Nvidia’s latest AI revenue forecast is drawing closer scrutiny as investors look beyond headline growth and focus on whether spending will translate into meaningful returns.
BNN Bloomberg spoke with Nicholas Mersch, portfolio manager at Purpose Investments, about how the AI trade is shifting toward profitability and which parts of the market may benefit next.
Key Takeaways
- Investors are shifting from AI spending toward returns, prioritizing revenue growth, margins and cash flow over capital outlays.
- Nvidia’s $1 trillion forecast appears largely in line with existing expectations, limiting its impact as a catalyst for the stock.
- Rising debt levels and declining free cash flow among hyperscalers are raising questions about the sustainability of AI investment.
- Supply chain segments such as memory, storage and optics are benefiting from near-term demand tied to AI infrastructure.
- Software remains in a prolonged repricing phase as AI disrupts traditional business models and pressures margins.
Read the full transcript below:
LINDSAY: At its GTC conference, Nvidia’s CEO announced its flagship AI processors would help generate $1 trillion in revenue by 2027. But while the $1 trillion projection made for a flashy headline, our next guest says when you break this number down, it doesn’t show much of an acceleration from the company’s previous forecast.
Joining us now is Nicholas Mersch, portfolio manager at Purpose Investments, to dive into this. Good to have you with us this morning.
NICHOLAS: Thanks so much for having me.
LINDSAY: You say the AI trade is maturing from hype to return on AI. Do you think this $1 trillion is enough of a return to get investors excited?
NICHOLAS: Yeah, that’s right. I think this trade overall is really shifting as people start to question when AI capex can be converted into durable revenue, margin and cash flow.
That’s why the Magnificent Seven have been under significant pressure and entered a 10 per cent correction as a group. The market is starting to favour those collecting the cheques instead of those writing them.
Early on, the narrative was that all of this was free cash flow-funded, but now we’re seeing two things happen. First, free cash flow is starting to decline, and second, there’s an increasing amount of debt being issued by mega-cap companies. At the same time, private credit is showing cracks and spreads are widening in fixed income markets.
The key question is how quickly AI is monetized and how much of that spending converts into long-term, durable moats. I think Amazon and Microsoft are positioned well, and Google is probably positioned the best to capture longer-term upside.
On the $1 trillion Nvidia number, it’s a flashy headline, and analysts will spend time digging into what it actually means. The previous forecast was $500 billion by the end of 2026, and now it’s $1 trillion by the end of 2027.
So when you look at the cumulative number, it’s actually pretty close to consensus. Investors were looking for an incremental boost, but this didn’t really provide that.
That said, Nvidia is no longer just selling chips — it’s selling full data centre solutions. There could be upside from other business lines tied to that ecosystem, which could support earnings.
But overall, while the growth is still impressive, the market is starting to expect these numbers, and Nvidia is running up against the law of large numbers.
LINDSAY: You compared Nvidia to some of the hyperscalers you think are well positioned. Do you think Nvidia itself is well positioned against its competition?
NICHOLAS: Absolutely. What Nvidia is doing with its next generation of GPUs is stacking multiple S-curves. In technology, you typically see rapid growth, then a plateau, and then another wave driven by new innovation.
At GTC, we saw another step forward, with compute capabilities increasing roughly 35 times from the previous generation.
But again, the hyperscalers funding this — the ones writing the cheques — are facing pressure. Free cash flow is declining, cash balances are coming down, and they’re issuing more debt. So the question is where the next phase of growth comes from.
Another key area to watch is which parts of the ecosystem benefit. There were comments around whether copper or optics will be the key scaling components.
We’re seeing different parts of the supply chain move, with Nvidia acting as a kind of kingmaker. Memory and storage companies like SanDisk, Micron and Western Digital are benefiting because more compute requires more supporting infrastructure.
So while Nvidia may be facing scale challenges, it remains the leader in this space.
LINDSAY: Before we move on, what about Nvidia’s partnerships, including autonomous vehicle deals with Lyft and Uber? Are these meaningful for the company?
NICHOLAS: These partnerships often drive short-term stock reactions, similar to what we saw earlier with companies tied to OpenAI.
Some of them can create lasting competitive advantages, though. In autonomous vehicles, it’s especially interesting because companies are trying to define their role in the ecosystem.
For Uber and Lyft, labour is a major cost. If autonomy reduces that, it could significantly improve margins by lowering cost per mile.
At the same time, companies like Tesla are building their own ecosystems, leveraging large amounts of data from vehicles already on the road.
Nvidia plays a central role here because all of this requires more compute power.
LINDSAY: Finally, I want to get your thoughts on the software sector. It’s taken a hit this year. Do you see a rebound ahead?
NICHOLAS: Software is still going through a messy repricing, and I don’t think that’s going to resolve quickly.
Agentic AI is challenging traditional seat-based SaaS models by shifting value toward outcomes, while also creating margin pressure through self-cannibalization and higher inference costs.
What that means is companies can now produce software at near-zero marginal cost, which makes distribution and being a system of record more important.
There will still be winners, but the broad selloff reflects real uncertainty.
Right now, there’s likely easier money to be made elsewhere. I’d focus more on the physical economy — hardware, materials and infrastructure — because AI is becoming something that’s consumed on an ongoing basis rather than built once and sold repeatedly.
LINDSAY: We’ll leave it there. Nicholas Mersch, portfolio manager at Purpose Investments. Thanks for your time.
---
This BNN Bloomberg summary and transcript of the March 17, 2026 interview with Nicholas Mersch are published with the assistance of AI. Original research, interview questions and added context was created by BNN Bloomberg journalists. An editor also reviewed this material before it was published to ensure its accuracy and adherence with BNN Bloomberg editorial policies and standards.
