Inside the MediaTek–Nvidia Collaboration

The partnership spans a full AI computing stack, creating engineering demand at every layer:

• AI PC processors: MediaTek co-designed the N1X processor powering Nvidia RTX Spark laptops, announced at GTC Taipei. The chip is built on TSMC’s 3nm process .
• Personal AI supercomputers: MediaTek co-designed the GB10 Grace Blackwell Superchip inside the DGX Spark personal AI supercomputer, which shipped publicly in October 2025 .
• Custom AI ASICs: MediaTek entered Nvidia’s custom ASIC ecosystem as an early adopter of NVLink Fusion, using its high-speed interconnect and design expertise to create cloud-scale AI silicon .
• Automotive solutions: The two companies are building joint in-vehicle cabin platforms using Nvidia GPU chiplets and AI IP, targeting production as early as 2026 .
• IoT and edge AI: Collaboration extends to edge AI frameworks including Nvidia TAO, bringing AI inference to lower-power devices .

Each initiative requires engineers specializing in chip design, advanced packaging, system integration, and AI software — roles that did not exist at this scale before the current AI buildout.

Huang’s Computex 2026 Argument: ‘Complete Nonsense’

Speaking at Computex 2026 in Taipei, Jensen Huang directly confronted the claim that AI eliminates software jobs :

“People talk about AI reducing jobs — complete nonsense. It’s causing more software engineers to be hired.”

Huang’s argument rests on a straightforward economic logic: when a single software engineer’s output multiplies through AI tools, companies want more of that productive output — not less . He likened AI’s effect to previous technology waves: “PCs made us more busy. The internet made us more busy, mobile made us more busy. AI will do the same” .

As supporting evidence, Huang pointed to GitHub commit volumes, which have nearly tripled, and AI-assisted coding usage skyrocketing from 300 million sessions in 2023 to nearly 1.4 billion in early 2026 . The implication: AI tools are amplifying engineering output so dramatically that demand for engineers is rising, not falling.

Reconciling the 107,000 Layoff Figure

Huang’s optimistic view sits against an uncomfortable backdrop: over 107,000 tech workers have been laid off in 2026 so far . Huang and his supporters reconcile the contradiction in two ways:

1. Many layoffs stem from post-pandemic restructuring and efficiency drives, not direct AI substitution. Tech firms notoriously over-hired during the 2020–2022 boom, and the current cuts echo the cyclical corrections that predated generative AI.
2. AI-specific hiring is growing faster than legacy roles are shrinking. MediaTek’s expansion — hiring a TSMC packaging veteran, targeting $1 billion in new data-center business, and staffing up for Nvidia co-design projects — illustrates how the AI infrastructure buildout generates categories of work (chiplet interconnect design, AI factory operations, automotive AI integration) that barely existed five years ago.

Huang’s core claim remains nuanced: he is not saying zero jobs will be lost, but rather that new categories of engineering work will absorb more talent than roles that get automated. MediaTek, Qualcomm, and the broader Nvidia partner ecosystem are currently demonstrating that pattern in real time.

In the end, the question is not whether AI will displace individual tasks, but whether the total demand for engineers building AI systems will outpace the reduction in traditional roles. For now, the chip industry’s hiring data leans in Huang’s direction.

Inside the MediaTek–Nvidia Collaboration

The partnership spans a full AI computing stack, creating engineering demand at every layer:

• AI PC processors: MediaTek co-designed the N1X processor powering Nvidia RTX Spark laptops, announced at GTC Taipei. The chip is built on TSMC’s 3nm process .
• Personal AI supercomputers: MediaTek co-designed the GB10 Grace Blackwell Superchip inside the DGX Spark personal AI supercomputer, which shipped publicly in October 2025 .
• Custom AI ASICs: MediaTek entered Nvidia’s custom ASIC ecosystem as an early adopter of NVLink Fusion, using its high-speed interconnect and design expertise to create cloud-scale AI silicon .
• Automotive solutions: The two companies are building joint in-vehicle cabin platforms using Nvidia GPU chiplets and AI IP, targeting production as early as 2026 .
• IoT and edge AI: Collaboration extends to edge AI frameworks including Nvidia TAO, bringing AI inference to lower-power devices .

Each initiative requires engineers specializing in chip design, advanced packaging, system integration, and AI software — roles that did not exist at this scale before the current AI buildout.

Huang’s Computex 2026 Argument: ‘Complete Nonsense’

Speaking at Computex 2026 in Taipei, Jensen Huang directly confronted the claim that AI eliminates software jobs :

“People talk about AI reducing jobs — complete nonsense. It’s causing more software engineers to be hired.”

Huang’s argument rests on a straightforward economic logic: when a single software engineer’s output multiplies through AI tools, companies want more of that productive output — not less . He likened AI’s effect to previous technology waves: “PCs made us more busy. The internet made us more busy, mobile made us more busy. AI will do the same” .

As supporting evidence, Huang pointed to GitHub commit volumes, which have nearly tripled, and AI-assisted coding usage skyrocketing from 300 million sessions in 2023 to nearly 1.4 billion in early 2026 . The implication: AI tools are amplifying engineering output so dramatically that demand for engineers is rising, not falling.

Reconciling the 107,000 Layoff Figure

Huang’s optimistic view sits against an uncomfortable backdrop: over 107,000 tech workers have been laid off in 2026 so far . Huang and his supporters reconcile the contradiction in two ways:

1. Many layoffs stem from post-pandemic restructuring and efficiency drives, not direct AI substitution. Tech firms notoriously over-hired during the 2020–2022 boom, and the current cuts echo the cyclical corrections that predated generative AI.
2. AI-specific hiring is growing faster than legacy roles are shrinking. MediaTek’s expansion — hiring a TSMC packaging veteran, targeting $1 billion in new data-center business, and staffing up for Nvidia co-design projects — illustrates how the AI infrastructure buildout generates categories of work (chiplet interconnect design, AI factory operations, automotive AI integration) that barely existed five years ago.

Huang’s core claim remains nuanced: he is not saying zero jobs will be lost, but rather that new categories of engineering work will absorb more talent than roles that get automated. MediaTek, Qualcomm, and the broader Nvidia partner ecosystem are currently demonstrating that pattern in real time.

In the end, the question is not whether AI will displace individual tasks, but whether the total demand for engineers building AI systems will outpace the reduction in traditional roles. For now, the chip industry’s hiring data leans in Huang’s direction.