Investors are rewarding cloud hosts like AWS for AI-driven growth, while punishing pure hype. Here's what that means for tech strategy. Act accordingly.
AI mania continues to sweep financial markets, but investors have become surprisingly selective. The latest earnings calls and trading patterns send a clear message: they love AI, as long as you’re a cloud host. Companies that own data centers, GPU clusters, and AI services are enjoying premium valuations, while firms that merely wrap AI into their products face tougher questions.
The shift is most visible at Amazon. In 2025, the company plans nearly $100 billion in capital expenditures, largely for data centers and AI infrastructure—and investors seem more enthusiastic than anxious. This marks a dramatic change from the old days, when rising capex would typically trigger fears of shrinking free cash flow. Today, such spending is seen as a necessary investment to capture AI-driven demand.
This article examines why cloud providers are becoming the most trusted AI beneficiaries, what it means for investors, and how technology leaders should adapt their AI strategies.
Why have investors warmed to massive infrastructure spending? The answer lies in revenue growth. AWS reported a 19% year-over-year increase in revenue in Q2 2024, according to Amazon’s earnings release. At AWS’s scale—with roughly 30% share of the global cloud infrastructure market in Q1 2024, according to Synergy Research Group—this level of growth demonstrates real demand, not just speculation.
Cloud hosts profit at every layer of the AI stack. They sell raw compute, managed AI services, and even foundation models. This diversified revenue model reduces risk. When one workload matures, another emerges. AI has created a perfect flywheel: training new models requires more GPUs, which requires more data center capacity, which generates more revenue for cloud providers.
Consider the broader numbers. Hyperscaler AI infrastructure spending rose 45% from 2023 to 2024. That’s not a rounding error. It reflects a coordinated bet that enterprise AI adoption will continue to explode. And because cloud contracts are typically multi-year, this capital spending translates into predictable future revenue.
Investors have updated their models accordingly. They now view capex as a leading indicator of future market share, rather than a drag on near-term profits.
Dan Ives of Wedbush Securities captured the current investor mindset in a June 2024 commentary: “The biggest companies in tech are going to be the ones that provide the picks and shovels for AI—cloud providers are the prime example.” That framing has become a guiding principle for many technology funds.
The picks-and-shovels analogy is historically powerful. During the California Gold Rush, miners often lost money, but merchants selling equipment made steady profits. In AI’s modern gold rush, cloud providers play the role of merchants. They don’t need to pick a winning AI application. They profit whenever any company succeeds in using AI, because every AI service runs on their infrastructure.
Cloud providers also enjoy an ecosystem advantage. AWS, for example, offers a wide range of AI and machine learning services, from SageMaker for model building to Bedrock for generative AI. These tools create a lock-in effect: once an enterprise builds its AI pipeline on AWS, migrating to another cloud becomes expensive and technically painful.
This is why investors are willing to fund enormous capex plans. They see cloud providers accumulating not just hardware, but also developer mindshare, data assets, and enterprise trust—all of which compound over time.
Not all AI companies are treated equally. In recent quarters, publicly traded companies that discuss AI without strong cloud or infrastructure assets have seen more skeptical reactions. The market wants evidence that AI creates real, recurring revenue, not just slideware promises.
Amazon’s CEO Andy Jassy spoke to this advantage on the Q4 2023 earnings call: “We’re optimistic that AWS can continue to grow at a healthy rate as more companies move to the cloud and use AI services built on top of AWS.” His emphasis on “AI services built on top of AWS” highlights the value of having both infrastructure and higher-level services.
Investors are also watching which companies actually own the compute. Pure software AI companies often rent capacity from cloud hosts. That makes them more vulnerable to price changes and capacity constraints. In contrast, cloud hosts control the supply of compute, and can adjust pricing to maintain margins.
Another differentiator is data gravity. Enterprises keep their data in cloud platforms for governance, security, and compliance reasons. When they decide to use AI, they naturally use AI services that can access that data without moving it. This data gravity makes cloud hosts even more entrenched, and it explains why investors prefer them.
For technology professionals, the market’s preference for cloud-host AI has practical implications. It validates a cloud-first strategy for AI initiatives. Instead of building proprietary in-house AI infrastructure from scratch, most organizations should leverage the same hyperscalers that investors are rewarding.
Here are some strategic actions to consider:
Cloud architecture also enables agility. Instead of maintaining expensive, underutilized GPU clusters on-premises, businesses can scale resources up or down based on real-time needs. This operational flexibility gives cloud-host AI companies an edge, because they can adapt to changing demand without massive upfront losses.
Finally, consider the security and compliance angle. Cloud providers invest heavily in certifications and governance frameworks, making it easier for enterprises to deploy AI in regulated industries. This trust factor is another reason why cloud hosts are the preferred AI infrastructure partners.
The 45% rise in hyperscaler AI infrastructure spending is impressive, but it also invites a key question: are we building too much capacity? If AI adoption slows, cloud providers could be stuck with unused data centers and massive depreciation costs. Yet the market’s reaction suggests that investors consider this risk acceptable.
Investor preference for cloud-host AI companies is rising, according to current trends. That’s not a temporary shift. The AI revolution is still in early stages, and the infrastructure buildout is likely to continue for years.
Still, there are risks to monitor. If AWS growth drops below double digits, or if cloud providers start cutting prices to compete, margins could compress. Businesses may also become more cautious about AI spending, which would slow cloud growth. But even in a downturn, established cloud providers are better positioned than smaller rivals.
One more factor supports sustained growth: optionality. Cloud platforms are increasingly offering not just AI compute, but also related services like vector databases, AI security tools, and model evaluation suites. This expansion creates new revenue streams that are less capital-intensive than raw compute.
The bottom line for investors: cloud providers remain the most reliable way to play the AI trend. The bottom line for technologists: align your AI strategy with a cloud provider’s roadmap, and you can benefit from the same tailwinds.
Investors love AI, as long as you’re a cloud host. This is not just a market quirk—it reflects a fundamental truth about how AI works. Training and running AI models requires massive, reliable, and scalable infrastructure, which cloud providers like AWS are uniquely equipped to deliver. The data confirms it: AWS grew 19% year over year, Amazon is spending $100 billion on infrastructure, and hyperscaler AI spending surged 45%.
For corporate leaders, the takeaway is clear. Build your AI strategy on top of the same cloud infrastructure that investors are rewarding. Focus on outcomes, not slogans. Use the cloud’s native AI services to accelerate time-to-market, control costs, and avoid the burdens of managing physical servers. In a market that favors infrastructure over hype, being a thoughtful cloud tenant is almost as good as being a cloud host.
Investors view cloud providers as the essential infrastructure layer of the AI boom. Because hyperscalers like AWS sell raw compute, managed AI services, and foundation models, they generate direct, measurable revenue from AI demand, while companies that merely add AI features often have less clear monetization and face tougher questions about competitive advantage.
Cloud providers profit from multiple layers of the AI stack: renting GPU instances, offering managed machine learning services, and providing platforms for building and deploying AI models. They also benefit from a flywheel effect—more AI training requires more data center capacity, which drives recurring cloud revenue through multi-year contracts.
Leaders should weigh capital requirements, scalability, and speed. Building your own AI infrastructure can offer control but requires massive upfront spending and ongoing expertise. Using a cloud host like AWS provides faster deployment, elastic scaling, and access to a broad AI toolset, making it the more practical choice for most organizations.
The spending is backed by actual revenue growth—AWS reported 19% year-over-year revenue growth in Q2 2024 and holds about 30% of the cloud infrastructure market. Hyperscaler AI infrastructure spending rose 45% from 2023 to 2024, but because contracts are multi-year and usage is accelerating, the capex is seen as a leading indicator of future market share rather than a speculative bubble.
Non-cloud companies should focus on clear use cases that demonstrate measurable ROI, rather than vague AI claims. They should leverage cloud AI services to keep costs variable and experimentation fast, and report concrete metrics like cost savings or revenue lift to reassure investors that AI initiatives are adding value.
