Big Tech’s AI Revenue Is Surging, but Suppliers Will Still Be the Bigger Winners

• As analysts call for AI capex to hit $1 trillion in 2027, Big Tech showed strong progress on AI monetization in Q2.
• Despite strong increases for AI revenue un rates, Big Tech’s capex to AI revenue ratios remain elevated from 1.5X to 4X.
• Higher AI capex extends the demand tailwind for suppliers which have already seen earnings growth multiples higher than the market and hyperscalers.

Big Tech is finally proving that AI can generate meaningful revenue. Azure, Google Cloud and AWS all accelerated in Q2, reaching tens of billions in dollars in annualized revenue.

Meanwhile, as is well-reported by now – Google and Amazon have posted negative free cash flow in Q2, as capex continues to rise faster than the cash generated from these investments. As discussed in the article AI Capex to Hit $1 Trillion – And Estimates Are Still Too Low, improving AI economics may not reduce spending, rather it could accelerate it to $7.6 trillion spent between 2026-2031.

This creates a question the market will be debating for years – if AI monetization is getting better, then why are Big Tech’s cash flows getting worse? In other words, why can’t monetizing AI offset these investments?

That is the kind of question that short sellers will salivate over, yet remember; short sellers are not the winners of the AI boom. Instead, the stock market winners have decisively been those who positioned in lesser-known suppliers.

In the analysis below, we spell out this important dynamic for AI investors. And as the I/O Fund has done for years and years, we also connect the dots as to who the real AI beneficiaries will be.

Cloud Growth Accelerates Across the Board in Q2 Driven by AI 

Last week’s Q2 earnings provided several important pieces of evidence that AI is driving cloud accelerations at Azure, Google Cloud and AWS, as AI run rates and key metrics continue to tick higher.

Microsoft Azure Growth Accelerates as AI Demand Outstrips Capacity 

Starting with Microsoft, we noted in our Q2 Big Tech earnings preview that its AI business hit a $37 billion run rate in Q3 FY2026, up 123% YoY and a nearly 3X increase versus early 2025. Despite that, Azure had failed to accelerate with growth consistently sitting near 39%-40%.

We noted that accelerating Azure growth would be key to beating expectations, and this is exactly what transpired. Azure accelerated to 43% growth, above guidance for 39-40%, with Microsoft expecting Azure to accelerate further to 45% growth next quarter.

Microsoft also disclosed that the number of Foundry customers operating at 1 trillion token run rates rose 4X YoY. Paid Copilot seats rose to 30 million, with net additions doubling QoQ. Microsoft also added $51 billion in commercial RPO, with no contribution this quarter from frontier labs, with the overall RPO figure hitting $678 billion, up 84% YoY.

Google Cloud Growth Accelerates to 82% YoY 

Google also put up impressive growth metrics, with Google Cloud accelerating to 82% YoY growth. This was a sharp acceleration compared to 63% growth in the prior quarter, with growth also over 2.5X higher than the 32% YoY achieved in Q2 2025. Meanwhile, Cloud operating profit soared 214% YoY to $8.8 billion as operating margin expanded 1,490 basis points YoY and 270 basis points QoQ to 35.6%. Cloud backlog increased over $50 billion sequentially and 385% YoY to $514 billion.

Additionally, as we recently noted in our latest piece on token processing, Google processed 3.2 quadrillion tokens in May, up 7X in one year and up a whopping 330X in two years. The company also said its first-party model token processing hit 22 billion per minute in Q2, matching Q1’s growth rate at a 6 billion increase sequentially.

AWS Posts Fastest Growth in Over 4 Years, Chips + AI Run Rate Hits $50B Combined 

The trend of accelerating cloud growth continued with Amazon’s AWS, which marked its fastest growth rate in more than 18 quarters in Q2. Revenue increased by 36.7% YoY to $42.2 billion, more than double the 17.5% growth in Q2 2025, and a more than 8 point acceleration versus 28% in Q1. AWS operating income also rose 58.9% YoY to $16.6 billion, while operating margin expanded 630 basis points YoY and 160 basis points QoQ to 39.4%.

In just one quarter, AWS’s AI business saw its annual revenue run rate increase from $15 billion to over $25 billion, growing triple digits YoY. Additionally, the run-rate of its chips business eclipsed $25 billion, also growing triple-digits YoY and 25% QoQ and led by its Trainium accelerators and Graviton CPUs. Combined, this represents a $50 billion run rate across chips and AI services, or 30% of AWS’s total $169 billion annualized run rate.

Notably, CEO Andy Jassy said there is a “real chance” that Amazon will sell Trainium chips to third-party data centers in the future, opening the door to increase its chips run rate further as the company indicated in April that a third-party approach could’ve seen its run rate already hit $50 billion.

Meta Advantage+ Reaches $75 Billion Run Rate Despite Slower Revenue Growth 

Of all the Big Tech reports, Meta was decidedly the least impressive growth-wise. Total revenue rose 28% YoY, decelerating considerably versus 33% in the prior quarter, while its Q3 guidance also implies a further deceleration to 22% growth.

Still, this doesn’t mean that Meta isn’t seeing significant AI monetization. The company notes that its end-to-end AI-powered Advantage+ solutions hit a $75 billion annual run rate during the quarter. This marks a 16.7% increase versus Q3 2025, and a 3.75X increase versus Q4 2024.

AI Capex-to-Revenue Ratios Are Sky High Across Big Tech 

Looking at quarterly AI revenue run rates versus capex as a simple barometer for AI ROI across Big Tech shows a marked improvement in ROI over the last two years, though run rates still lag capex by a wide degree.

As noted above, AWS saw its AI services and chips businesses both hit a $25 billion run rate in Q2, or ~$6.25 billion per quarter each (assuming run rate is calculated quarterly). Combined, both AI-driven businesses would be contributing $12.5 billion in quarterly run rate revenue in Q2, versus capex at $53.1 billion, or a 4.25X capex to AI revenue ratio.

Looking back to Q1 2024 when Amazon revealed it reached a multi-billion run rate for AI, and assuming this would be roughly $4 billion annualized or $1 billion quarterly, this would give a 13.9X capex to AI revenue ratio. This suggests AWS has meaningfully improved its AI ROI over the last two years as its AI businesses ramp despite quarterly capex scaling 4X from $13.9 billion to $53.1 billion.

For Microsoft, no new update was given for its exact AI run rate in Azure, though extrapolating from FQ3’s update of reaching a $37 billion run rate projects FQ4’s run rate to be around the $45 billion range. This would correspond to quarterly AI revenue of $11.2-11.3 billion, giving a capex to revenue ratio of 3.64X this past quarter. Looking back to FQ2 2024 (ending Jan 2024), Microsoft’s AI run rate is estimated to be around $4.75 billion, or around $1.19 billion quarterly, while capex was $11.5 billion. This corresponds to a capex to AI revenue ratio of 9.75X.

Unlike peers, Alphabet has not been very upfront with its AI run rate, though assuming AI contributed a similar proportion of YoY growth as Microsoft in Q2, this would roughly project Google Cloud’s AI run rate to be ~$47 billion, or ~$11.7 billion per quarter. This would correspond to a capex to AI revenue ratio of 3.84X. Looking back to Q2 2024 where Alphabet first disclosed its YTD AI revenue in the billions, and assuming this would be approximately $1.5 billion that quarter, this would give a capex to AI revenue ratio of 8.8X.

For Meta, Advantage+ reached a $75 billion run rate, implying a quarterly run rate of $18.75 billion. Compared to Q2 capex of $31.1 billion, this corresponds to a capex to AI revenue ratio of 1.66X. Looking back to Q4 2024’s $20 billion run rate, or $5 billion quarterly, offers a capex to AI revenue ratio of 2.84X.

Free Cash Flow Headwinds Mount as Infrastructure Spending Outpaces Revenue 

Prior to this earnings season, we published our two-part series on capex and free cash flows, Big Tech’s Free Cash Flow is Turning Negative – Who’s Next?. In this, we discussed how capex growth was outpacing both operating cash flow growth and the scale of AI monetization, concluding that Big Tech was at high risk of going FCF negative this year and next.

Q2’s reports showed this dynamic arise earlier than expected, as Google saw FCF go negative for the first time since its IPO, followed by Amazon the week after, while Meta held on by its fingertips this quarter.

Looking ahead to 2027, FCF pressure remains, as both Google and Meta are forecast to see FCF drop deeper into negative territory with Amazon also on the brink of negative FCF. On the flipside, Microsoft is expected to be relatively well-insulated with FCF remaining resilient through next year at $46.2 billion estimated in 2027.

In Q2, Amazon put it rather bluntly to investors – short term FCF headwinds are necessary and being incurred as capacity must be built out to meet demand, but this is expected to create stronger long-term FCF growth once it hits the break-even point for servers.

Semiconductor and Hyperscaler Earnings Growth Is Rapidly Diverging 

As hyperscalers spend massive sums on AI capex, I/O Fund’s strategy has centered around investing in the companies that are receiving that cash rather than those spending it. Looking at the growth of semiconductor earnings versus hyperscalers and the overall market provides strong validation as to why we have taken this approach.

Data from JP Morgan shows that annual semiconductor earnings growth has outpaced hyperscalers for multiple years, and that divergence is only widening over time. Semiconductor earnings growth moderately exceeded the ~40% among hyperscalers in 2024, but then increased to ~50% in 2025, more than 2.5X the 19% growth hyperscalers achieved. That gap is expected to become much larger in 2026, with semiconductor earnings expected to rise by 97%, nearly 6X faster than hyperscaler earnings.

Semiconductor PEG Ratio Drops Steeply, Supporting AI Supplier Thesis 

While shares of many semiconductor stocks have soared, soaring earnings have also put significant downward pressure on valuations. This is demonstrated through the trailing PEG ratio, which divides P/E ratios by EPS growth to help assess whether multiples are in line with earnings growth.

The trailing PEG ratio among chip stocks has fallen from well above 2 to nearly 1 over the past year, lower than other subsectors tied to the AI buildout. While these estimates are as of early June, semiconductor benchmarks are down meaningfully in that span, likely pushing trailing PEG ratios down further. As inference demand is likely to lead to even higher levels of capex, semiconductor earnings can receive another strong tailwind in 2027, reinforcing the merits of our semiconductor-focused approach.

Conclusion

Big Tech is finally proving that AI can generate meaningful revenue, but Q2 also showed why stronger monetization may not translate into stronger near-term cash flows. Azure, Google Cloud, and AWS all accelerated, and AI businesses across both cloud and core products are reaching run rates well into the tens of billions.

The key takeaway is that improving AI economics may reinforce the spending cycle rather than slow it down. From what is being communicated from Big Tech management teams, revenue will be reinvested into more data center capacity. This could lead to a few more years of elevated capex and free cash flow pressure.

From our vantage point, however, the focus on Big Tech is a distraction. While hyperscalers grab the flashy headlines, the more compelling opportunities are the enabling technologies standing directly in the path of a nearly $1 trillion wind tunnel of annual capex.

Our latest 90-page Top 20 AI Stocks for Q3 2026 report offers investors a comprehensive deep dive into the AI stack, mapping out the companies best positioned to capture capex spend. Previous winners identified in the report include Bloom Energy up 1150% since our first entry, Micron up 210% since our entry a few months ago, a lesser-known networking stock up 370% since November. All of this from the team that first identified Nvidia as an AI stock in 2018, up 6700% since our first entry.

Don’t miss out on the AI trade. Subscribe Now.

Please note: The I/O Fund conducts research and draws conclusions for the company’s portfolio. We then share that information with our readers and offer real-time trade notifications. This is not a guarantee of a stock’s performance and it is not financial advice. Please consult your personal financial advisor before buying any stock in the companies mentioned in this analysis. 

Leo Miller, AI and Semiconductor Investment Writer at I/O Fund, contributed to this analysis.

👉🏻 Share with a Fellow Investor
Help someone else benefit from this insight.

Recommended Reading:

• Token Growth is Surging – Here Are the Beneficiaries
• AI Token Demand is Shattering Forecasts
• Nvidia and Google Are Crowding TSMC’s N3 Node – Can Intel Fill the Gap?
• Intel vs TSMC: How CoWoS Packaging Constraints Could Create an Opportunity for Intel Foundry