Question 1

What is Microsoft's Maia AI chip?

Accepted Answer

Maia is Microsoft's in-house AI processor, built to power its data centres and the Copilot AI assistant. The goal is to reduce reliance on third-party suppliers like Nvidia for the heavy computational work behind AI. The next generation, Maia 300, is reportedly targeted for a 2027 rollout, reflecting a multi-year engineering effort rather than a quick switch.

Question 2

What is custom silicon in AI?

Accepted Answer

Custom silicon refers to chips designed in-house by a company for its own specific needs, rather than bought off-the-shelf from a chipmaker. In AI, this means building processors tailored to run workloads like large language models or assistants such as Copilot. It can offer better cost control and supply independence, though it requires significant upfront investment and engineering time.

Question 3

What does hyperscaler mean in the AI chip market?

Accepted Answer

A hyperscaler is a large cloud computing company, such as Microsoft, that operates vast data centre networks to deliver AI and cloud services at scale. Several hyperscalers are now developing custom AI chips to reduce dependence on dominant suppliers like Nvidia. This growing trend could gradually reshape how AI compute is sourced across the industry.

Question 4

Is Nvidia at risk from Microsoft's Maia chip?

Accepted Answer

Not immediately. Nvidia's customer base spans many cloud providers, enterprises and research institutions, which helps cushion any gradual reduction in orders from a single hyperscaler. Nvidia also benefits from years of software and ecosystem lock-in that custom chips may struggle to replicate quickly, though the long-term direction is worth watching.

Question 5

Why is Microsoft building its own AI chips?

Accepted Answer

Microsoft wants greater control over costs, supply and performance as AI workloads grow, since chips are often the biggest expense in running data centres. Developing chips in-house could eventually reduce the margin paid to external suppliers. However, the upfront investment is substantial and may weigh on near-term margins before any benefits show through.

Question 6

How does Broadcom benefit from custom AI chips?

Accepted Answer

Broadcom designs custom AI silicon for large technology firms, effectively enabling the kind of in-house chip strategy Microsoft is pursuing. If more hyperscalers follow this path, Broadcom could see rising demand for its chip-design expertise. This positions it in an interesting middle ground between Nvidia's off-the-shelf dominance and hyperscalers' push for independence.

Question 7

Is investing in AI chip companies risky?

Accepted Answer

Yes, this theme carries genuine uncertainty. Chip development timelines can slip, the Maia 300's 2027 target could shift, and investor sentiment around AI infrastructure spending can be volatile and disconnected from fundamentals. As with all investing, capital is at risk and values can fall as well as rise.

Question 8

When will Microsoft's Maia 300 chip launch?

Accepted Answer

Reports suggest the Maia 300 is targeted for a rollout around 2027, signalling a long-term engineering effort rather than a quick pivot. Building custom silicon from scratch requires years of design, testing and manufacturing coordination, so this timeline could shift depending on technical or supply chain challenges.

Question 9

What is driving the shift away from Nvidia chips?

Accepted Answer

Large cloud providers want more control over AI infrastructure costs and supply, especially as chips represent the single biggest expense in running data centres. Building internal alternatives, even for part of their workloads, can improve negotiating leverage and reduce exposure to supply bottlenecks. This is part of a wider pattern among hyperscalers, not just Microsoft.

Question 10

What should investors watch with the Microsoft Maia story?

Accepted Answer

Investors may want to track capital expenditure patterns across Microsoft, Nvidia and Broadcom over coming reporting cycles, as these figures could offer early signals of how the custom silicon trend is playing out. Watching how much Microsoft's spending shifts toward internal chip development versus external purchases could reveal effects on profitability not obvious from headline revenue.

Question 11

How can I invest in AI chip stocks like Microsoft and Nvidia?

Accepted Answer

You can build exposure to this theme through Nemo, the ADGM-regulated investing platform, which offers commission-free access to US stocks including Microsoft, Nvidia and Broadcom. Fractional shares start from just $1, making it possible to build a diversified position without a large upfront sum, supported by AI-driven research through Nemo AI.

Question 12

What AI research tools does Nemo offer beginners?

Accepted Answer

Nemo provides AI-driven research through Nemo AI, giving beginner investors real-time insights to help evaluate themes like custom AI silicon. Combined with commission-free trading and fractional shares from $1, this makes it easier to research and build a diversified portfolio around companies shaping the AI chip market.

Why Microsoft Is Betting Big on Its Own AI Chips

Why Microsoft Is Betting Big on Its Own AI Chips

Recent insights

Why ASML's Bulging Backlog Is a Chip Supercycle Tell

Waymo Crosses The Border Into Driverless Territory

Why a $6.3bn Deal Just Reshaped Data Security

Why Boeing's Delivery Logjam Could Finally Break

Neutron's Slip to 2027 Puts Rocket Lab's Growth Story to the Test

The Hyperscaler Revolt Against Big Tech Hardware

Microsoft Wants Its Own Silicon, But Nvidia's Kingdom Might Prove Harder to Crack

The Trouble With Reinventing the Wheel

Who Really Profits From the Feud?

Deep Dive

Market & Opportunity

Key Companies

Primary Risk Factors

Growth Catalysts

How to invest in this opportunity

Frequently Asked Questions

Recent insights

Why ASML's Bulging Backlog Is a Chip Supercycle Tell

Waymo Crosses The Border Into Driverless Territory

Why a $6.3bn Deal Just Reshaped Data Security

Why Boeing's Delivery Logjam Could Finally Break

Neutron's Slip to 2027 Puts Rocket Lab's Growth Story to the Test

Why ASML's Bulging Backlog Is a Chip Supercycle Tell

Waymo Crosses The Border Into Driverless Territory

Why a $6.3bn Deal Just Reshaped Data Security

Why Boeing's Delivery Logjam Could Finally Break

Neutron's Slip to 2027 Puts Rocket Lab's Growth Story to the Test