Nvidia has had a knack of besting analyst expectations and being punished for it. After another record-setting quarterly earnings report, which smashed the ‘whisper estimates’ laid out by Wall Street analysts in recent days, it seems to have broke the streak.
On Wednesday, the chipmaker upgraded the company’s current quarter revenue, promising investors that it would exceed $100 billion in revenue for the first time in its Q3 2027 earnings report. But not only that, it did something else it never did before: It laid out a forecast for its next fiscal year.
And just like that, a modest decline in $NVDA turned into a wave of praise, with the company’s stock soaring nearly 5% to fresh all-time highs.
What were the results?
Going into the ‘Super Bowl of Earnings’, most on Wall Street had no doubts that Nvidia was going to beat. The question was ultimately by how much. When the rooster came to roost, its Q2 2027 results didn’t just best the LSEG analyst numbers, but the most outlandish of ‘whisper’ figures heard on Wall Street:
- Revenue: $96.2 billion (vs. $92.165 billion expected)
- Earnings per share: $2.22 (vs. $2.10 expected)
Data center revenue, which is the largest portion of Nvidia’s revenue, more than doubled. That showing emboldened the company to issue a bold forecast for its current quarter, Q3 2027. It’s expecting $108 billion in revenue, plus or minus 2%, blowing past the centi-billion milestone.
But its forecast didn’t stop here. On its earnings call that followed, Nvidia EVP and CFO Colette Kress laid out the picture for its fiscal year 2028, which doesn’t even start until next year. The company’s confidence is what helped turn the stock, which was down about 0.5% after the strong report, around in the after hours session.
FY 2028 is looking bright
Nvidia management said that it expects FY 2028 revenue to grow 70% year-over-year; a historic growth rate for a company of this size, with margins this thick, and so much untapped potential.
Laying out the picture was one way that Nvidia hoped to stress its staying power amid worries that AI spending is settling. It also had more data in tow on the earnings call:
- Management said that the cloud-industry backlog now exceeds $2 trillion and it expects capital expenditures to continue to grow.
- The company is broadening its book of business beyond AI labs and hyperscalers, banking on a future where businesses and governments become large customers, briefly touching on a partnership with the U.S. Government to build data centers.)
- It’s devising new forms of financing in partnership with Wall Street institutions to help appeal to “non-hyperscalers”, with the long-term goal of making it half of its data center business.
However, despite the positive news and reaction that followed in Nvidia stock, the company did have one piece of negative news: It’s “capacity constrained.”
There’s always a catch
Like many semiconductor players, Nvidia bemoaned “extreme pricing conditions in memory” which are forcing it to rethink its growth forecast even as it prepares to roll out new products from its Vera Rubin product line.
It now expects lower gross margins in Q3 and Q4 of its 2027 fiscal year, with an improvement to follow in 2028, during which the company plans to raise prices.
Nvidia management said it’s engaging suppliers to attempt to blunt the impact of the memory shortage, but CEO Jensen Huang said that if not for the supply constraint, the fiscal year 2028 outlook would be “a lot higher.”
That might be something to take in stride, especially at the rate that Nvidia is rattling off new products. It’s new Vera CPU, Rubin GPU, and other equipment is turning Nvidia from a one-trick pony into a one-stop shop for high-performance, rack-scale AI systems.
That’s obviously something to look forward to for Nvidia investors, which have watched other semiconductor players take off while the world’s most valuable company has posted more tepid gains this year. While the iShares Semiconductor ETF is up over 64% this year, Nvidia stock is up just 13%.
Could competition and higher prices threaten Nvidia?
Nvidia’s strong forecast is enough to get investors excited, especially if it’s a baseline from which the company can improve upon in fiscal year 2028.
Per the chipmaker, it has made more revenue per gigawatt of product with every new generation:
- Hopper: $18 billion per gigawatt
- Grace Blackwell: $25 billion per gigawatt
- Vera Rubin: $40 billion per gigawatt
And asked about whether that trend could hold now and into the future, CEO Jensen Huang said that he thought that the trend would hold. He also sounded an unconcerned tone about the rate at which Nvidia hardware could be replaced by competitors or even chips made by many of the company’s current customers.
“They’re going to be utilizing Nvidia compute for a lot of their computing [for quite a long period of time.]”
However, even though Nvidia is the market leader, it has to strike the balance necessary to juice its revenue, but also ward off competition (or not create unnecessary competitors by pushing current clients to rush ‘custom silicon’ replacements).
Thankfully for them, this goliath has a commanding lead over GPU players like Advanced Micro Devices, is still years ahead of XPU builders like Broadcom, and even more inference-focused players like Cerebras.
And with all competitors facing the same component shortages, it really feels like the AI accelerator market isn’t Nvidia’s to win. At this point, it remains theirs to lose.
