Nvidia expects revenue to grow about 70% in fiscal 2028, a forecast that could force investors to rethink how long the AI spending boom can last.

For months, one question has haunted the artificial-intelligence trade.

How long can the spending continue?

Technology giants have poured hundreds of billions of dollars into data centers, AI accelerators and infrastructure. Investors have watched Nvidia transform extraordinary demand into extraordinary revenue growth. But eventually, every boom raises the same question: when does the growth slow?

Nvidia has just delivered a striking answer.

Not yet.

The company expects revenue to grow by approximately 70% in fiscal 2028, Chief Financial Officer Colette Kress said during the company's post-earnings conference call. The forecast is dramatically stronger than Wall Street's previous expectations, which implied roughly 44% growth for that fiscal year.

That may be the most important message from Nvidia's latest earnings.

The company is not merely predicting another strong quarter.

It is signaling that the current AI infrastructure cycle could remain extraordinarily powerful for several more years.

Nvidia's second quarter delivered numbers that would have been almost impossible for a semiconductor company to achieve a few years ago.

Revenue reached $96.2 billion, more than doubling from a year earlier.

Data Center revenue alone came in at $89 billion, up 117% year over year. Adjusted earnings reached $2.22 per share, beating the consensus estimate of roughly $2.09.

The company then forecast around $108 billion in revenue for the next quarter.

Those figures were already enough to challenge concerns that AI spending had begun to peak.

The fiscal 2028 outlook goes further.

A 70% growth expectation suggests Nvidia believes its customers are still in the early stages of building massive AI computing infrastructure.

It also suggests that the demand curve for advanced computing could remain steep even after the industry has already invested enormous sums.

That is what makes the forecast so important.

Investors were becoming increasingly aware of the sheer size of the AI investment cycle.

Alphabet, Amazon, Microsoft and Meta have been committing enormous amounts of capital to AI infrastructure.

AI startups are consuming ever-growing amounts of computing power.

Governments are building sovereign AI capacity.

And Nvidia has become the primary beneficiary of much of that spending.

At some point, simple arithmetic suggested the growth rate would have to slow.

The company itself now appears to be pushing that date further into the future.

According to Reuters, Nvidia's projected 70% revenue increase for the fiscal year ending January 2028 is significantly above Wall Street expectations.

That does not mean 70% growth is guaranteed.

Forecasts can change.

Supply can tighten.

Customer spending can slow.

Geopolitical restrictions can intervene.

But management's willingness to make such an aggressive forecast demonstrates an unusually high level of confidence.

One reason Nvidia remains optimistic is that the customer base is becoming broader.

In the early stages of the AI infrastructure boom, a handful of major cloud providers drove much of the spending.

Now Nvidia says demand is coming from a wider range of customers, including AI startups, frontier laboratories, enterprises, sovereign buyers and industrial users.

That broadening matters.

If AI spending depended entirely on a few hyperscalers, investors could reasonably worry that the largest customers might eventually reach a saturation point.

But if thousands of companies and governments are only beginning to deploy AI infrastructure, the potential market becomes much larger.

This is the logic behind Nvidia's bullish forecast.

The AI boom is evolving from a concentrated technology buildout into a broader computing platform.

Nvidia's numbers also highlight a deeper transformation in the global economy.

AI data centers are increasingly being treated like industrial infrastructure.

They require specialized chips.

They require huge quantities of electricity.

They require advanced cooling.

They need high-speed networking.

They require enormous buildings.

And they depend on global supply chains.

Once viewed primarily as technology facilities, they are beginning to resemble a new class of industrial assets.

That creates a powerful long-term demand cycle for Nvidia.

Every new generation of AI applications can create another reason for companies to build more computing capacity.

The resulting infrastructure spending can continue even if individual AI models become more efficient, because lower costs can actually encourage more usage.

Another important part of the bullish outlook is the company's upcoming Vera Rubin architecture.

Nvidia says Rubin is already in production and has purchase orders from major customers. The company has described the rollout as potentially its fastest product ramp yet.

This matters because Nvidia's competitive advantage is not based solely on what its current chips can do.

The company operates a rapid product cycle in which each new generation promises greater performance and efficiency.

That creates a recurring upgrade opportunity.

Customers that built AI clusters using earlier generations may eventually need newer systems to remain competitive.

Nvidia is therefore not simply selling a single product.

It is trying to create a continuous infrastructure upgrade cycle.

The AI buildout is also creating opportunities outside Nvidia's traditional GPU franchise.

Networking has become one of Nvidia's fastest-growing businesses because increasingly large AI clusters need extremely high-speed communication between processors.

That means the company's future revenue is becoming more diversified.

The more complicated AI systems become, the more infrastructure customers need around the central accelerator.

Nvidia's position allows it to sell parts of that broader architecture rather than depending solely on GPU demand.

This could become particularly valuable as hyperscalers develop their own custom chips.

A custom processor may compete with Nvidia's accelerator in certain applications.

But it still needs high-speed connections, systems architecture and networking.

Nvidia wants to remain central to those pieces.

The rise of custom AI chips is one of the strongest arguments against Nvidia's long-term dominance.

Google has developed its own AI accelerators.

Amazon and Microsoft are investing in custom silicon.

Meta is designing hardware around its specific workloads.

Other startups are pursuing specialized approaches.

On paper, this could reduce Nvidia's share of the accelerator market.

But Nvidia's strategy is becoming increasingly broader.

The company is positioning itself as an AI platform rather than simply a GPU supplier.

That means customers can potentially use custom processors while still relying on Nvidia's networking and interconnect technologies.

The competitive battle is therefore becoming more complicated than a simple comparison of chip performance.

Ironically, Nvidia's biggest immediate problem is not demand.

It is supply.

The company continues to warn about limitations involving memory and other critical components. Recent reporting indicates Nvidia expects gross margins to decline toward roughly 71%–72% in an upcoming quarter as memory costs rise faster than previously expected.

This creates an unusual dynamic.

The company is saying customers want more systems than it can currently deliver.

That is obviously bullish for demand.

But it can also create financial pressure.

Higher component prices mean Nvidia may have to spend more to satisfy the same amount of customer demand.

The company is therefore trading some margin strength for increased scale.

If revenue grows quickly enough, that can still be an excellent trade.

There is one major market Nvidia is not fully counting on.

China.

The company does not currently include expected Data Center revenue from China in its forward outlook because of U.S. export restrictions and uncertainty over future shipments.

That is significant for two reasons.

First, China represents a potentially large customer market.

Second, excluding China from forecasts arguably makes Nvidia's growth expectations more conservative.

If restrictions were eased, the company could gain an additional demand source.

If restrictions become more aggressive, Nvidia's opportunity in China could shrink further.

For now, management appears to be building its long-range outlook without depending on that business.

Nvidia's forecast does not prove there is no bubble in AI.

A company can experience huge real demand while investors simultaneously overpay for other companies in the same ecosystem.

But it does make the simplest “AI spending is about to collapse” argument much harder to defend.

Nvidia's management is effectively telling investors that customers are still planning enormous purchases well into 2028.

And the numbers backing that confidence are not hypothetical.

The company just reported nearly $100 billion in quarterly revenue.

Its Data Center division produced $89 billion.

It guided the next quarter toward $108 billion.

And it now expects roughly 70% annual revenue growth in fiscal 2028.

Those are extraordinary levels of commercial demand.

A longer AI growth runway changes the stock's valuation debate.

If Nvidia's earnings grow far faster for several more years than investors previously expected, today's valuation may look more reasonable in hindsight.

But there is an equally important counterargument.

The more successful Nvidia becomes, the harder it is to maintain gigantic percentage increases.

Nvidia is already one of the world's largest companies.

A 70% increase in revenue at this scale represents an enormous addition in absolute dollars.

That requires customers to keep spending at extraordinary levels.

It also means the company must continue executing flawlessly across manufacturing, software, networking, product launches and supply-chain management.

There is very little room for operational mistakes.

Nvidia's outlook may also provide a fresh catalyst for companies elsewhere in the AI ecosystem.

Strong forward demand supports the business cases of semiconductor manufacturers, memory suppliers, networking companies and data-center operators.

It supports utilities and power developers building capacity for AI infrastructure.

It supports optical networking and cooling businesses.

It supports cloud providers that can monetize AI workloads.

In other words, Nvidia's revenue forecast is increasingly becoming an economic indicator for the entire AI infrastructure economy.

A bullish Nvidia forecast can therefore move capital far beyond the semiconductor sector.

Perhaps the biggest detail is that the company did not frame 2028 as an endpoint.

The forecast implies growth remains extremely strong, but Nvidia has not suggested that AI demand is approaching a final saturation point.

Instead, management continues to describe an ecosystem expanding into new types of customers and workloads.

AI agents.

Physical AI.

Sovereign computing.

Enterprise deployment.

Advanced reasoning models.

All of these applications can require more computation.

The argument is essentially that AI is moving from experimentation into production.

And once AI becomes economically useful, the demand for compute becomes recurring.

Nvidia's confidence should not be mistaken for certainty.

The AI market could face a spending slowdown.

Customers could discover that some projects have weaker returns than expected.

New competitors could gain share.

Custom silicon could expand faster than Nvidia expects.

Export restrictions could tighten.

Supply bottlenecks could worsen.

Or AI models could become much more computationally efficient.

Any of those factors could reduce the company's growth rate.

The 70% figure is therefore a management forecast, not a guarantee.

But forecasts matter because they reveal how management sees the market.

And Nvidia clearly sees a market with several enormous years still ahead.

The company's latest outlook represents one of the strongest corporate endorsements yet of the idea that the AI infrastructure cycle still has room to run.

After years of explosive spending, the market was beginning to ask whether the boom was finally approaching a ceiling.

Nvidia has pushed that ceiling further away.

Its customers are still buying.

Its next-generation products are already attracting orders.

Its networking business is accelerating.

Its customer base is broadening.

And management expects revenue growth of approximately 70% in fiscal 2028—far above what Wall Street previously anticipated.

That does not eliminate the risks.

But it does change the debate.

The question is no longer simply whether the AI boom can survive another quarter.

It is whether the world is entering several more years of extraordinary spending on artificial intelligence.

Nvidia's answer is unmistakable.

The company believes the AI spending wave is not nearing the shore. It is still building.

And if that forecast proves correct, 2028 could look less like the end of the AI boom and more like another chapter in what becomes one of the largest technology investment cycles in history.

Source basis: Bloomberg/Yahoo Finance reporting, Reuters coverage and Nvidia's official fiscal Q2 2027 results and conference-call commentary.