Trillion-Dollar Investments in AI Technology

Most of the market is still pricing AI like it’s 2023. LLMs, chatbots, GPUs, parameter races. That cycle is over. The next trillion-dollar AI wave? It won’t come from bigger models. It’ll come from solving what’s really holding AI back: • Bandwidth bottlenecks • Latency at scale • Data liquidity across silos • Lack of domain-specific process intelligence The models are good enough. The plumbing isn’t. ⸻ And here’s what almost everyone’s missing: AI Agents are about to obliterate the SaaS business model. The old way: 1 seat = 1 user = $X/month The new way: 100s of autonomous agents doing work, at scale, around the clock. TAM is no longer limited by employee headcount. It scales with how much work your customer needs done. That’s a fundamental shift. Software isn’t a tool anymore—it’s a synthetic workforce. ⸻ But here’s the catch: The winners won’t be generic AI wrappers. They’ll be companies that understand: • The regulatory pain in pharma • The operational bottlenecks in energy • The complexity of compliance in finance • The friction in real-world workflows This cycle belongs to those who go deep: Build vertically. Automate intelligently. Embed agents in real industry processes. ⸻ The infrastructure arms race is just beginning. We’ll need trillions in investment—bandwidth, cooling, compute, interoperability. It could be the biggest public-private infrastructure buildout in history. And the outcome? A new class of AI-native platforms. A new software growth model. A new digital labor force. As Jensen Huang put it: “AI Agents are probably the next robotics industry—and likely to be a multi-trillion-dollar opportunity.” ⸻ Don’t follow the last cycle. Design for the next one. Solve friction. Go vertical. Build agentic workflows. And remember: This time, output—not users—defines value. #AI #ArtificialIntelligence #AgenticAI #SaaS #FutureOfWork #EnterpriseAI #DigitalTransformation #AIInfrastructure #LLMs

Last week, Amazon’s CEO reaffirmed their $100B investment into AI and data center infrastructure despite market volatility, tariff uncertainty, and macro hiccups. The message couldn’t be clearer: the AI CapEx race is still very much on, and it’s all about customers. They want more tools, more infrastructure, and more models. Demand isn’t slowing. AI capabilities aren’t slowing. Today's AI is the worst version of AI we will ever have; tomorrow's will be better. Total AI infrastructure spend this year? Over $300B across the major players. 🔹 Amazon: $100B into AWS and Trainium 🔹 Microsoft: $80B into Azure 🔹 Google: $75B into Gemini and cloud 🔹 Meta: $62.5B to power the next wave of AI-driven engagement This isn’t just public cloud. This is the foundational layer of the next era of compute, intelligence, and global connectivity. What we’re watching isn’t incremental—it’s transformational. AI is pulling forward CapEx at a velocity we haven’t seen since the internet and telecom buildouts. From chips to racks to data centers next to volcanoes—yes, really—massive bets are being placed to win the future. This is the moment.

With Microsoft talking about a $6.5 trillion ecosystem TAM for AI and OpenAI talking about a $7 trillion AI infrastructure buildout, we at Canalys (part of Omdia) are collecting the receipts. For example, 2023 global data center capex, including all market segments, was $325 billion. In 2024 that rose to $466 billion (43% growth y/y), primarily driven by AI server investments. Omdia's five-year capex growth outlook is 18.3% CAGR reaching $753 billion in 2028. This growth is highly concentrated across the largest platform vendors. For example: --> The top 15 service providers (listed below) grew 81% y/y. --> The big four - Amazon Web Services (AWS), Google, Meta, and Microsoft - the growth was even higher at 94%. These top companies have publicly disclosed $325 billion spend in their financial reporting causing some alarm among investors. Data center capacity continues to expand exponentially: 11 million new sq. feet has come online with another 50 million sq. ft anticipated. The AI surge has driven third-party data center operators to fill the gap where hyperscalers need capacity but have no time to build. Driving this growth is server capex which grew 88% y/y in 2024 reaching a record $229 billion. NVIDIA is flirting with a $4 trillion valuation today.