AI summary by TwinMind
Looking Back At ATI Technologies
Asianometry
16m 41s
Company Origins (1983-1985)
- Kwakuen Ho (KY Ho) - Hong Kong immigrant vacationed in Toronto 1983, decided to relocate
- Family lost everything in Chinese Civil War, fled to Hong Kong
- Ho sold vegetables, father worked in factory
- 1970s: earned scholarship to National Chengkong University, Tainan
- 1974: graduated electrical engineering degree
- Worked at Control Data Systems, National Semiconductor in Hong Kong
- Canadian job market challenges - electronics industry immature, far from Silicon Valley
- 1985: Founded Array Technologies Industry (ATI) with two other Hong Kong immigrants
- Co-founders: Bani Lao, Li Lao
- Initial funding: $300,000 CAD life savings, 6 employees
Early Struggles & Breakthrough (1985-1988)
- Vision: color monitors would make PCs mainstream, needed powerful graphics hardware
- Funding crisis after 3 months - founders' money exhausted
- UOB (Singaporean bank) provided $300,000 CAD loan, later increased to $1.5M CAD
- Business model: OEM graphics controllers, then integrated graphics cards for IBM, Commodore
- Breakthrough: Commodore partnership - selling 7,000+ chips/week
- First year revenue: $10M CAD
- Went public on Toronto Stock Exchange
Product Success & Market Position (1988-1996)
- EGA Wonder and VGA Wonder series - major hits
- VGA Wonder (1988): breakthrough graphics power, market compatibility
- Drove company to $60M CAD sales
- Claimed "world's biggest graphics board maker"
- Industry standards participation
- Co-founder of Video Electronics Standards Association (9 founding members)
- Established 800x600 resolution video display format
- Mach series 2D graphics accelerators
- Mach 8 (1991): first product processing graphics without CPU, optimized for Windows
- Mach 32 (1992): combined graphics controller + accelerator on single chip
3D Graphics Transition (1996-2000)
- 1996: Quake game sparked 3D graphics revolution
- Acquired Kubota Graphics design team for high-end workstation expertise
- 3D Rage chip (originally 264 GT)
- Named to compete with S3's Verge 3D chip
- Lower performance than NVIDIA but good value positioning
- Rage Mobility (1998): defeated S3 and NeoMagic in notebook market
- Market dynamics: 3dfx fell 1998, NVIDIA led performance, ATI "fast follower"
Intel Competition & Strategic Response (1998-2000)
- Intel i-740 graphics card (1998): failed due to performance issues
- Intel pivoted to integrated graphics in CPUs - "good enough" for low-end
- ATI most threatened by Intel's encroachment (value-oriented position)
- Strategic response: diversify across industries
- Low-end gaming to workstations to gaming consoles
ArtX Acquisition & Console Market (2000)
- $453M acquisition of ArtX (70-person startup)
- ArtX founder: Wei Yen (Taiwanese-American, ex-Silicon Graphics)
- Nintendo relationship: designed N64 graphics, contracted for GameCube successor
- ATI logos on GameCube chip, later produced for Wii and Xbox 360
- ArtX CEO became ATI COO, expanded into new industries
- Reorganized into two design teams for faster product releases (9-month cycles)
NVIDIA Competition Era (2000-2005)
- GeForce II GTS (2000): 25M transistors, unprecedented performance
- ATI Radeon response: redesigned transform/lighting engine 20% faster
- NVIDIA Detonator 3 drivers significantly improved older GeForce 2 performance
- Consistent ATI weakness: inferior driver quality vs NVIDIA
Radeon 8500 Missed Opportunity (2001)
- Built on TSMC 150nm process (same as GeForce III)
- TrueForm tessellation engine innovation
- Automatically inserted triangles for smoother, natural surfaces
- No frame rate compromise
- John Carmack praised 8500 performance
- Failure factors: buggy early drivers, benchmark cheating scandal
- Developers didn't adopt TrueForm, feature gradually removed
Financial Struggles & AMD Acquisition (2005-2006)
- 2005 financials: $2.2B revenue (Xbox 360 shipments), only $16.9M profit
- Value cards sold well but yielded minimal profits
- Intel dependency: 60-70% chipset revenues from Intel platforms
- Strategic concern: pincer movement between NVIDIA (high-end) and Intel (low-end)
- AMD acquisition: $5.4B offer accepted
- AMD first attempted NVIDIA acquisition, rejected by Jensen Huang (CEO control issue)
- Rationale: CPU-GPU convergence expectation, AMD's "Fusion APU" strategy
Post-Acquisition Challenges (2006-2008)
- Cultural integration problems: "green" (AMD) vs "red" (ATI) sides
- Fusion APU delays: over 2 years behind schedule
- Acquisition price too high, non-GPU businesses overvalued
- December 2007: AMD write-down of acquisition value
- 2007 AMD financial loss: $3.3B loss vs $6B revenue (>50% of revenue)
- 2008 Q2: additional $1B impairment charges
- Debt burden: $2.5B debt issued for acquisition, $5B total debt by 2008
- 2008 global financial crisis: PC sales cratered
- Near bankruptcy, saved by Abu Dhabi financial infusion
Current Status & Legacy
- AMD spun off fabs to GlobalFoundries
- CPU revival through chiplet innovations
- GPU market: NVIDIA still leads performance, benefits from deep learning/CUDA dominance
- AMD Radeon: continues good value positioning, open-sourced drivers
- ATI brand phased out 2010 after 25 years
- Legacy: "always a survivor"
