The AI scale advantage: On the road on the US West Coast

When Satya Nadella assumed the role of Microsoft CEO in 2014, he wrote that “Our industry does not respect tradition – it only respects innovation.” A decade later, those words resonate more strongly than ever. In 2025, US equity markets were dominated by a handful of technology giants, while artificial intelligence (AI) reshaped the competitive landscape at an unprecedented pace. AI-related stocks were the primary engine of market gains, yet scepticism is mounting around the vast capital being committed to infrastructure build-outs. 

A recent trip to the West Coast provided a timely opportunity to engage directly with some of the world’s leading technology companies. Discussions centred on how these businesses are navigating a period of profound transformation and, crucially for long-term investors, the strategies they are deploying to sustain innovation and maintain leadership in an increasingly competitive landscape.

Microsoft’s strategic position

Microsoft continues to demonstrate its leadership in the AI era, building on a long history of technological innovation. During our recent visit to the company’s headquarters in Redmond, we observed strong and accelerating demand for its AI-driven solutions. 

Rather than forecasting demand for each individual use case, Microsoft focuses on building broad, adaptable platforms that can support many forms of AI at once. For example, the same Azure infrastructure underpins Office Copilot, enterprise data tools and third party AI applications. This shared, flexible technology stack allows Microsoft to meet demand wherever it emerges across the AI ecosystem, giving it a clear advantage over competitors backing narrower, more specialised growth paths.

Further reinforcing this outlook, Microsoft’s remaining performance obligations – the value of contracted products and services not yet recognised as revenue – have surged to nearly $400bn. This reflects exceptionally strong near term demand and validates its accelerated infrastructure expansion. With most contracts averaging around two years, the urgency to scale capacity is clear, ensuring Microsoft can meet customer commitments and fully capitalise on this growth opportunity.

In the current environment, Microsoft is well positioned, remaining highly selective in the contracts it accepts. It now holds roughly a 27% stake in OpenAI’s for profit arm, reflecting both the scale of its long term commitment and its strategic position at the centre of the company’s future development. This partnership underscores Microsoft’s ambition to lead in artificial general intelligence, even if that technology remains some way off. Efficiency gains are also emerging, with token throughput for GPT-4.1 and GPT-5 up over 30% per GPU compared to older generations. 

Importantly, Microsoft advocates a portfolio approach to AI adoption. Rather than concentrating resources on a single proof of concept, the company encourages starting several small AI projects simultaneously to identify the most promising initiatives. This approach helps identify the most valuable ideas quickly, spreads risk, and speeds up innovation, reinforcing Microsoft’s ability to remain agile and ahead of the curve in a rapidly evolving landscape.

Amazon’s approach to scale

We had an equally engaging discussion with Amazon, another of the so-called ‘hyperscalers’. Regarding the datacentre build-out, Amazon’s flexibility is paying off. Its strategy of building extra capacity, rather than giving up market share, has paid off because it keeps its options open. Capacity can be added, slowed or paused as needed. Crucially, the most expensive part – the chips – comes last, giving the company valuable flexibility if conditions change. Given Amazon’s cash flow, balance sheet and existing businesses, it can absorb costs in an overbuild scenario. This is reminiscent of its e-commerce investment throughout the Covid era, where it outpaced competitors.

Custom application-specific integrated circuits (ASICs) have long played a key role in semiconductor innovation. Amazon has been using its own custom ASICs for Amazon Web Services (AWS) for years. The company recently unveiled its latest version, Trainium 3, at its AWS re:Invent 2025 conference. According to Amazon, it delivers more than 4× overall compute performance and greater energy efficiency than the previous generation, resulting in better price-performance. In short, it delivers more AI output for the same spend. Given its market leadership, particularly across AWS and e-commerce, Amazon has excellent leverage over suppliers and remains focused on long-term planning – a key differentiator when conditions tighten.

Both Microsoft and Amazon show strong financial resilience through balanced investment strategies, substantial cash reserves with the flexibility to scale when it matters. In contrast, a company like Oracle’s approach, while ambitious, depends heavily on long-term lease commitments and higher leverage, which introduces greater financial risk. Overall, Microsoft and Amazon appear better positioned for sustainable growth, while Oracle faces a more complex risk profile.

Resilience amid disruption

Beyond the AI hyperscalers and infrastructure build-out, one of the big talking points has been whether software-as-a-service businesses are “dead” given the rise of AI-based applications. Adobe is one company that has come under scrutiny and share price pressure. From our recent conversation in San Jose, it is clear that Adobe remains laser-focused on building a differentiated ecosystem. Its future offerings will undoubtedly be AI-centric, but creator retention rates remain stable, supporting seat growth – an increase in the number of licensed users or subscriptions within organisations – and driving customer value over time. Competition has always existed. Investors once feared mobile and the web would end Adobe, yet the company continues to evolve.

For many enterprise customers, the sticking point with AI models is commercial safety. This is where Adobe should shine, enabling brands to train models on their own data rather than relying on internet-trained large-language models (LLMs). Customer segments are more diverse than a decade ago, now including legal, finance and accounting departments. In October, Adobe launched Firefly Foundry, an enterprise-level generative AI (GenAI) platform that lets businesses build customised models trained on approved, company owned content.

Adobe likens its approach to the iPhone ecosystem: a core platform with built-in apps and the ability to add others. It has also adapted its processes to accelerate innovation, moving from large engineering groups to smaller, autonomous pods fostering agility and accountability.

Speed, data and discipline at Intuit

Intuit is a leading provider of accounting and tax software that helps individuals and businesses manage their finances. Meeting with its Chief Technology Officer Alex Balazs, we heard how deeply AI is embedded in the company’s strategy. Machine learning initiatives since 2018 have delivered the greatest impact, and a pivotal decision a decade ago to clean and structure data now enables personalised, GenAI-powered experiences. Intuit’s culture reinforces this edge, hiring heavily at early career levels – a generation Balazs calls “AI native”. New joiners complete an intensive bootcamp and are encouraged to challenge norms. Their approach: write only essential code and let AI handle the rest, accelerating development and adaptability.

Strategically, CEO Sasan Goodarzi predicted that AI would disrupt business logic – the rules and processes that govern how a company operates and makes decisions – rather than core platforms, which are the foundational systems supporting those operations. Intuit’s position is strong, with 40 years of proprietary data, an advantage many start-ups lack. “Intuit intelligence” resembles a smarter GPT, grounded in fact-checked data. Industry dynamics amplify the opportunity. As AI reshapes the competitive landscape and eliminates niches for smaller players, Intuit seeks to capitalise on these shifts through its agility and scale. Speed, proprietary data and an AI-ready talent model reinforce control of destiny and align with our investment framework.

Balazs likens today’s environment to the internet boom of 1999 – only at 10× speed. For us, this acceleration underscores why scale data assets, disciplined architecture and a culture of experimentation remain critical to compounding value over time.

ASML’s AI edge

We also met ASML’s Silicon Valley team, acquired through Brion in 2007, which leads the company’s most complex computational challenges, including big data and AI-driven metrology. This team heads ASML’s advanced big data initiatives and takes on the toughest algorithmic problems. Meeting with this group, one of the most intricate parts of the company, was fascinating and underscored its critical role in shaping ASML’s future.

Our visit reinforced two key points. First, TSMC’s extraordinary influence over the entire semiconductor ecosystem. And second, how much room there still is to make these systems run better. Most of the improvements so far have focused on the early stages of making a wafer, rather than fine tuning each individual chip in real time. Even small gains in yield really matter, so although progress will be gradual, the long term opportunity is substantial.

Opportunities and caution

During our trip, we gained a clear sense of the prevailing bias towards overinvestment in AI infrastructure and valuable insights from the businesses we met. From our perspective, AI is undoubtedly a transformational technology, one that will deliver productivity gains and improved decision quality through data insights. That said, we remain cautious about current infrastructure spending trends, mindful that no investment boom has ever perfectly matched supply with demand. Amazon and Microsoft stand out as industry titans with the financial strength and innovative capability to maintain leadership positions.

On the software side, we believe the bear case has been overstated. AI-native challengers are seeking to disrupt incumbents, but leading firms such as Adobe and Intuit have responded decisively, embedding AI at the heart of their strategies. Their robust customer relationships and extensive proprietary data provide a strong advantage.

Confident long-term outlook

Innovation is not confined to technology. We also met with several healthcare companies, including Edwards Lifesciences, Intuitive Surgical and ResMed. This sector has long been out of favour compared to technology, but we are confident that many of these businesses are developing innovative treatments that will benefit patients globally. Regardless of sector, we came away from this trip more convinced than ever that the US is home to many of the world’s most innovative and high-quality businesses – those positioned to thrive meaningfully over the long term.