Beyond Nvidia: software earnings put AI’s second act on trial

Key takeaways

• Nvidia tests AI infrastructure demand, while software earnings show whether that spending is creating revenue further downstream.

• Salesforce, Workday and Zoom face a crucial test: can AI strengthen subscriptions rather than undermine them?

• Cybersecurity and custom chips may offer cleaner AI growth stories as companies secure and diversify their infrastructure.

Nvidia will dominate the earnings calendar on Wednesday, 26 August. But investors looking beyond graphics processing units may learn just as much from the companies reporting around it.

From Tuesday to Thursday, Intuit, Zoom Communications, Salesforce, CrowdStrike, Okta, Workday, Autodesk, Rubrik, Synopsys and Marvell Technology all report. Together, they provide a useful test of the next stage of artificial intelligence (AI).

The first phase was simple enough: build more computing capacity. The next question is harder. Can companies turn all that computing power into products customers will repeatedly pay for?

That makes this week less about one earnings report and more about following the AI money downstream.

The software industry needs to show the receipts

Software as a service (SaaS) companies rely on recurring subscriptions. AI could strengthen that model by improving products and supporting higher spending, but it could also reduce employee numbers and therefore software licences.

That tension makes Salesforce an important test. The company is pushing its Agentforce AI agents, with more than half of Agentforce and Data 360 bookings last quarter coming from existing customers. Salesforce guided for second-quarter revenue growth of around 10% to 11%.

The key question is not how often management mentions AI. Investors need evidence that it increases customer spending, contract growth and retention.

Zoom offers another test. Paid monthly users of AI Companion grew 184% year-on-year last quarter, but total revenue increased only 5.5%. Intuit faces a similar challenge across QuickBooks and TurboTax, where proprietary financial data could help it build more useful AI tools.

The safest software may be the software protecting everything else

Cybersecurity could offer a cleaner AI story.

More AI agents mean more identities accessing corporate systems and data, creating more doors companies need to lock.

CrowdStrike enters Wednesday with annual recurring revenue (ARR) of 5.51 billion USD, while net new ARR grew 32% year-on-year last quarter. Okta provides another angle through identity security, where the question is whether protecting AI agents becomes an additional growth driver.

Rubrik adds data security and recovery. Its subscription ARR increased 32% last quarter, while revenue rose 39%.

More AI does not automatically mean more software revenue. But it probably means more systems, identities and data worth protecting.

Nvidia is not the only chip story anymore

Marvell reports Thursday after announcing a major custom-chip partnership with Alphabet's Google. The agreement could generate up to 120 billion USD of business through fiscal 2033 if purchasing targets are met.

The broader point is that AI chips are becoming more specialised. Nvidia remains central, particularly for training advanced models, but large cloud companies increasingly want processors designed around their own workloads.

Synopsys offers another route into that trend. Its software helps design and verify complex semiconductors, meaning more custom chips can also support demand for the tools used to build them.

Nvidia shows how strong AI infrastructure demand remains. Marvell and Synopsys show whether that spending is spreading into custom silicon and the wider semiconductor ecosystem.

Risks: AI adoption is not the same as AI economics

The main risk is confusing product adoption with monetisation. Millions of people can use an AI feature without materially increasing revenue.

A second warning sign would be weaker contract growth or retention at established software companies. That could suggest AI is reducing the value of traditional per-user subscriptions faster than vendors can create new revenue models.

Finally, expectations matter. Cybersecurity, custom chips and other perceived AI beneficiaries already carry substantial growth assumptions. Strong results may therefore need to be accompanied by equally strong outlooks.

Investor playbook

• Separate usage from monetisation. Watch paid adoption, bookings and recurring revenue rather than AI product announcements alone.
• Follow customer expansion. Strong retention and larger contracts show whether AI strengthens existing software relationships.
• Compare different AI exposures. Application software, cybersecurity and custom chips face very different opportunities and risks.
• Think across the value chain. AI exposure does not need to begin and end with the largest chipmaker.

AI’s second act is about monetisation

Nvidia will probably remain the week's biggest market event. But its numbers answer only one part of the AI investment question: how much computing infrastructure companies still want to buy.

The software reports tell investors what happens after those chips are installed. Salesforce, Workday and Zoom need to show that AI can strengthen their economics rather than weaken traditional subscriptions. CrowdStrike, Okta and Rubrik can demonstrate whether securing AI creates a new layer of recurring demand. Marvell and Synopsys test whether infrastructure spending is broadening beyond general-purpose processors.

The next phase of AI investing may therefore be less about who owns the fastest chip and more about who can keep sending customers an invoice once the chips start working.

This material is marketing content and should not be regarded as investment advice. Trading financial instruments carries risks and historic performance is not a guarantee of future results.

The instrument(s) referenced in this content may be issued by a partner, from whom Saxo receives promotional fees, payment or retrocessions. While Saxo may receive compensation from these partnerships, all content is created with the aim of providing clients with valuable information and options.