What You'll Learn
- What Computex 2026 is — the dates, venues, theme "AI Together," and why it matters for the global tech industry
- Nvidia's Vera Rubin platform — 336B transistors, 288GB HBM4, and Jensen Huang's $1 trillion revenue vision
- AMD's $10 billion Taiwan bet — Lisa Su's massive investment in AI infrastructure and what it signals
- Intel's Nova Lake preview — the 52-core desktop chip and what it means for the CPU wars
What Is Computex 2026?
Computex 2026 is the world's premier technology trade show, and this year it arrives at a moment when artificial intelligence has become the single most transformative force in global business. Running from June 2-5 in Taipei, Taiwan, the event carries the theme "AI Together" — a signal that the industry has moved beyond solo AI launches into a phase of deep ecosystem collaboration.
The scale is staggering. 1,500 exhibitors from 33 countries and regions will fill 6,000 booths across four venues: TaiNEX 1 and 2, TWTC Hall 1, and the Taipei International Convention Center (TICC). Stretching across Taipei's Nangang and Xinyi districts, the show is one of the world's largest multi-venue technology showcases dedicated to AI and next-generation computing.
Three core pillars anchor the event: AI & Computing, Robotics & Mobility, and Next-Gen Tech. From semiconductor design and advanced computing infrastructure to robotics systems and real-world AI applications, Computex presents a vertically integrated technology ecosystem spanning research, manufacturing, and deployment. The show runs 9:30 AM to 5:30 PM on June 2-4, and 9:30 AM to 3:30 PM on June 5.
Jensen Huang's Keynote: The $1 Trillion Vision
The headline event at Computex 2026 is Jensen Huang's keynote at GTC Taipei, scheduled for Monday, June 1, at 11:00 AM Taipei time at the Taipei Music Center. This is not just another product launch — it is the unveiling of Nvidia's vision for the next decade of AI infrastructure.
At GTC 2026 in March, Huang projected $1 trillion in cumulative revenue from Nvidia's Blackwell and Vera Rubin chips through 2027 — doubling the $500 billion forecast he had delivered just a year earlier. "I see through 2027 at least $1 trillion of demand for Blackwell and Rubin," Huang told attendees. The projection underscored the explosive growth in AI computing demand that has outpaced even the most bullish forecasts.
Huang arrived in Taiwan early, on May 23, ahead of the show. His first stop was a meeting with TSMC Chairman C.C. Wei to discuss Vera Rubin production capacity. The visit signaled just how critical Taiwan's supply chain has become to Nvidia's AI ambitions. Each NVL72 configuration connects 36 Vera CPUs and 72 Rubin GPUs, contains nearly two million parts, and draws on roughly 150 Taiwanese suppliers.
At Computex, Huang is expected to provide live updates on Vera Rubin's rollout, demo next-generation AI applications, and outline the company's physical AI and robotics strategy. Nvidia's GTC Taipei sessions will cover AI factories and scaling infrastructure, agentic AI and reasoning AI, AI for science, and physical AI and robotics.
Vera Rubin: The Chip That Could Redefine AI
The Vera Rubin NVL72 is Nvidia's most ambitious platform to date. Built on TSMC's 3nm process, the Rubin GPU features a dual-die design with 336 billion transistors — a 1.6x increase over Blackwell's 208 billion. Each GPU is equipped with 288GB of HBM4 memory delivering 22 TB/s of bandwidth, nearly tripling Blackwell's 8 TB/s on HBM3e.
The performance numbers are staggering. A single Rubin GPU delivers 50 PFLOPS of FP4 inference throughput. Scaling to the full NVL72 rack with 72 GPUs, the system achieves 3,600 PFLOPS (3.6 EFLOPS) of inference and 2,520 PFLOPS of training. That represents roughly 5x the inference performance of Blackwell per GPU, and 65x the performance of the prior Hopper generation at the rack level.
But Rubin is more than just a GPU. It is a seven-chip platform that includes the Rubin GPU, the Vera CPU (with 88 Olympus cores and 1.5TB LPDDR5X), the Groq 3 LPU for low-latency inference, NVLink 6 switches delivering 3.6 TB/s per GPU, ConnectX-9 SuperNICs at 1.6 Tb/s, BlueField-4 DPUs, and Spectrum-6 Ethernet switches. Nvidia is not just selling chips — it is selling complete AI factory blueprints.
The Vera Rubin platform ships in the second half of 2026. Nvidia had previously forecast $500 billion in demand through 2026 for Blackwell and Rubin; the updated $1 trillion projection through 2027 includes contributions from the new Vera CPU, Groq 3 LPU, and storage rack products. Jensen Huang has described it as the biggest product ramp in Nvidia's history.
| Specification | Rubin R100 GPU | Vera Rubin NVL72 |
|---|---|---|
| GPU Count | 1 (discrete) | 72 Rubin GPUs |
| Transistors | 336 billion | 24+ trillion (system) |
| GPU Memory | 288GB HBM4 | 20.7TB HBM4 |
| Memory Bandwidth | 22 TB/s | 1,580 TB/s |
| FP4 Inference | 50 PFLOPS | 3,600 PFLOPS (3.6 EFLOPS) |
| FP4 Training | 35 PFLOPS | 2,520 PFLOPS (2.5 EFLOPS) |
| Process Node | TSMC 3nm (N3P) | TSMC 3nm (N3P) |
| NVLink Bandwidth | 3.6 TB/s per GPU | 260 TB/s aggregate |
AMD's $10 Billion Taiwan Bet
AMD is not just attending Computex — it is making a massive statement. On May 21, 2026, AMD announced more than $10 billion in investments across Taiwan's AI ecosystem, aiming to deepen strategic partnerships and scale advanced packaging capabilities for AI infrastructure.
The investment signals AMD's determination to compete head-to-head with Nvidia in the AI chip race. AMD CEO Lisa Su arrived in Taiwan ahead of Computex, and the company is expected to showcase its next-generation AI accelerators and provide updates on the EPYC Venice CPU roadmap — a Zen 6-based processor with up to 256 cores planned for 2026.
AMD also announced its Advancing AI 2026 event for July 22-23 in San Francisco, where the company will detail its AI roadmap and next-gen platform plans. Biostar has already teased next-gen AMD motherboards for Computex, hinting at early looks at the Zen 6 desktop platform — though the full Ryzen "Olympic Ridge" Zen 6 desktop launch may slip to 2027.
The $10 billion commitment puts AMD in direct competition with Nvidia's trillion-dollar vision. While Nvidia dominates with an estimated 81% AI chip market share, AMD's aggressive investment in Taiwan's ecosystem — the same supply chain Nvidia depends on — shows that the AI chip war is far from over.
AMD Stock Surges 18% After Crushing Earnings: The $10.25 Billion Quarter That Proved Lisa Su's AI Bet Is Paying Off
Intel's Nova Lake and the CPU Wars
Intel is also making a major play at Computex 2026. The company is expected to preview Nova Lake, a 52-core desktop processor built on Intel's new architecture, with a commercial launch slated for the second half of 2026. Nova Lake engineering samples have already started shipping, with early benchmarks claiming up to 2x multi-core gains over the current Arrow Lake generation.
Intel's Computex showcase will also feature Clearwater Forest data center processors, Arrow Lake Refresh mobile chips, and the upcoming chipset 900 series motherboards with a new LGA1954 socket. The platform supports up to 48 PCIe lanes and DDR6 memory, positioning Intel to compete with AMD's AM5 platform on multiple fronts.
Intel's presence at Computex is critical for the company's recovery narrative. After losing ground to AMD in data centers and struggling to compete with Nvidia in AI accelerators, Intel needs to demonstrate that its 18A process node and new architectures can deliver competitive products. The Nova Lake preview will be a key test of that strategy.
The AI Chip Arms Race: Nvidia vs AMD vs Intel
Computex 2026 arrives at a pivotal moment in the AI chip industry. Nvidia dominates with an estimated 81% market share, but AMD's $10 billion Taiwan investment and Intel's Nova Lake preview show that competitors are not backing down. The global AI compute sector has entered a massive structural transformation, driven by a $700 billion hyperscaler spending wave.
The Vera Rubin platform's specs tell the story of why competition matters. With 5x inference performance over Blackwell and 65x over Hopper at the rack level, Nvidia is pushing the boundaries of what AI systems can do. But AMD's Zen 6 EPYC Venice with 256 cores and Intel's Nova Lake with 52 cores show that the CPU side of the equation is also evolving rapidly.
The supply chain dynamics add another layer of complexity. Both Nvidia and AMD depend on TSMC for advanced packaging, and Taiwan's semiconductor ecosystem is under increasing pressure. Jensen Huang's early arrival in Taiwan and his meeting with TSMC's chairman signal that securing production capacity is just as important as chip design in the AI race.
What to Watch at Computex 2026
For investors, analysts, and tech enthusiasts, Computex 2026 offers several critical moments:
June 1 — Jensen Huang's GTC Taipei Keynote: The main event. Expect Vera Rubin production updates, new AI platform demos, and possibly a preview of the next-generation Feynman architecture targeting 2028.
June 2-5 — AMD and Intel Showcases: AMD will likely detail its AI accelerator roadmap and Taiwan investment specifics. Intel will preview Nova Lake performance and 18A process milestones.
Throughout — Robotics and Physical AI: The show's new robotics zones will feature the latest in autonomous systems, humanoid robots, and AI-powered manufacturing — areas where Nvidia, AMD, and Intel are all investing heavily.
Market Impact: Computex announcements typically move chip stocks significantly. Nvidia, AMD, Intel, TSMC, and supply chain players like Foxconn, ASE, and SK Hynix could all see trading activity around key announcements.
Conclusion
Computex 2026 is more than a trade show — it is the annual temperature check of the global AI industry. With Nvidia projecting $1 trillion in chip revenue, AMD pouring $10 billion into Taiwan, and Intel previewing its next-generation architectures, the stage is set for a pivotal week in Taipei. The theme "AI Together" captures the moment: no single company can build the AI future alone, and Computex is where the ecosystem comes together to make it happen.
For anyone following the AI chip race — whether as an investor, developer, or technology enthusiast — Computex 2026 is the event to watch. The announcements made in Taipei this week will shape the AI landscape for years to come.
Last Updated: May 30, 2026 | Sources: NVIDIA Blog, TechCrunch, CNBC, Reuters, AMD Investor Relations, TAITRA Computex Official Site