big customer of TSMC - one of their  

top ten customers, actually. Meanwhile,  Xilinx used TSMC's rival foundry UMC. But when UMC started to wither - a  rumored issue at its 65 nanometer   node delayed several Xilinx  products - Xilinx defected. In February 2010, they announced that they had  added Samsung and TSMC as their new leading edge   partners. In an attempt to win the deal  over Samsung and their 40-nanometer node,   TSMC gave Xilinx access to its  then-leading edge 28-nanometer node.

This pissed off Altera, which now needed  a technical advantage they can leverage   against Xilinx in the high-end FPGA  space. This brought them to Intel. In 2013, Altera agreed to make FPGAs  on Intel's 14-nanometer node - which   followed 22 nanometers. Some trailing edge  stuff would still be made at TSMC. CEO   John Daane cited Intel's process  lead as the reason for the move: > "Intel’s 14-nm is a second  generation FinFET process,   while others are just starting  to implement their first" At the start, TSMC did not take ICF all  that seriously. I heard second-hand of  

a prominent former executive  there being quite dismissive.   So the Altera defection came as  a real shock for the Taiwanese. Daniel Nenni recalled in a blog  post on his SemiWiki website: > I was having coffee with a friend in TSMC Fab  12 when it was announced. If my memory serves   it was Dr. Morris Chang who made the announcement  and it honestly felt like parents were divorcing. In the company's next earnings call on April  2013, Morris Chang said this about the loss: > We have gained many customers in the  last few years, but I really hate to   lose even a part of an old one. We  want them all really. I regret it.  

And because of this we have thoroughly  critiqued ourselves ... it's a lesson   to us and ... we'll try our very best not to  let similar kinds of things to happen again Chang then addresses Intel Custom  Foundry's growing impact to TSMC itself.

> I still view Intel as a selective picker  among customers. As a foundry competitor   they will pick their targets and so on ... And  I don't view them as a general competitor ... > But they are a very serious  competitor to our customers.   That really I would say applies even  greater pressure on us than Intel as   a direct foundry competitor. They are a  very serious competitor to our customers.

You see what he did there?  Morris is a very careful speaker.   He dismisses ICF as a general foundry competitor,   and then shifts the message again to his core  point: Intel competes with its own customers. To me, it says everything: Intel Custom Foundry's  progress was indeed making an impact on TSMC.

## 14 Nanometers Altera was without a doubt ICF’s biggest  win yet. And it was not the only one. These early wins were promising  enough to Intel senior management   to slowly loosen the reins on the foundry's  sector restrictions. New business came in   from Panasonic and then LG, which contracted  ICF to make specific system-on-chips for them. Switching to a new foundry always takes time. But   it took a few months for Altera to even  get started on their 14-nanometer design. This was because another FPGA project -  the Arria 10, built on TSMC's 20-nanometer   process node - had sucked out design  resources. The Intel Custom Foundry  

team had to dedicate many of their own design  resources to ready the design for the fab. The design delay turned out to be fortunate  because Intel's 14-nanometer node was   a year late. The node's technology  definition had been too aggressive. Not fatally so, but aggressive enough to force   additional experiment cycles and thus  extra time to mature into good yield. At the end of 2012, Intel announced  that 14-nanometer should be ready by   2013. It did not reach high  volume production until mid  

to late 2014 - entirely missing  the two-year Moore's Law cadence. The year-long delay also gave foundries like  TSMC and GlobalFoundries - whose 14-nanometer   node was being done in collaboration with  Samsung - some time to close the gap. But like I said, not a fatal error. Just  took some extra time to get out. Once mature,   the node was fine - a second-generation FinFET  process denser than TSMC’s first-gen FinFET   16-nanometer. The main issue was the  impact on the node that came after it. ## 10 Nanometers In 2013, Intel and Apple met once more for  discussions on a possible foundry deal. Apple was again in the market. Perhaps with the  

intention to dual-source their next  iPhone chip, the A9. Rumors abounded   in the financial media that Intel was  targeting Apple as their big customer. They were probably just kicking the tires.  Intel Custom Foundry had gotten a few   customers by now and was learning from  them, but it was still a bit too early. The A9 second-source deal  eventually went to Samsung.

Critically in those meetings, Apple  indicated that it wasn't interested   in the Intel 14-nanometer node, which in 2013  was then going through its slow maturation. But rather the node after that: 10-nanometer,   originally due in 2016. Had it ramped on  time, it would have done so 2 years ahead   of TSMC's N7 node - which it was quite  similar to. But famously, it didn't.

What went wrong with 10-nanometer?  Intel has never said publicly,   and the story is probably best reserved  for another day. This is what I understand. Intel develops its process nodes  with two teams working in parallel. One team works on what would be 14 nanometer  gen+1 while the other works on 10, gen+2. So by the time senior management realized that  14-nanometer had a serious problem - perhaps   some time in 2014 - the definitions for  10-nanometer had already been solidified. And 10-nanometer was a full 50%  area shrink - with a 2.7x increase  

in transistor density - on top of what  had already been a very aggressive 14. SemiAnalysis's Dylan Patel added in a  2022 report that Intel’s use of cobalt   in the vias and interconnects  played a major factor too. The   deposition tools for that were not ready  yet, but Intel blazed ahead despite warnings.

And meanwhile, EUV - which I remind you that  Intel funded and always believed in - was   not ready in 2014 or even 2015. It would not  be inserted into a process node until 2019,   when TSMC did it for their N7+ node. Even  if it was available earlier than that,   it wouldn’t have been trivial  to redesign the node to add EUV.

Long story short, it was too much.  A step of hubris that took them off   the cliff. But Intel's Product group had  already taken the assumptions Technology   group made for what 10-nanometer  was promised to offer and developed   products with them. Only to find out  that those promises cannot be kept.

It was like building a 50-story  skyscraper only to find out on the   20th floor that the foundations can't  support the whole building. You can't   just tweak this. You must tear it all  down and start again - a 2-3 year process. Once the news about 10-nanometer's delays  started reaching outside parties in 2015,   the backlog of available business for the  foundry rapidly began to dry up. By the time  

10-nanometer fabbed Intel chips finally started  to ship in 2019, the foundry no longer existed. ## The Altera Acquisition And then in June 2015, Intel  announced that it would buy its   biggest foundry customer Altera for $16.7 billion. CEO Brian Krzanich’s galaxy brain thinking  behind buying Altera was that they can   integrate FPGAs into their data center CPUs. Such  an arrangement could let people offload certain   workloads to those FPGAs - which meant that they  wouldn’t need to buy a GPU to do those workloads. A relatively sound idea. And  indeed after AMD acquired Xilinx,   there have come out some interesting  SoCs with similar arrangements.

But for Intel, the work of merging these  two previously discrete chips - CPUs and   FPGAs - turned out to be too much.  The strategic notion soon dissolved   and the mooted product never made it to market. From the perspective of Intel Custom Foundry,   the Altera acquisition was a disaster.  Assimilating Intel’s marquee external   foundry customer sends a bad message to  all of the foundry’s potential customers. It implied to those potential customers -  guys like Qualcomm and such who were only   starting to open up to an Intel foundry  - that using ICF might mean Intel later   trying to buy them too. Not to mention all  the competition issues Morris talked about.

## End Alone, the Altera buy was a bad  move. Combined with 10-nanometer's   ongoing failure, it was crippling. In 2015, a number of major retirements,  including that of General Manager Sunit   Rikhi - who had run Intel Custom Foundry for  nearly eight years - delivered the final blow. He was replaced by someone more amenable to  Technology's perspectives. In other words, making   the people in the Technology look good, rather  than advocating for the actual foundry customers. So when the customers inevitably started to get  angry with their inability to get through Foundry,   senior management arranged for those  customers to work directly with the   Technology people. In other words, Intel  Custom Foundry got cut out of the loop.

With the writing on the wall now obvious,  the division - which once employed over   1,100 people around the world - quietly  began to dissolve. By 2018 it was no more. ## Conclusion Circumstances have greatly changed between Intel Custom Foundry in 2008 and  Intel Foundry Services in 2021. It seems like Pat Gelsinger's approach  upon returning to Intel was to rebuild from   scratch. I think that was a mistake. Basically  throwing out the years of progress made by ICF. Intel Custom Foundry was no joke. Over eight  years, they built up a book of 10+ customers   and a billion plus dollars of contracted  revenue. If TSMC considered them a real  

competitive threat, then we should take  their lessons and progress seriously too. So I will share a few takeaways of my own. First,   the foundry service needs to have organizational  support from across the whole organization. It   should not be left awkwardly stuck between  Technology and Product. At the very least,   it should have the autonomy  to modify its own technology.

Second, Intel's most significant technological  pull is a working leading-edge node. That is   what gets people in the door. Intel  Custom Foundry could lean on Intel's   multi-year process node lead on TSMC,  Samsung and the rest of the industry. That lead no longer exists, and retaking  it will be extremely difficult. Get 18A   and 14A working right as a node for  foundry customers - meaning more   than just for making CPUs. That will not be easy.

Third, the only thing that closes the book  is bankruptcy. Prepare for the long run. Let   investors know that this is a very long  term story. TSMC was not built in four,   five or even ten years. Competing  with them will be a ten year endeavor. Fourth, a little humility. Intel Custom  Foundry felt like an unusually humble part   of the company. Perhaps because it was a ragtag  side project that had to serve outside customers. Perhaps because they hired over half of their  workforce out of the fabless-foundry ecosystem.

Perhaps because they closely studied Morris Chang  whenever he talked about the foundry business. Bring that humility to the rest of the company.

2025-01-16 21:55

Show Video

Other news

Best Elon Musk Inventions That Are Coming in 2025 2025-02-13 12:19
Ford CEO Announces All New Lithium Iron Phosphate (LFP) batteries without Degradation 2025-02-11 19:21
China’s New AI Model DeepSeek Shocks ENTIRE Tech Industry...American CEOs in Shock! 2025-02-09 15:01