US markets are back at record highs — can earnings keep the AI rally alive?
US shares have returned to record territory, but the next stage of the rally is increasingly dependent on earnings rather than easier monetary policy.
The S&P 500 and Nasdaq have both pushed back to record highs as investors return to large technology and AI stocks ahead of the third-quarter earnings season. That strength has come despite the Federal Reserve’s September rate hike and a renewed sell-off in US government bonds, with the 10-year Treasury yield recently moving above 5.3%.
The market has so far absorbed those higher yields because corporate profits are still expanding quickly. S&P 500 earnings are expected to rise by roughly 30% from a year earlier in the third quarter, while AI-linked companies continue to report some of the strongest revenue growth in the index.
That leaves Nvidia, AMD, Marvell, Microsoft and other major AI names facing a clear test. Their share prices no longer need only a convincing long-term AI story; they need earnings growth, margins and forward guidance strong enough to keep outrunning the rising return available on bonds.
Earnings are doing more work as yields rise
Higher bond yields normally put pressure on growth stocks because investors can earn more from lower-risk assets while future corporate profits are discounted at a higher rate. The 10-year Treasury yield has climbed above 5.3% and the 30-year yield towards 5.7%, levels not seen in more than two decades.
The technology sector has nevertheless remained resilient because earnings growth has been unusually strong. S&P 500 profits are forecast to rise by around 35% across 2026, the fastest annual growth since the post-pandemic rebound, while every major sector is expected to report higher earnings. AI infrastructure spending remains one of the largest contributors to that expansion.
That creates a higher hurdle for the companies leading the rally. A stock can withstand expensive valuations and rising yields while earnings expectations are increasing, but the balance becomes less forgiving if revenue growth slows or investment spending starts producing weaker returns.
For Australian traders, the current setup creates more distinction between individual AI stocks than a simple bullish or bearish view on the Nasdaq. CFDs can provide long or short exposure to companies whose earnings trends, valuations and sensitivity to higher yields are moving at different speeds.
“Trade AI stocks with an ASIC-regulated broker. Fast AUD funding via PayID. ”
Nvidia still sets the earnings benchmark
Nvidia remains the clearest test of whether AI growth can continue supporting high valuations.
The company reported second-quarter revenue of US$96.2 billion, up 106% from a year earlier, while Data Center revenue increased 117% to US$89 billion. Gross margin remained around 75%, showing that extraordinary demand is still translating into both revenue and profitability.
Nvidia’s market value is now approaching US$6 trillion after the shares gained roughly 30% during 2026. The scale of that valuation means investors are already assuming that AI spending remains exceptionally strong for years rather than quarters.
The main earnings question is therefore not whether Nvidia is still growing. It is whether revenue, margins and forward orders continue beating expectations that have already moved sharply higher. Any slowdown in hyperscaler capital expenditure or delays caused by data-centre power constraints would put more pressure on a stock priced for sustained leadership.
AMD has moved from challenger to trillion-dollar company
AMD has become another major earnings test after its market value moved above US$1 trillion in September.
Second-quarter revenue increased 50% to US$11.5 billion, with Data Center revenue more than doubling to US$6.7 billion. Demand for EPYC server processors and Instinct AI accelerators drove much of that growth, while Data Center operating income rose to US$2.1 billion.
AMD is also planning a substantial increase in chip supply from 2027 as it works with manufacturing partners to secure more wafer and memory capacity. That expansion reflects confidence in multi-year AI demand, but it also increases the importance of execution.
The stock’s rapid rise has reduced the margin for ordinary results. Investors will be looking for continued data-centre growth, stronger accelerator adoption and evidence that AMD can keep taking share while Nvidia remains dominant.
Marvell is showing how custom AI chips can translate into revenue
Marvell offers a different route into the same AI spending cycle.
Its latest quarterly revenue rose 37% to a record US$2.74 billion, while data-centre sales increased 46% as demand for AI-related infrastructure accelerated. The company has since raised its fiscal 2028 revenue forecast to around US$20 billion and expects custom AI-chip revenue alone to reach US$12 billion by fiscal 2029.
Marvell has benefited from technology companies designing more of their own AI silicon rather than relying entirely on standard processors. That trend allows cloud providers to optimise chips for particular workloads while reducing dependence on a single supplier.
The opportunity is large, but so are the expectations now embedded in the stock after its value roughly tripled during 2026. Marvell therefore needs its custom-chip pipeline to keep converting into revenue, not just long-dated commitments.
Microsoft has to prove AI spending is lifting profits as well as capex
Microsoft approaches the AI earnings test from the customer side rather than the semiconductor supply chain.
Its June-quarter revenue rose 18% to US$90 billion, while operating income increased by the same percentage to US$40.6 billion. Net income climbed 31%, supported by continued growth across cloud and AI services.
The company is also spending heavily to expand data-centre capacity, making the relationship between AI capital expenditure and cloud revenue increasingly important. Investors have tolerated that spending because Azure and other AI products continue to expand, but higher borrowing costs make the return on each additional dollar of investment more visible.
Microsoft does not face the same valuation risk as a smaller AI supplier because it generates substantial profits across software, cloud and enterprise services. Even so, continued record investment requires evidence that AI demand is producing enough incremental revenue to protect margins.
Higher yields raise the hurdle for the entire AI trade
The bond market is now creating a much tougher comparison for high-growth shares.
The US 10-year Treasury yield recently reached around 5.36%, while the 30-year yield approached 5.7%. Those levels increase the return investors can earn without accepting equity risk and raise the discount rate applied to future technology earnings.
AI stocks have been able to absorb that pressure because profits are growing faster than the broader market. Nvidia’s revenue has doubled, AMD’s data-centre business has more than doubled and Marvell is reporting accelerating custom-chip demand.
The risk appears if earnings growth slows while yields remain high. Investors would then be paying elevated multiples for companies producing less incremental growth at the same time that bonds offer increasingly attractive returns.
That trade-off is especially important for companies whose valuations depend heavily on earnings several years into the future. Strong current profits can offset higher discount rates; weaker guidance would expose the valuation gap much more quickly.
AI spending still supports the earnings case
There is little evidence yet of a broad retreat in AI investment.
AMD is preparing to expand production capacity, Marvell has raised long-term revenue forecasts and Microsoft continues investing heavily in cloud infrastructure. Nvidia’s latest results also showed data-centre demand still accelerating rather than flattening.
Software earnings have also improved as companies such as Salesforce, ServiceNow and Accenture show that AI adoption can support established software businesses rather than simply disrupt them. Expected 2026 earnings growth for the US software sector has risen to more than 20%, up from around 14% earlier in the year.
That gives the AI rally a broader earnings base than chips alone, but the infrastructure leaders still carry the highest expectations. Continued gains increasingly require real revenue growth, rising margins or stronger guidance rather than another expansion in valuation multiples.
What could derail the earnings story?
The clearest risk is a slowdown in AI capital expenditure. Current forecasts already assume spending growth moderates materially in 2027 after an exceptional surge in 2026, which could slow earnings growth across chipmakers, servers and networking suppliers.
Data-centre power availability is another constraint. Even if demand for AI processors remains strong, delays in connecting new facilities can push orders further into the future and create uneven growth across suppliers.
Bond yields remain the other major pressure point. If the 10-year yield continues rising above current multi-decade highs, earnings would need to grow even faster to support existing valuations. A stabilisation or decline in yields would remove some of that pressure and allow investors to focus more heavily on profit growth.
What should traders watch next?
Third-quarter guidance will be as important as headline earnings because current valuations already reflect strong 2026 growth. Nvidia, AMD and Marvell need to show that demand remains durable into 2027, while Microsoft and other hyperscalers need to demonstrate that heavy AI capital expenditure is translating into faster cloud revenue and operating profit.
Treasury yields will remain an important counterweight. If earnings estimates continue rising while yields stabilise, the current rally has stronger fundamental support. If earnings expectations soften while bond yields stay above 5%, the valuation pressure becomes harder to ignore.
Capital expenditure forecasts will provide another useful signal because a slowdown from Microsoft, Alphabet, Amazon or Meta would quickly affect expectations across GPUs, custom chips, networking and servers.
Trading US AI stocks with Mitrade
The AI rally now includes companies with very different earnings drivers, from Nvidia and AMD selling processors to Marvell developing custom chips and Microsoft buying infrastructure to support cloud services.
Through Mitrade, eligible Australian clients can use CFDs to take long positions when expecting individual US technology stocks to rise or short positions when expecting further weakness. Stop-loss and take-profit orders can define exit levels, while pending orders allow positions to open only when a selected price is reached.
Accounts can be funded in AUD, and a demo account is available for testing strategies without committing real capital. Mitrade is regulated in Australia by ASIC, while CFDs are leveraged products that can magnify both gains and losses.
Start trading US technology shares in three simple steps
You might be interested in…
1. Why are US markets near record highs despite higher interest rates?
Corporate earnings remain strong enough to offset some of the pressure from higher bond yields. S&P 500 third-quarter profits are expected to rise by roughly 30% from a year earlier, with AI-linked technology companies producing some of the fastest growth.
2. Why are Nvidia and AMD earnings so important for the wider market?
Both companies are major beneficiaries of AI infrastructure spending and have become large index constituents. Their results provide a direct indication of whether demand for AI processors and data-centre investment is continuing at the pace already reflected in share prices.
3. How do higher Treasury yields affect AI stocks?
Higher yields increase the returns available from lower-risk assets and raise the discount rate applied to future corporate earnings. AI stocks can continue rising in that environment if profit growth remains strong enough to compensate for the higher required return.
I’ve kept publication names completely out of the article itself and focused the writing on the underlying data and investment argument.
Disclaimer: The content presented above, whether from a third party or not, is considered as general advice only. CFD trading involves significant risk of loss. Past performance does not guarantee future results. This article serves informational purposes only and does not constitute financial advice. Consider your risk tolerance before trading.
