$NVDA #NVIDIAEarnings

NVIDIA's $800 Billion Wave: The Flow of Money and New Competitors

Big Cloud Companies' Investment Spending of $800 Billion in 2026 and $1.3 Trillion in 2027 is Flowing to NVIDIA, But the Money Doesn't Stop There

2026 marks a period when capital spending on AI infrastructure reaches its highest level in history. The total investment of the five largest cloud providers rose from $162.3 billion in 2022 to $448.3 billion in 2025. Analysts expect this figure to reach approximately $750-800 billion in 2026 and $1.3 trillion in 2027.

Nvidia: A Giant Generating $1 Billion a Day

NVIDIA's second-quarter revenue reached $96.2 billion, with the company generating over $1 billion in revenue per day. The data center segment accounted for 92% of total revenue, at $89 billion.

However, the truly shocking figure was NVIDIA's first-ever guidance of 70% growth for fiscal year 2028. CEO Jensen Huang emphasized that this figure was not based on a 70% increase in demand, but rather on the maximum growth capacity of supply capacity. "Our actual demand is well over 70%, but current supply guarantees this growth," said Huang, adding that "the entire supply chain is operating at full capacity."

The Distribution of Money: An Ecosystem from Memory to Energy

The money flowing into NVIDIA doesn't stop there. The company's supplier commitments jumped from $119 billion to $279 billion in a single quarter. A large portion of this money flows into the supply chain, primarily memory.

Huang notes that the revenue opportunity per data center increases with each new architecture: $180 billion for Hopper, $250 billion for Grace Blackwell, and $400 billion for Vera Rubin. This shows that growth is not just about selling more GPUs, but about each new generation of AI data centers generating more revenue for NVIDIA.

Apple's Challenge: Targeting the H100 with the M7 Ultra

Apple has declared itself a rival to NVIDIA. Its new M7 Ultra chip promises performance that can compete with the NVIDIA H100. Combining two M7 Max dies, the M7 Ultra offers 800 GB/s bandwidth and the capacity to run models with 70 billion parameters.

However, Apple's own AI server chip, Baltra, has been delayed until 2026. This situation makes Apple dependent on NVIDIA H100 and H200 GPUs via Google Cloud for the most demanding AI tasks. Apple's shift towards semiconductor acquisitions demonstrates the company's strategic transformation in this area.

Tesla's Ambitious Entry: AI5 and TeraFab

Tesla is also joining the race. Elon Musk claims that the AI5 chip will offer "2-3 times better" performance compared to NVIDIA and will cost around 10% less. Musk states that he has memorized the chip's physical design and that all of Tesla is tied to this chip.

Tesla, along with SpaceX and Intel, launched the TeraFab project, described as "the largest chip manufacturing facility in history." However, analysts point out that NVIDIA's dominance not only in hardware but also in software and the entire ecosystem is the biggest obstacle for Tesla.

The flow of money is clear: from large cloud companies to NVIDIA, and from there to memory manufacturers, optical connectivity, cooling, and energy. Hardware still has a long way to go. While the entry of new competitors like Apple and Tesla challenges NVIDIA's dominance, the company's strategic moves in its supply chain and its 70% growth guidance demonstrate just how strong the demand for AI infrastructure is.

DYOR 🔎 NFA ✔️