August underscored the same theme we’ve been pounding the table on: artificial intelligence is not just a trend – it’s the central force reshaping earnings, regulation, infrastructure, and even the power grid.
From NVIDIA’s blowout results to Google’s antitrust ruling, to MIT’s cautionary report, and finally the scramble for electrons to feed AI workloads, the past month has given us another set of tea leaves to sift through on both sides of the long-term thesis.
NVIDIA Earnings: The AI Toll Collector Keeps Printing
NVIDIA has once again put up numbers that looked more like a hyperscaler than a chip company. I’ll get right to the punchline: This much growth, at this scale, with these margins, is unprecedented, likely in the history of the world. While we can debate which inning of the buildout we are in (I believe we are in the 5th or 6th inning), the AI theme is still alive and well.
For Q2 FY26, company revenue hit $46.7B (+56% YoY), with data centre sales coming in at $41.1B, by far the bulk of the business. Gross margins came in at 72%, underscoring the software-like economics of selling the picks and shovels for the AI buildout. Guidance for Q3 of $54B suggests the demand curve remains exponential.
The company’s Blackwell microarchitecture drove sequential growth of 17%, with networking called out as a standout line item. No H20 GPU shipments were booked into China due to export restrictions, but even with that headwind, the order book remains sold out into 2026. NVIDIA also authorized a $60B buyback, a reminder that this company throws off so much cash it can both fund hyper-aggressive R&D roadmaps and still return capital to shareholders.
Key Takeaway: The cadence here matters. Rubin – the next generation after Blackwell – is already in flight, with expectations of doubling performance. In short, NVIDIA owns the full ecosystem (GPU, networking, software), and it’s forcing the rest of the supply chain to catch up.
For now, it looks like no one is catching them. Not even close.
Google’s Antitrust Case: The Gavel Falls on Distribution
The Department of Justice’s long-running search case against Google finally landed in August with remedies that, while stopping short of a breakup, strike directly at its distribution model. Google will not be forced to spin off Chrome or Android, but it can no longer pay billions for exclusive default search placement, and it must share its search data index with rivals.
While the market largely cheered this ruling as favourable (which it is, overall), there are underlying implications we need to pay attention to. For the AI space, the ruling is critical. It explicitly bars Google from extending exclusivity practices into AI search and assistants. That means competitors like OpenAI, Microsoft, or even Apple can get distribution opportunities that were previously locked. Access to query data could also narrow the training gap for challengers, improving their AI models.
Key Takeaway: Alphabet stock rallied on relief that Chrome and Android were spared, but the long-term implication is that Google will have to effectively disrupt itself in the new age of AI, which it very well may pull off. Competition in AI search is now codified into law, and that creates tailwinds for emerging answer engines and AI assistants.
MIT Report: Pilots Are Cheap, but 95% Still Fail
A new MIT study dropped in August with a sobering reminder: efficiency gains in AI mean nothing if enterprises can’t actually implement them. The market saw this as a warning sign, but we’re still early here. MIT found that more than 95% of AI pilot projects fail to scale into production. This is less about the math and more about the messy reality of organizational adoption.
The report highlights three culprits:
- Data quality and fragmentation: Companies underestimate how hard it is to centralize clean, labelled datasets.
- Talent bottlenecks: There simply aren’t enough machine learning engineers and AI product managers who can shepherd pilots into workflows.
- Integration pain: Legacy IT stacks are brittle. Plugging AI into them is like trying to install a Tesla battery into a horse-drawn carriage.
This failure rate matters because it tempers the narrative of linear AI adoption. Enterprises aren’t going from zero to fully autonomous overnight; they’re stubbing toes, burning budgets, and circling back.
But here’s the paradox: these failures may actually increase long-term demand for infrastructure. Every failed pilot is still compute-consumed, and every retrial drives more workloads back to the hyperscalers and GPU clouds. Over the long term, these pilots will materialize into longer-term commitments, but we need more proof.
Key Takeaway: Efficiency gains at the model level don’t guarantee efficiency at the company level. For now, AI is still a land grab of experiments. Investors should still focus on the picks-and-shovels providers (semiconductors, cloud and data infrastructure) who get paid whether the pilots succeed or not. Application revenue is coming, but it’s not here yet (outside of Microsoft, OpenAI, and Anthropic).
Megacaps Earnings: The Atlas of the S&P500
Stop me if you’ve heard this one before: the Mag7 continues to carry the market. In 2024, they accounted for ~73% of all S&P 500 earnings growth. In 2025, they’re on track to contribute 50%, despite making up just 25% of earnings weight and ~39% of market cap.
Once again, AI hyperscalers and infrastructure companies continue to drive results. Microsoft reported Azure growth of 39%, with AI workloads serving as the key driver. Meta posted 22% revenue growth and 36% higher profits as AI-driven engagement boosted ad yields. Amazon grew 12% with AWS at $30.9B revenue, still expanding. Google Cloud surged 32% YoY to a $54B run-rate. NVIDIA’s results speak for themselves.
Key Takeaway: Breadth remains historically narrow. Just two names, Microsoft and NVIDIA, represented 42% of S&P 500 gains in the first half of 2025. This concentration is not just narrative; it’s fundamental. Until earnings power broadens, the megacaps will remain both the market’s ballast and its beta.
Energy Complex: The New Limiting Factor
If semiconductors are the brains of AI, electricity is the bloodstream, and we’re starting to see strain. Data centres consumed ~460 terrawatt-hours (TWh) last year, and forecasts have that doubling by 2026. A single hyperscale AI campus now demands over a gigawatt of power, equivalent to a nuclear reactor or one million homes.
The near-term bridge is natural gas. Utilities in Virginia, Texas, and the Southeastern United States are rushing to add capacity, with Duke Energy saying over half its project queue is now data centres. Roughly two-thirds of new planned gas plants are directly tied to AI-driven demand.
The long-term play is nuclear. We’ve now seen Big Tech cement deals measured in gigawatts: Amazon’s $18B contract for 1.9 GW of nuclear power, Meta’s 1.1 GW agreement in Illinois, Microsoft funding the restart of Three Mile Island for 0.8 GW, and Google backing small modular reactor (SMR) projects. These are staggering numbers, and they underscore that powering AI sustainably is now as important as building the GPUs themselves.
Key Takeaway: Bottlenecks are already here. Dominion in Virginia had to slow new data centre hookups, and DOE models forecast potential hundreds of hours of shortages in hotspots by 2030 if new supply isn’t built. Energy is no longer a side story; it’s a core risk and opportunity in the AI trade.
Wrapping Up
August reminded us that AI’s influence is everywhere: NVIDIA minting record profits, Google forced to open its gates, AI model efficiency breakthroughs broadening adoption, megacaps carrying the earnings load, and energy grids bending under new demand curves.
The secular thesis remains intact: more compute, more adoption, more power. Volatility will come, but the long-term direction is clear. We remain overweight the infrastructure (NVIDIA, Broadcom, Arista), the hyperscalers (Microsoft, Google, Amazon, Meta), and selective energy plays (Vistra, Constellation, nuclear developers) that enable this transition.
The AI rocket is still on the launchpad, and August showed us the boosters are firing.
Strong Convictions. Loosely Held.
-Nicholas Mersch, CFA
The content of this document is for informational purposes only and is not being provided in the context of an offering of any securities described herein, nor is it a recommendation or solicitation to buy, hold or sell any security. The information is not investment advice, nor is it tailored to the needs or circumstances of any investor. Information contained in this document is not, and under no circumstances is it to be construed as, an offering memorandum, prospectus, advertisement or public offering of securities. No securities commission or similar regulatory authority has reviewed this document, and any representation to the contrary is an offence. Information contained in this document is believed to be accurate and reliable; however, we cannot guarantee that it is complete or current at all times. The information provided is subject to change without notice.
Commissions, trailing commissions, management fees and expenses all may be associated with investment funds. Please read the prospectus before investing. If the securities are purchased or sold on a stock exchange, you may pay more or receive less than the current net asset value. Investment funds are not guaranteed, their values change frequently, and past performance may not be repeated. Certain statements in this document are forward-looking.
Forward-looking statements (“FLS”) are statements that are predictive in nature, depend on or refer to future events or conditions, or that include words such as “may,” “will,” “should,” “could,” “expect,” “anticipate,” intend,” “plan,” “believe,” “estimate” or other similar expressions. Statements that look forward in time or include anything other than historical information are subject to risks and uncertainties, and actual results, actions or events could differ materially from those set forth in the FLS. FLS are not guarantees of future performance and are, by their nature, based on numerous assumptions. Although the FLS contained in this document are based upon what Purpose Investments and the portfolio manager believe to be reasonable assumptions, Purpose Investments and the portfolio manager cannot assure that actual results will be consistent with these FLS. The reader is cautioned to consider the FLS carefully and not to place undue reliance on the FLS. Unless required by applicable law, it is not undertaken, and specifically disclaimed, that there is any intention or obligation to update or revise FLS, whether as a result of new information, future events or otherwise.