### Why the chart still shows acceleration despite real bottlenecks
The graphic is a **smoothed, long-term historical view** (1900–2025) of calculations per second per constant dollar. It compresses decades of progress into one clean line. Every previous computing paradigm hit massive bottlenecks (mechanical parts, vacuum tube reliability, transistor scaling limits, power/heat, etc.) and yet the overall curve kept bending upward.
Current bottlenecks (power grid constraints, transformer shortages, data center construction delays of 2–5 years, chip supply, etc.) are **very real and painful in 2025–2027**. They are exactly what we’ve been discussing with the hyperscaler capex numbers and the recent earnings reactions. These frictions create short-term lumpiness and slower-than-hoped deployment.
### Will the bottlenecks be solved in time?
**Yes, they are expected to be solved — just not overnight.** Here’s why the long-term acceleration thesis (the 10x/year slope) remains intact:
- **Historical pattern repeats**: Every computing era solved its own bottlenecks through innovation. We are now in the middle of the biggest shift yet — from general-purpose chips to highly specialized GPUs, TPUs, custom ASICs, and eventually optical/neuromorphic computing. This is what drives the steeper part of the curve on the right.
- **Solutions already in motion**:
- Massive new power projects (gas peakers, small modular nuclear reactors, renewables + battery storage) are being fast-tracked in the US and elsewhere.
- Hyperscalers are signing long-term power purchase agreements and even building their own generation capacity.
- Software and algorithmic efficiency gains continue to multiply effective compute (one of the reasons the chart keeps rising even when raw hardware growth slows).
- New chip architectures and 3D stacking are delivering big performance-per-watt jumps.
- **Economic incentive is enormous**: The hyperscalers are spending hundreds of billions precisely because the ROI on more compute is so high. That money is pulling forward solutions to the bottlenecks faster than in previous eras.
### Bottom line
The bottlenecks are **real and will cause volatility and slower near-term rollout** (exactly what we saw in the April 29 earnings reactions and the capex fatigue). But they are viewed as **temporary speed bumps**, not structural limits. The chart’s long-term view assumes these will be overcome — just as every previous bottleneck was — so the overall exponential trend in compute-per-dollar continues.
This is why investors (All-In gang, Chamath, Diamandis, etc.) remain net bullish on the AI supercycle even while acknowledging the current pain. The acceleration isn’t guaranteed to be perfectly smooth, but the historical evidence and economic forces strongly favor it continuing.
Strong earnings
With 28% of stocks reported, 84% of companies have delivered higher-than-expected EPS. This is above the 5- and 10-year averages. In aggregate, companies are reporting EPS that are 12.3% above estimates. This is also higher than the 5- and 10-year averages.
Also, tech is at an all-time high, but its overall P/E is still lower than the October 2025 high.
Big tech earnings- MSFT, GOOGL, AMZN, META
**Summary of the April 29 hyperscaler earnings reaction**
Four hyperscalers reported earnings within minutes of each other on Wednesday, but the immediate after-hours reaction was sharply split. Only Alphabet rose right away, jumping roughly 7 percent, while Microsoft, Meta, and Amazon all sold off despite each company beating consensus estimates on both revenue and earnings.
Alphabet posted the cleanest and most impressive result in the group. Revenue rose 22 percent to 109.9 billion dollars and EPS jumped 82 percent to 5.11 dollars. Google Cloud accelerated sharply to 63 percent growth with revenue of 20 billion dollars and the backlog nearly doubled quarter-over-quarter to 460 billion dollars. The market rewarded the clear evidence that AI monetization is accelerating faster than expected and that Alphabet is turning heavy capex into visible cloud and ad cash flow.
Microsoft reported a solid beat with revenue at 82.9 billion dollars and EPS at 4.27 dollars. Azure growth reaccelerated to 40 percent and the AI business reached a 37 billion dollar annual run rate, up 123 percent year over year. Yet the stock gapped lower as investors focused on continued high capex and uncertainty around free cash flow timing.
Amazon delivered AWS growth at its fastest pace in 15 quarters with the segment up 28 percent to 37.6 billion dollars. Overall net sales rose 17 percent to 181.5 billion dollars and retail margins held up well. The stock initially sold off but later rallied as investors viewed the print as validation that the massive capex plan is beginning to pay off without major surprises.
Meta beat on revenue and EPS with strong ad business performance, but raised full-year 2026 capex guidance materially higher to 125–145 billion dollars. The increased AI infrastructure spend weighed on margin and free cash flow expectations, causing the stock to sell off sharply.
The clear market takeaway is that investors are now benchmarking every hyperscaler’s capex efficiency against Alphabet’s performance. Companies that cannot demonstrate comparable visible returns on their spending are being punished in the short term, even when underlying AI demand signals remain strong.
In short, the earnings cluster removed some near-term doubt about AI monetization and provided a modest tailwind to futures today, even if the capex debate is far from over. The move remains volatile and any fresh Iran headline or follow-through from individual stocks can still swing it quickly.
Oil up, futures up?
As for both oil and futures being up earlier today, futures are up despite surging oil because the oil spike is being read as a short-term geopolitical premium rather than a broad economic negative.
Oil prices jumped this morning on renewed fears of deeper U.S.-Iran conflict and continued disruptions in the Strait of Hormuz. Normally that would be a clear headwind for stocks (higher input costs, inflation worries).
But today the market is treating the oil move as **supportive** for two main reasons:
- **Energy sector leadership** — Higher crude directly boosts profits for oil producers, service companies, and refiners. That strength is spilling over and lifting the broader indices in a classic sector rotation.
- **Contained-risk narrative** — Traders are viewing this as another layer of the same known Iran story rather than a fresh escalation that changes the long-term outlook. The focus remains on the strong AI/cloud demand signals from yesterday’s hyperscaler earnings (especially Alphabet and Amazon) and the view that any oil-driven inflation will be temporary.
In short, the market is pricing the oil surge more as a commodity tailwind for energy names and a “risk-on” signal that the conflict remains contained, rather than a broad negative for growth stocks. This dynamic has let futures push higher even as oil climbed.
The move remains volatile — any real escalation or de-escalation headline can flip it quickly as we are seeing during the trading day at the time of this writing.