NextFin News - Nvidia shares rose in extended trading on Tuesday after the chipmaker reported second-quarter revenue of $96.2 billion, up 106% from a year ago, and issued its first-ever year-ahead forecast of roughly 70% revenue growth for fiscal 2028. The move came as Federal Reserve Chair Kevin Warsh prepared for his debut Jackson Hole speech with the policy committee split and the 30-year Treasury yield at a 19-year high.
The Beat Was Never in Doubt. The Guidance Was.
Nvidia reported revenue of $96.22 billion for the quarter ended July 26, up 18% from the prior quarter and more than double the $46.74 billion it posted a year earlier. Data Center revenue reached $89.0 billion, up 117% year over year, while GAAP and non-GAAP gross margins both held at 75.0%. GAAP diluted earnings per share came in at $2.46, up from $2.39 in the prior quarter and $1.08 a year earlier; non-GAAP EPS was $2.22. GAAP net income of $59.7 billion represented a 126% year-over-year increase, and GAAP operating income climbed 124% to $63.7 billion.
The company guided third-quarter revenue to $108.0 billion, plus or minus 2%, with gross margins of 74.0% plus or minus 50 basis points. Then it did something it does not normally do: it looked a full year ahead. Nvidia's preliminary expectation for fiscal 2028 is roughly 70% revenue growth — a figure that, applied to this year's expected revenue, implies sales in the range of $690 billion to $700 billion, more than $100 billion above the roughly $570 billion analysts had been modeling. The magnitude of the guidance shift is the story within the story: the market had been underwriting a 44% to 45% growth rate for fiscal 2028, and the company handed investors a number 25 percentage points higher.
"AI has reached its inflection point. It's doing useful work. Its tokens are productive and profitable. Now, compute is revenue," said Jensen Huang, Nvidia's founder and chief executive. Asked about the ambitious outlook, he added: "Even though our demand is much greater than 70%, our supply allows us to confidently deliver 70%."
Shares rose about 0.8% to $213.70 in the first wave of after-hours trading following the regular-session close of $212.03, then extended the advance to roughly 4% to 5% as the fiscal 2028 projection and an expanded partnership with Amazon sank in. The move snapped a flat stretch for the stock since its previous earnings call and broke a pattern of negative day-after reactions in four of the past five quarters despite upside surprises. The company has now beaten estimates for 15 consecutive quarters, yet the stock had been treading water — a divergence that said more about investor psychology than about the company's execution.
The earnings release landed one day before Chair Kevin Warsh takes the stage at the Federal Reserve Bank of Kansas City's annual symposium in Jackson Hole, Wyoming, and the backdrop could hardly be more charged. The 30-year Treasury yield reached 5.34% last week, a 19-year high, and the Treasury Department stepped into the bond market on August 19 with an enlarged buyback program for longer-dated debt. Minutes from the Fed's late-July meeting showed three of the committee's 12 voting members dissented in favor of a rate increase. Markets are pricing a 64% probability that the Fed holds its 3.50%-3.75% target range at the September meeting, with a 36% chance of a 25-basis-point hike.
Two developments, one tension: the AI engine that powered the last two years of equity gains just delivered another blowout quarter and raised the bar for itself, while the interest-rate regime that prices those gains is showing fresh signs of strain. The question is which force wins.
Why the Bond Market Is the Real Story
The 30-year Treasury yield's climb to 5.34% — a 19-year high — is not just a number on a chart. It is the discount rate that prices every long-duration asset, from growth stocks to AI infrastructure projects financed over decades. When the long end of the curve moves, the present value of future cash flows moves with it. A 100-basis-point rise in the discount rate can compress the fair value of a long-duration growth asset by double-digit percentages even if the company's earnings forecasts never change.
The Treasury Department's response is telling. On August 19, it announced it would double buybacks of longer-dated bonds, funded by issuing more short-term debt. The stated goal was to support liquidity in the long end of the market. The subtext was harder to ignore: borrowing costs for the federal government at 30 years had reached levels not seen since before the 2008 financial crisis, and policymakers were acting to contain them. Yields initially fell on the news, then rebounded to finish the week higher — a sign that the market was not fully reassured by the intervention. The move also shifts the government's own funding mix toward shorter maturities, a trade-off that buys near-term relief at the cost of refinancing risk down the road.
This creates a transmission channel that runs straight through Nvidia's customers. The hyperscalers and AI labs buying billions in GPUs finance their data centers with a mix of equity and debt. A 30-year yield above 5% raises the hurdle rate for new projects, slows the pace of capital deployment, and forces a harder look at whether AI inference revenue can cover the cost of the compute. The earnings beat says demand is strong today. The bond market asks whether it will be strong enough tomorrow, at a higher cost of capital.
There is a second channel, through the Fed itself. With core PCE running at 3.7% annually in July — well above the Fed's 2% target — and three voting members dissenting in favor of a hike, the policy path is not the easy-money environment that powered the last AI rally. Warsh's Jackson Hole debut on Friday, August 28, at 8:00 AM ET is his first chance to reconcile a growth-friendly equity market with an inflation problem that refuses to fully cool. The symposium, running August 27 to 29 at Jackson Lake Lodge, draws roughly 120 officials and economists from more than 70 countries; this year's theme, "Financial Innovation: Implications for Payments and Policy," gives the chair room to speak in principles rather than commitments. Markets will parse every clause for a hint on whether the next move is up, down, or nowhere.
Cyclical or Structural? Both, and That Is the Problem
The central analytical question is whether the current environment represents a cyclical fluctuation or a structural shift. The answer, uncomfortably, is both — and they point in opposite directions. Getting this wrong flips the conclusion, so the two forces need to be separated rather than blended.
On the structural side, the AI infrastructure buildout has the hallmarks of a regime change rather than a cycle. Nvidia's own guidance — $108 billion next quarter, plus a 70% full-year growth target for fiscal 2028 that implies roughly $700 billion in revenue — describes a multi-year expansion that does not simply mean-revert. The customer base has broadened from a handful of frontier labs to what Huang called a golden age of new AI labs and startups, plus sovereign and enterprise demand coming online. The partnerships Nvidia announced with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR to mobilize over $500 billion of third-party capital for AI infrastructure are evidence that financing is being institutionalized, not improvised. When asset managers are building dedicated financing vehicles for AI data centers, the buildout has moved from a venture-backed sprint to a balance-sheet marathon.
On the cyclical side, the interest-rate environment is the classic mean-reverting force. The Fed funds rate sits at 3.50%-3.75%, and the market is split almost evenly on whether the next move is a hold or a hike. If inflation proves sticky, rates stay higher for longer and the discount-rate pressure on growth stocks persists. If the labor market weakens, the Fed cuts and the pressure eases. Either way, the rate cycle turns; it does not stay at an extreme. History is full of periods when a transformative technology arrived into a hostile rate environment — the buildout proceeded, but the equity multiples did not get to participate in full.
The two forces interact through one mechanism: duration. Nvidia's earnings are structural; the multiple investors will pay for those earnings is cyclical. A company can deliver 100% revenue growth and still see its stock go nowhere if the discount rate applied to those future cash flows rises faster than the cash flows themselves. That is the trap embedded in the current setup: the business can be right and the stock can still disappoint.
The Counter-Thesis: A Bubble Wearing a Growth Costume
The strongest case against the bullish read comes from the valuation camp, and it deserves a full hearing. The argument runs as follows: Nvidia's roughly $5 trillion market capitalization prices in years of flawless execution. The company has guided to $108 billion in revenue for next quarter, but that number assumes the hyperscalers keep spending at a pace that, annualized, would put AI infrastructure capex well above $1 trillion. At some point, the return on that spending has to materialize as actual AI software revenue, not just as revenue for the chip supplier. If the software monetization lags, the capex cycle ends in a digestion period, and the suppliers absorb the damage first.
The macro backdrop adds weight to this view. Christopher Waller, a Fed governor who dissented in favor of a rate cut at the July 2025 meeting, argued in his Jackson Hole speech that year that "proper risk management means the FOMC should be cutting the policy rate now." A year later, the committee's own minutes show three voting members pushing in the opposite direction — toward a hike. If the Fed is split on whether to cut or raise with core PCE at 3.7%, the macro backdrop is not the easy-money environment that powered the last AI rally. The reversal in the committee's center of gravity is itself evidence that the inflation problem has more legs than the equity market wants to admit.
The bear case also points to the stock's recent behavior: flat since the last earnings call, negative day-after reactions in four of the past five quarters despite beats. That is the profile of a stock that has moved from momentum-driven to event-driven, where each quarter becomes a binary test rather than a step in a trend. In that regime, the reward for being right shrinks and the penalty for a miss grows.
The rebuttal: the bear case conflates a stock-price cycle with a business cycle. Nvidia's revenue doubled year over year on 75% gross margins. The customers are not speculative startups alone; they include the largest technology companies in the world, plus sovereign wealth and infrastructure funds now committing capital through the newly announced financing platforms. The risk is real, but it is a risk to the stock's multiple, not to the underlying demand — at least for the next several quarters. The fiscal 2028 guidance, volunteered without precedent, is management's way of forcing the market to underwrite the full cycle rather than trade the quarter.
The falsifying signal: if Nvidia's Data Center revenue growth decelerates below 50% year over year for two consecutive quarters while gross margin compresses below 70%, the structural-growth thesis breaks. That would signal that hyperscaler spending is hitting a wall, not just pausing. Until then, the burden of proof sits with the bears.
What to Watch Next
The earnings report and the Jackson Hole setup are two sides of the same coin: one shows an economy where AI investment is accelerating, the other shows a financing environment where the cost of that investment is rising. Investors should treat them as a pair, not as separate stories.
In the short term, the stock reaction will hinge on whether the market reads the 70% fiscal 2028 revenue growth projection as a new floor or as a peak. A decisive break above $220 would signal that guidance has reset the bar; a failure to hold $212 would suggest the beat was priced in. The 52-week range of $164.07 to $236.54 puts the stock in the upper half of its yearly territory, but still well below its record — room to run if the multiple expands, and room to fall if it contracts.
In the medium term, the key variable is the 30-year Treasury yield. If it stabilizes below 5%, the discount-rate pressure on growth stocks eases and Nvidia's multiple can expand. If it pushes toward 5.5%, the pressure returns regardless of how well the company executes. The Treasury's buyback program is the first line of defense, but it is a liquidity tool, not a solution to the inflation problem driving the yield higher.
In the long term, the structural question is whether AI inference revenue catches up with infrastructure spending. If it does, the current capex cycle looks like the buildout phase of a new computing platform. If it does not, the capex cycle ends in a consolidation, and the suppliers absorb the damage first. The $500 billion in third-party financing Nvidia helped line up is the bridge between those two outcomes — it buys time for the software economics to catch up.
What to watch this week:
- Chair Kevin Warsh's Jackson Hole keynote on Friday, August 28, at 8:00 AM ET — his first at the symposium as Fed chair.
- Any shift in Fed funds futures pricing after the speech; the September meeting currently prices a 64% hold and 36% hike.
- Whether the 30-year yield holds below 5.34% after the Treasury's buyback intervention.
- Hyperscaler capex commentary in the weeks ahead, which will confirm or deny the sustainability of Nvidia's $89 billion Data Center quarter.
Nvidia just proved the AI boom is still accelerating. The bond market is about to test whether it can afford to keep going.
Explore more exclusive insights at nextfin.ai.
