Amazon Reports Tomorrow: The $200 Billion Confirmation
AWS is growing at 28%, the AI buildout is accelerating, and the 2027 capex signal could be the real shocker. The Bargo 2.5x model explains why.
Amazon reports Q2 2026 earnings tomorrow after the close. The Street expects EPS of $1.81 on roughly $182B in revenue. But like every hyperscaler this season, the print is really about one thing: the capex trajectory.
Amazon already guided to $200B in 2026 capex, a 50% increase from 2025. The question is not whether they reaffirm. The question is what they signal for 2027. The Bargo 2.5x tokenomics model says the answer matters enormously.
Amazon has been sliding for two weeks, from $261 at the end of June to $231. The stock is below its 50-day SMA, RSI 37, and approaching oversold territory. The selloff is not an Amazon-specific story. It's the entire hyperscaler complex repricing as the market questions whether the capex is worth it.
The Bargo 2.5x Prediction Model: A Quick Primer
The Bargo tokenomics model is a deterministic structural model of the AI economy. It doesn't predict stock prices. It models the physical flows: how many tokens the world consumes, how much GPU capacity that requires, and whether the hyperscalers are building enough to meet demand.
The model gets its name from its core finding: global token demand is growing at 2.5x per year. This is independently verified from company disclosures. Google's token volume grew 18.4x. OpenAI's API volume grew 5.1x. Gemini's monthly active users grew 2.1x. The capability-derived cross-check, using AI benchmark data completely independent of any revenue series, says the growth could be even faster at 3.2x.
The model has three layers that drive its capex predictions:
Layer 1: Kappa. Kappa measures what percentage of the installed GPU fleet's theoretical capacity is actually being used for tokens. It started at 0.5% in early 2025 and has climbed to 2.33% today. The gap between 0.5% and 2.33% is the demand story. As AI models improve and adoption grows, more of the fleet lights up. At 2.33%, there is still enormous headroom before the fleet is saturated.
Layer 2: Supply-cover. The model compares projected token demand against projected GPU supply. Supply-cover is 1.17x right now, meaning demand is pressing against capacity. When supply-cover is tight, hyperscalers have to build. Through 2027, the model keeps supply-cover between 1.08x and 1.40x. That is tight enough to force continued spending.
Layer 3: The NVDA lead-lag. NVIDIA datacenter revenue leads hyperscaler capex by roughly 3 quarters, with a peak correlation of +0.37. NVDA's recent prints were strong. The model says capex should be accelerating now, and GOOGL already confirmed it with a raise to $195–205B.
Together, these three layers tell a simple story: if demand is growing at 2.5x, supply is tight, and the leading indicator says accelerate, then the hyperscalers must raise capex. They are not choosing to spend. They are responding to physics.
The AWS Engine
AWS grew 28% year-over-year last quarter, the fastest pace in 15 quarters. At a $150B annualized run rate, the cloud division alone justifies the capex. Customer commitments already cover a substantial portion of 2026 AWS infrastructure spending, meaning the capacity is pre-sold.
The 2.5x model projects the inference revenue pool growing from $183B annualized today to $384B by Q4 2026. AWS is the primary platform where that inference happens. Every dollar of capex deployed into AWS AI infrastructure is backstopped by real customer demand.
The 2027 Signal
GOOGL already said 2027 capex will "significantly increase." Microsoft guided Q4 FY26 above $40B in a single quarter. Amazon's CFO will almost certainly signal 2027 capex above $200B. Morgan Stanley recently raised its 2027 hyperscaler capex forecast to $1.2 trillion. The 2027 number is the real information.
The 2.5x model agrees. The fleet needs to grow from 47.6 GW to 75.9 GW by end-2027. The inference revenue pool scales to over $1 trillion annualized. Supply-cover stays tight at 1.40x. At those numbers, $200B in 2026 is not the ceiling. It's the floor.
What the Model Expects
| Signal | Reading |
|---|---|
| Token demand growth | 2.5x/year |
| Supply-cover (Q4 2026) | 1.17x (tight) |
| B200 project IRR | 35.7% |
| B200 equity IRR | 77.5% |
| AWS growth | 28% YoY |
| AMZN current capex guide | $200B (2026) |
| GOOGL 2027 signal | "Significantly increase" |
| Model capex prediction | Reaffirm $200B for 2026, signal $220–240B for 2027 |
The model assigns a roughly 60% probability that Amazon reaffirms the $200B 2026 number but signals $220–240B for 2027. The 2026 number is already enormous, a 50% increase from 2025, and the shock value is already priced in. The forward trajectory is the new information.
The Overhang
Amazon's Q1 2026 free cash flow was negative $18.2B. Capex consumed 24% of revenue. Total debt is $210B against $143B in cash. The credit default swap market is pricing stress across all hyperscalers. If AWS growth decelerates or the AI revenue story stalls, the capex becomes a liability rather than an asset.
But the 2.5x model says the demand is real. Token consumption is growing at 2.5x per year. The inference revenue pool is scaling toward $1 trillion. B200 GPUs deployed at AWS earn 35.7% project IRRs. At those numbers, the capex is not spending. It's investment.
What to Watch Tomorrow
The 2.5x model says the market's obsession with the 2026 capex number is missing the point. The real signal is the 2027 trajectory. If Amazon's CFO says anything that implies 2027 capex above $220B, and AWS growth holds at 28% or better, the stock can rally even with the $200B number. If AWS decelerates, the capex suddenly looks like a problem, and the selloff in the chart above continues.
More research at bargo.ai/research.