Tech Giants: Starting from Bankrupt Companies

91: Chapter 91: The Spark Ignites a Prairie Fire, Mobilization for the First BattleThe Five Major Assault Groups: Architecture and Goals (August 18)

The following morning, the first general technical conference of the Xingchen Academy of Sciences was held in the campus’s largest conference room. Aside from the five Academicians, the core backbone of various Spark chip departments and expert representatives from domestic partner enterprises like Empyrean gathered together, totaling over two hundred people.

The top-level architecture diagram of the ”Flint Project” was displayed on a massive projection screen, consisting of five parallel assault groups. Like five sharp arrows, they pointed directly at the core fortresses of the chip industry chain.

1. EDA Software Assault Group (Leader: Li Weimin)

Ultimate Goal: Within 90 days, complete the prototype development and successful tape-out verification of a domestic EDA toolchain supporting the full 16nm process (digital front-end design, logic synthesis, layout and routing, physical verification, and analog circuit design).

Key Battles:

Timing Closure Algorithm: Reconstruct the Static Timing Analysis (STA) engine to solve accurate modeling of interconnect delay, crosstalk, and voltage drop under deep sub-micron conditions.

Physical Verification Engine: Rewrite the Design Rule Check (DRC) and Layout Versus Schematic (LVS) tools to ensure a perfect connection between design and manufacturing rules.

Parallel Computing Framework: Utilize domestic computing platforms to achieve large-scale distributed parallelism of the tools, enhancing efficiency.

Team: Li Weimin personally leads the core algorithm group (10 people); integrates 80 members of the original Spark EDA team; and 30 senior engineers stationed from Empyrean to collaborate on the assault.

2. Lithography and Precision Equipment Assault Group (Leader: André Dupont)

Ultimate Goal: Within 6 months, complete the integration of a functional prototype and testing of key modules for a domestic 16nm immersion lithography machine (prototype code-named ”Lightseeker-1”).

Sub-item Breakthroughs:

Optical System: Dupont is personally responsible for this. Based on his non-mainstream designs, it uses new lighting modules and simplified projection objective lenses to significantly reduce system complexity and cost while ensuring 16nm resolution.

Precision Machinery: Collaborate with Shanghai Micro Electronics Equipment and the Tsinghua Department of Precision Instrument to tackle ultra-high precision dual-workpiece stages and laser interferometer measurement systems.

Control System: Partner with Wuhan Huazhong Numerical Control to develop dedicated multi-axis linkage nanometer-level motion control software for the lithography machine.

Collaboration: Shanghai Micro Electronics Equipment (Overall Integration), Changchun Institute of Optics, Fine Mechanics and Physics (Optical Component Processing), Harbin Institute of Technology (Precision Machinery).

3. chip Materials Assault Group (Leader: Robert Chen)

Ultimate Goal: Within 12 months, achieve stable domestic mass production and supply of the three core materials required for the 16nm process: 12-inch monocrystalline silicon wafers, ArF immersion photoresist, and CMP (Chemical Mechanical Polishing) slurry.

Technical Breakthrough Points:

Utilize Chen’s exclusive ALD technology to develop new high-k gate dielectric materials, building reserves for future advanced processes.

Deeply collaborate with domestic material enterprises to establish a full-chain joint laboratory covering material synthesis, purification, and application testing.

Partner Enterprises: Zhonghuan Semiconductor (Silicon Wafers), Nanda Optoelectronics (Photoresist), Anji Microelectronics (Polishing Slurry).

4. Integrated Circuit Manufacturing Process Assault Group (Leader: Zhang Mingyuan)

Ultimate Goal: Within 14 months, build a completely de-Americanized 16nm integrated circuit process demonstration R&D line within the Xingchen Academy of Sciences.

Technical Route Assessment:

Parallel assessment of the feasibility of FinFET, FD-SOI, and GAA paths under domestic equipment and material conditions.

Pragmatic First-Phase Goal: Utilize existing domestic equipment (such as the Shanghai Micro Electronics Equipment 90nm lithography machine) and domestic materials to achieve complete tape-out of a 90nm logic chip, opening up the entire chain.

Significance: This line is not just a technical verification platform, but a real testing ground and feedback iteration loop for all-domestic equipment and materials.

5. chip Architecture and Ecosystem Assault Group (Leader: James Wilson)

Ultimate Goal: Design a brand-new, fully independent intellectual property general-purpose computing chip instruction set architecture, tentatively named ”Flint Architecture (Flink ISA)”.

Core Features:

Native support for scalar, vector, and matrix mixed operations, oriented toward the future trend of AI and high-performance computing convergence.

Hardware-level security and Trusted Execution Environment (TEE) design.

Modular and scalable, supporting full-scenario coverage from IoT devices to supercomputers.

Ecosystem Construction: Simultaneously launch the ”Flint Architecture” dedicated compiler (in collaboration with the CAS Institute of Software), basic operating system adaptation (in collaboration with domestic OS vendors), and porting of key basic software libraries.

The 90-Day Battle: The Life-and-Death Line (Afternoon of August 18)

After the group discussions, Li Weimin walked alone to the large whiteboard at the front of the conference room and picked up a black marker. The room fell silent instantly as everyone’s gaze focused on that pen.

He didn’t draw any complex block diagrams; he simply used clear and powerful strokes to write down a schedule that was simple to the point of being brutal:

【EDA Toolchain 90-Day Assault Battle】

Days 1-30: Core Algorithm Reconstruction Period.

Goal: Complete the mathematical modeling and algorithm rewriting for the three core engines: timing analysis, logic synthesis, and physical verification.

Output: Core algorithm prototype code library.

Days 31-60: Toolchain Integration and Verification Environment Setup Period.

Goal: Integrate the new algorithm engines with the existing tool framework (such as graphical interface, database, and process file interface) to form a runnable EDA tool prototype. Simultaneously, set up a simulation verification cluster based on domestic servers.

Output: Preliminary runnable toolchain prototype, test case sets, and verification environment.

Days 61-90: Actual Tape-out Verification Period.

Goal: Use the self-developed EDA toolchain to complete the full process from design to GDSII layout data generation for the revised ”flint no. 2” chip (16nm, code-named ”Flint-1A”). Deliver the data to a domestic foundry (preliminarily selected as Huali Microelectronics) for trial production tape-out (MPW).

Output: Complete layout data for tape-out; the first batch of engineering samples.

Life-and-Death Line: November 20, 2017. Tape-out data delivery must be completed.

Finishing the last stroke, Li Weimin turned around and slowly capped the pen. This 72-year-old man’s gaze was like lightning, sweeping across every face below—some young, some mature, but all etched with tension and determination.

Su Chen spoke up, asking the question everyone wanted to ask but didn’t dare: ”Dean Li, if... 90 days pass and we fail, if the tape-out data isn’t ready, or the chip doesn’t work after tape-out, what then?”

The conference room was so quiet one could hear a pin drop.

Li Weimin was silent for a few seconds. Those few seconds felt as long as a century. Then, he spoke slowly, his voice not loud but carrying a decisive, unquestionable power that reached everyone’s ears clearly:

“There are no ‘ifs’.”

He paused, his gaze seemingly piercing through the walls to look further away, or perhaps returning to his own magnificent yet regret-filled academic career:

“I, Li Weimin, am seventy-two years old this year. In my life, I have seen algorithms go from clumsy to exquisite, and I have seen EDA go from an auxiliary tool to a yoke that chokes our throats. These ninety days may be the final battle of my career.”

His gaze refocused, becoming incredibly sharp, as if burning with an invisible flame:

“I will not allow myself to go to my grave with the regret that ‘Chinese people cannot build chips without American software.’ You,” he pointed at everyone below, “should not face your children and grandchildren with such doubt and humiliation either.”

“So, there are no ‘ifs.’ There is only the necessity of completion.”

A Conversation Deep into the Night: Spark and Inheritance

The meeting ended late at night. The crowd dispersed to rest and prepare for tomorrow’s official start of the war. Su Chen and the five Academicians did not leave; they went to an open-air terrace in the campus facing the sea. The sea breeze blew with a salty, damp scent, and in the distance on the pitch-black sea, the lights of fishing boats flickered faintly.

A waiter brought hot tea and some simple snacks and then withdrew. Only the six of them remained on the terrace, under a sky full of quiet stars.

André Dupont held his teacup without drinking, just staring at the stars in a daze. After a long time, he asked softly, “Su, do you know what finally made me decide to come here?”

Su Chen looked at him, gesturing for him to continue.

“It wasn’t because of those final unpleasant arguments at ASML.” Dupont’s voice was very soft, as if he were reminiscing. “It was because in the last two years before I left, I participated in an internal project to evaluate how to ‘selectively’ reduce the performance of equipment exported to certain regions by fine-tuning software parameters or restricting components. At that moment, I realized that the optical perfection I had pursued all my life, the technology I was proud of that could make light accurate to the picometer, was becoming a tool for... selective blindness. It was being used to decide who is qualified to ‘see’ the future. This goes against the meaning of light, and the meaning of science.”

Robert Chen took over, his tone much more pragmatic: “In my final years at Applied Materials, I was forced to fill out dozens of so-called ‘Customer Risk Assessment Reports’ every year, focusing on Chinese semiconductor companies. The report templates were full of leading questions: ‘Do they have the intent to steal technology?’ ‘Is the background of their R&D personnel suspicious?’. Once, I wrote in a report: ‘If we always speculate with the greatest malice against a country with the most engineers and the largest market in the world, and refuse to sell them any advanced tools, the only outcome will be to force them to build a completely independent system that may eventually compete with us.’ My supervisor crossed that out and said, ‘That’s not for you to consider.’” He gave a bitter smile. “You see, the prophecy came true. I just didn’t expect I would become part of this ‘other system’.”

Zhang Mingyuan gazed into the distance at the darkness and said slowly, “Intel used to be a synonym for innovation. Moore’s Law was our faith. But now... bureaucracy, path dependency, and an obsession with huge profits have made that former temple sluggish and conservative. They would rather patch up old architectures than risk opening a new path. I left because I didn’t want to watch the giant that once led the world become a boulder blocking progress.”

James Wilson took a sip of tea, speaking with a touch of British humor: “The situation at ARM isn’t as tragic, but it’s equally frustrating. After SoftBank’s acquisition, everything became a numbers game: market share, royalty growth rates, IoT connection numbers... The beauty of the architecture itself, the elegance of computing, and thinking about the computing paradigm for the next decade have all taken a back seat. When technical decisions are entirely dictated by spreadsheets, I knew it was time to leave.”

Li Weimin had been silent, but now he spoke, his voice carrying the vicissitudes of someone who had seen through the world: “Our generation experienced the golden age of science after World War II and witnessed how computers and semiconductors reshaped the world. We once believed that science was a bridge to a better world. But now, this bridge is being artificially blocked by roadblocks, and some are even trying to tear it down. The five of us come from different countries and different fields, but sitting here now, the reason is probably the same: we do not agree that ‘blockades’ should define the future of science. We do not believe that a passport should determine who is qualified to participate in humanity’s most cutting-edge intellectual exploration.”

Su Chen listened quietly, his heart surging with emotion. He raised his teacup, using tea in place of wine:

“To science. To ideals. To those things that cannot be blocked, no matter how many Iron Curtains or bans are imposed—human curiosity, creativity, and our shared aspiration for a better future.”

The five Academicians smiled at each other. In that smile, there was sentiment, relief, and even more, a sense of burning from having found a battlefield and a mission once again. They all raised their cups.

“To science.”

“To the future.”

The crisp sound of teacups clinking rang out softly in the night breeze of the South China Sea, like the striking of Flint, sparking the first tiny but destined to spread prairie fire.

Behind them, the lights in several buildings within the campus remained bright throughout the night. Those were the algorithm engineers of the EDA group working late into the night, the young people of the equipment group digesting the massive amount of optical drawings brought by Dupont, and the materials group building the first ALD experimental chamber overnight...

The Flint fire has been lit; the momentum of the prairie fire is unstoppable.