Interview with Market Analyst: Oracle Forms Epic Head-and-Shoulders Bottom—Pullback Represents Final Buying Opportunity
TechFlow Selected TechFlow Selected
Interview with Market Analyst: Oracle Forms Epic Head-and-Shoulders Bottom—Pullback Represents Final Buying Opportunity
I believe technology stocks—especially leading blue-chip operators—still have a long way to go before their best days are behind them.
Navigating Web3 tides with focused insightsCompiled & Translated by TechFlow
Guest: Thomas Hughes, Market Analyst
Host: Jessica Mitacek
Podcast Source: MarketBeat
Original Title: Wall Street Just Gave a Dire Warning. (Most Aren't Ready)
Air Date: June 25, 2026
Key Takeaways
Analyst Thomas Hughes argues that although hyperscale tech companies and AI infrastructure firms have taken on approximately $750 billion in new debt over the past 18 months, their accumulated backlog orders during the same period have reached roughly $2.1 trillion—nearly three times the amount of new debt. This suggests that the current market panic over “burning cash” is better understood as upfront investment to fulfill already-signed demand. According to Hughes, meaningful revenue recognition is unlikely until late 2027 or early 2028, when the first generation of AI data centers begins coming online. Therefore, volatility will persist over the next few quarters—but the long-term upward thesis for large-cap tech stocks and core infrastructure companies remains intact.
Key Insights Summary
Market Panic and Misinterpretation of “Cash Burning”
- “The market’s hesitation stems primarily from concerns that data center expansions and massive capital expenditures will strain these companies’ debt levels, balance sheets, and cash flows.”
- “Over the past 18 months, hyperscale cloud providers and AI infrastructure companies have added roughly $750 billion in new debt to advance data center construction. Investors naturally react with concern as free cash flow rapidly declines from elevated levels—creating headwinds for both market sentiment and stock performance.”
- “AI development carries risk—but it’s an execution risk. It’s more likely to manifest as temporary price corrections rather than a full-blown collapse in share prices. Over the longer term, these companies are still likely to continue trending upward.”
Backlog Orders Are the More Critical Leading Indicator
- “What most clearly distinguishes this not as a ‘story-driven emerging-tech narrative,’ but as a systemic technology upgrade, is the backlog. This round of spending, debt expansion, and data center construction isn’t a bet on hypothetical future business—it’s to deliver already-contracted work.”
- “This means the long-term logic is that today’s debt-financed capital will be converted into future revenue roughly three times larger than the debt itself.”
- “Over the coming years, we should see robust growth, accompanied by gradually declining debt levels, restored free cash flow, and accelerated buybacks.”
Revenue Recognition Timing Gap and Market Volatility
- “Overall, I think these companies have explained things quite well—because contract recognition cycles are inherently back-loaded, with much of the revenue dependent on compute capacity yet to be released. Keep in mind that existing compute capacity is already nearly fully utilized—that’s why pricing is rising.”
- “Although the backlog currently stands at $2.1 trillion, a large portion won’t be recognized as revenue immediately—it may take until late 2027 or even early 2028, after these new data centers go live, before revenue conversion becomes meaningfully visible.”
- “Volatility will persist over the next several quarters. The AI bubble is likely to expand further, creating substantial upside opportunities between earnings seasons; meanwhile, the market will continuously find new reasons to worry, remain cautious, or simply take profits—leading to frequent pullbacks.”
- “In the current environment, weakening prices look more like buying opportunities.”
Oracle’s Central Role
- “For me, Oracle represents one of the highest-quality AI narratives.”
- “It’s essentially a hyperscaler serving other hyperscalers—providing high-capacity, high-performance compute services to major cloud providers, Meta, various AI labs, and numerous enterprise customers. Its core database business—already deeply embedded across the entire cloud ecosystem—is integrated into major cloud providers’ systems and networks.”
- “Judging by its current price action, I believe Oracle has already formed a fairly clear bottom. The market has severely oversold the stock, overreacted to risks, and the technical structure is now highly conducive to a rebound—the only missing ingredient is a true catalyst to flip sentiment.”
Chips and the AI Super-Cycle
- “Micron’s production capacity is already booked through the end of next year—and this earnings report could push that timeline further into 2028. That continues to support price expectations and directly boosts its current business performance. What supports Micron isn’t just volume—it’s also the pricing power it commands.”
- “The semiconductor industry is currently in a massive super-cycle—initially driven by inventory normalization, then further amplified by AI. As long as supply-side capacity cannot catch up with demand, this cycle will persist for many years.”
- “AI investment creates new capacity and new technologies—and those outputs, in turn, boost efficiency, enabling the industry to invest more capital into the next round of upgrades. In other words, each investment cycle drives technological advancement, which then fuels the next, larger investment cycle—repeating endlessly. At least for now, AI’s flywheel effect appears to have no definitive endpoint.”
Long-Term Outlook
- “This cycle is far from over. I believe tech stocks—especially leading blue-chip operators—still have a long runway ahead. AI will drive continuous technological change, but these companies are already positioned advantageously to ride that wave forward.”
- “We haven’t even completed building the first generation of AI data centers yet. At current pace, the first-gen facilities won’t begin going live until next year.”
- “Once we truly enter the application phase—when enterprises begin monetizing these technologies—we’ll start seeing revenue and profit materialize.”
- “They possess sufficient capital, scale, and execution capability to make this happen.”
Opening: Why Is the Market Suddenly Nervous?
Host Jessica Mitacek: In just two days, mega-cap tech stocks shed over $2.5 trillion in market value. Headlines almost universally focus on how the scale of AI-related spending by these tech giants has become impossible to ignore—but is this sell-off an overreaction? Or does a deeper, more important story lie beneath the surface? Joining us today is MarketBeat analyst Thomas Hughes, who’ll break down what’s really happening with hyperscale cloud providers—and highlight three companies worth watching closely. Thomas, let’s start with the market’s current reaction.
Thomas Hughes:
The market peaked just a few weeks ago and is still working to regain upward momentum. The core reason for hesitation is concern that data center expansions and massive capex will strain these companies’ debt, balance sheets, and cash flows.
Host Jessica Mitacek: This pullback is indeed very visible on last week’s heat map—almost entirely concentrated in the mega-cap tech sector. For example, Oracle fell over 14% in the past seven trading days, Google dropped nearly 4%, and the broader mega-tech zone turned almost uniformly red. The issue is that these companies had just experienced a sharp rally—so everyone wants to know: what actually triggered this correction? The most common explanation lately has been “cash burning,” especially given the massive capex figures disclosed by hyperscale cloud providers in their Q2 earnings. How do you view this issue?
Thomas Hughes:
At a high level, hyperscale cloud providers and AI infrastructure companies have taken on roughly $750 billion in new debt over the past 18 months to advance data center construction, including some dilutive financing actions. They’re drawing down free cash flow while simultaneously adding debt—which then erodes free cash flow further, as cash must be used to service debt.
These companies previously engaged in massive share buybacks, but funds formerly earmarked for buybacks are now being redirected toward debt servicing—explaining the sharp compression in cash flow. Investors naturally grow nervous when they see cash flow falling rapidly from peak levels—creating headwinds for both market sentiment and stock performance.
Host Jessica Mitacek: I saw a chart from Yahoo Finance today that illustrates exactly this shift. The issue isn’t just how much was spent last quarter—it’s that these companies have been pouring enormous capital into AI expansion for multiple consecutive quarters. The real question is whether these investments will ever pay off. Is this an overly aggressive, high-risk bet—or will it ultimately yield substantial returns? Could this cash burn pose a material risk to hyperscale cloud providers?
Thomas Hughes:
Risk exists—because the sums involved are genuinely enormous. But ultimately, this is more of an execution risk: they must build these systems, bring them online, and successfully monetize them.
Still, compared to early-stage emerging tech companies, this risk profile is considerably lower—because this isn’t unproven technology requiring a new business model from scratch. It’s an evolution of the existing tech stack. Nearly all top-tier companies are already participating in—and collaboratively driving—this transformation. From this perspective, risk is significantly mitigated.
The technological direction is crystal clear. The question isn’t whether this path is viable—but whether it can be built, launched, and monetized. So my view is: Risk exists—but it’s an execution risk. It’s more likely to show up as temporary price adjustments—not a total collapse in stock prices. Over the long term, these companies are still likely to trend higher.
Host Jessica Mitacek: Markets have grown accustomed to these companies consistently outperforming—and to their ongoing buybacks, expansions, and above-market growth. Now they’ve entered a new investment cycle, pouring massive capital into AI infrastructure—a shift that understandably unsettles investors. One point I’d emphasize: many retail investors perceive stocks like Google or Nvidia as “boring”—but the reality is that virtually every investor holds these assets indirectly. Even if you don’t own them directly, if you hold a 401(k) or any broad ETF, you’re almost certainly exposed. So everyone wants to know: what does this mean for their portfolio?
Thomas Hughes:
That’s absolutely true—nearly everyone holds these stocks. They’re among the largest sources of revenue and cash flow, with exceptionally clear growth trajectories. Even investors without direct holdings likely own them via ETFs or mutual funds. So many people assume they have no exposure—yet they probably do.
Beyond “Cash Burning”: Backlog Orders Matter More
Host Jessica Mitacek: Do you see today’s “cash burning” as a genuine risk investors should worry about—or does it actually point to a deeper, longer-term logic?
Thomas Hughes:
I believe it reflects a longer-term logic. The strongest evidence this isn’t a ‘story-driven emerging-tech narrative,’ but a systemic tech upgrade, lies in the backlog.
This round of spending, debt expansion, and data center construction isn’t a speculative bet on future business—it’s to fulfill already-contracted work. As debt balloons, backlog orders are growing even faster. Across the chain—from hyperscale cloud providers to AI infrastructure firms, chipmakers to data center hardware suppliers to compute leasing platforms—the total backlog now stands at roughly $2.1 trillion—about three times the new debt taken on.
This implies that the long-term thesis is that today’s debt-financed capital will convert into future revenue approximately three times larger than the debt itself. We should see robust growth over the coming years—while debt levels gradually decline, free cash flow recovers, and buybacks accelerate once again.
Also note: these backlog contracts are typically multi-year agreements. We estimate this $2.1 trillion backlog is likely sufficient to sustain growth for the next three to five years. Once this batch is digested, new contracts will follow. By then, data centers will be built—and no new debt will be needed, only higher-margin contracts rolling forward. So from an AI long-term outlook perspective, I see tremendous strength.
Host Jessica Mitacek: Here’s a critical point: your mention of backlog orders doesn’t apply only to companies physically building data centers or selling equipment and components, right? We know Google, Amazon, and Meta are investing heavily—essentially transferring money to infrastructure providers (“the shovel sellers”). So this backlog isn’t just in the infrastructure chain—it’s also present within the hyperscalers themselves?
Thomas Hughes:
Absolutely—and it permeates the entire value chain. It’s not just reflected in AI GPUs, nor just in GPU assembly components and data center infrastructure—it’s embodied in actual compute capacity itself.
Major hyperscalers—and many large software service providers—have already begun signing, and continue to sign, future compute capacity contracts. What truly underpins the entire growth story is their demand for compute capacity—and that demand, in turn, pulls forward infrastructure investment.
So where we stand today is fundamentally an infrastructure build-out phase. The long-term investment thesis shifts from “building compute” to “using compute.” Once compute is widely deployed, backlog orders will materially convert into revenue and profit.
When Will Backlog Convert to Revenue?
Host Jessica Mitacek: There’s no doubt that demand for AI compute is real—whether from enterprises, corporations, or everyday investors, AI usage is becoming increasingly widespread. With demand clearly growing, hyperscaler expansions are easy to understand. Google is a classic example: it’s investing massively in AI—but investors most want to know: how will this translate into Google’s own revenue? What’s its monetization path?
Thomas Hughes:
Google is an excellent case study. As charts show, it’s not just Google—entire MAG 7, including Nvidia, has seen solid pullbacks recently. But to me, this looks like a perfectly normal market mechanism between earnings seasons. After the last earnings cycle, markets rose too quickly and too sharply—naturally setting the stage for a correction.
So I see the market as essentially recharging for the next leg up. Within the next two to four weeks, we’ll enter another earnings window—and reports from these companies are likely to remain strong, continuing to validate several key trends: sustained capex, ongoing data center capacity builds, and persistent end-market demand for compute.
Host Jessica Mitacek: Lately I’ve heard a recurring theme: while these earnings reports disclose AI expansion spending—and thus support the “AI investment story”—the real question is whether hyperscalers adequately explain on their earnings calls how today’s investments will translate into their own future revenue.
Thomas Hughes:
I think, overall, they’ve done a reasonably good job explaining—because contract recognition cycles are inherently back-loaded, with much dependent on compute capacity yet to be released. You need to recognize that current compute capacity is already near full utilization—that’s why pricing is rising.
So many contracts lock in future compute and future technologies. A significant portion of these data centers will be built on next-generation platforms—like newly launched architectures, or AMD’s MI450 series, which hasn’t yet rolled out broadly. In other words, although the backlog currently stands at $2.1 trillion, much of it won’t be recognized as revenue immediately—it may take until late 2027 or even early 2028, after these new data centers go live, before conversion into tangible revenue becomes meaningfully visible.
Host Jessica Mitacek: This timing is crucial for investors. If revenue realization is still one to two years away, will the market continue oscillating in the interim? Much of the mega-cap tech market cap erosion may stem precisely from investors’ unwillingness to bear risk during this waiting period. Do you expect concerns around high spending and slow revenue conversion to persist for the next 1–2 years?
Thomas Hughes:
I believe volatility will continue over the next several quarters. The AI bubble is likely to expand further—generating sizable upside opportunities between earnings seasons; meanwhile, the market will keep finding new reasons to worry, stay cautious, or simply book profits—leading to frequent pullbacks.
But in the current environment, I view price weakness more as a buying opportunity. Because catalysts for future stock appreciation aren’t limited to today’s earnings and guidance—they include the eventual conversion of this backlog into cash flow and profit. Once that process begins, the stock will gain longer-term support.
Stock #1: Oracle
Host Jessica Mitacek: We just discussed Google. Next, let’s examine several companies you’re watching closely—one being Oracle. You recently published an article on its recent pullback, and it’s a textbook example of a high-investment expansion company.
Thomas Hughes:
Oracle is arguably one of the most representative names in this story. It’s aggressively using debt leverage—but its backlog growth ranks among the strongest I track.
For me, Oracle represents one of the highest-quality AI narratives. It’s not merely transforming from a traditional tech firm into a modern cloud and AI company—it’s evolving from a relatively niche participant into a core, blue-chip player at the heart of the AI and data center ecosystem. Put simply, Oracle is a hyperscaler serving other hyperscalers.
It provides high-capacity, high-performance compute services to other major cloud providers, Meta, various AI labs, and numerous enterprise customers. And that’s just its cloud business. Its core database business—already deeply embedded across the entire cloud ecosystem—is integrated into major cloud providers’ systems and networks. It’s also one of the world’s most widely adopted and accessible databases—making it a foundational element across the AI infrastructure and industry value chain.
Host Jessica Mitacek: Compared to Google or other MAG 7 companies, does Oracle carry higher risk? After all, its starting balance sheet isn’t as strong as those mega-giants’.
Thomas Hughes:
I don’t view its risk as higher. In fact, even before this expansion cycle began, its balance sheet was already quite healthy—and it had been actively repurchasing shares. Some financial metrics do raise concerns, but those are partially offset by strong cash flow and aggressive buybacks.
I think the biggest near-term concern is debt expansion—which will likely weigh on results for roughly the next 12–18 months, possibly longer. But as I’ve emphasized earlier, starting next year, Oracle will progressively convert backlog into revenue—and that conversion will accelerate over subsequent quarters, quickly absorbing the debt burden. So, I believe the downward pressure currently weighing on the stock will reverse over the next few years—ultimately showing up as upward price movement.
Host Jessica Mitacek: Looking purely at the chart, Oracle has fallen roughly 15% recently—and this volatility has persisted throughout the year. It hit an all-time high last September and has since traded in a high-volatility range. So investors’ key question is: Is this the start of a deeper downtrend—or has it reached a bottom worth accumulating?
Thomas Hughes:
In my view, this is not the beginning of a downtrend—but a clear buying opportunity. Oracle’s rally last year was fundamentally driven by rapid backlog expansion and re-rating. The subsequent pullback stemmed largely from market concerns over debt.
But judging by current price action, I believe Oracle has already formed a fairly clear bottom. Although it’s still pulling back over the past few weeks, what I’m seeing is a classic head-and-shoulders-bottom pattern. The market has severely oversold the stock, overreacted to risks, and the technical structure is now highly conducive to a rebound—the only missing ingredient is a true catalyst to flip sentiment.
Oracle reports mid-cycle, having just released earnings—so its next report is still several weeks away and later than other hyperscalers. But in my view, other giants’ earnings will serve as Oracle’s catalyst—since part of their disclosed capex flows directly to Oracle to fund its own hyperscale expansion.
Stock #2: Micron and the Chip Chain
Host Jessica Mitacek: While Oracle won’t report new earnings soon, the market’s most anticipated report today is Micron’s. We’re recording this episode before Micron closes—so the report isn’t out yet. Still, everyone’s eager to see how much of this massive spending flows to chip and memory companies like Micron. How does this report connect to Oracle—or to the broader AI investment story?
Thomas Hughes:
Massive capital is flowing into the chip sector—not just GPUs, but also memory chips like Micron’s. Micron’s stock just hit a new high this week—though it pulled back slightly ahead of earnings. To me, the chart remains extremely strong.
The market currently expects Micron to demonstrate exceptionally high demand—and potentially push out its timeline for meeting capacity requirements even further. Under current conditions, Micron’s capacity is already booked through year-end next year—and this earnings report could extend that timeline into 2028. That continues to support price expectations and directly boosts its current business performance. What supports Micron isn’t just volume—it’s also the pricing power it commands.
Zooming out to the broader semiconductor industry, AI is driving demand across the board. Virtually every chip type imaginable plays a role in data center construction. Longer term, AI applications will extend into IoT and diverse AI use cases—because GPUs are merely the computational “brain,” while vast numbers of other chips are needed to interconnect GPUs, link servers, tie together data centers, manage power control devices, drive actuators, and support the myriad servers powering global IoT infrastructure.
Ultimately, AI will evolve into “AI in the physical world” and IoT—using this computational brain to remotely execute tasks globally.
Stock #3: How Long Can the AI Super-Cycle Last?
Host Jessica Mitacek: That explains why more and more investors are adopting AI—once they start using it, they realize it genuinely creates value. But the market’s next big question is: how long can this super-growth cycle last? Especially in semiconductors—will this high-visibility phase peak soon?
Thomas Hughes:
The semiconductor industry is in a massive super-cycle—initially driven by inventory normalization, then further amplified by AI. As long as supply-side capacity cannot meet demand, this cycle will persist for many years.
Focusing solely on memory chips, it will likely take at least another year for the industry to significantly ramp capacity and satisfy current demand. That means chipmakers’ good times will likely continue for the next year.
But looking further ahead, AI is generating a self-reinforcing flywheel effect. AI investment creates new capacity and new technologies—and those outputs, in turn, boost efficiency, enabling the industry to invest more capital into the next round of upgrades. In other words, each investment cycle drives technological advancement, which then fuels the next, larger investment cycle—repeating endlessly.
At least for now, AI’s flywheel effect appears to have no definitive endpoint. The industry remains in an extremely early stage—we haven’t even completed building the first generation of AI data centers yet. At current pace, first-gen facilities won’t begin going live until next year.
Host Jessica Mitacek: That brings us back to another key market concern: everyone knows money is flowing into first-gen AI data centers—but once built, won’t they already be obsolete? Will they be fully utilized as hyperscalers anticipate—or will the market pivot to newer generations before construction finishes? The risk of rapid infrastructure depreciation is another major source of investor anxiety. How would you address that concern?
Thomas Hughes:
In tech, once a data center is built, better products will inevitably emerge—that’s certain. But new technologies take time to arrive. You need to recognize that this AI data center boom has been brewing for years—yet hasn’t yet delivered large-scale outcomes.
We’ve seen early signs—like strong demand for infrastructure products from certain companies—but those data centers themselves aren’t fully built yet. So large-scale AI adoption still lies ahead.
Once we truly enter the application phase—when enterprises begin monetizing these technologies—we’ll start seeing revenue and profit materialize; and during that process, capital will gradually shift toward next-generation products. At that point, a new super-cycle may well form.
Host Jessica Mitacek: So your conclusion remains: bullish on MAG 7, bullish on hyperscalers, and bullish on the entire AI value chain’s long-term prospects. You don’t believe this story is nearing its end—or that risks have grown large enough to undermine these companies’ futures.
Thomas Hughes:
This cycle is far from over. I believe tech stocks—especially leading blue-chip operators—still have a long runway ahead. AI will drive continuous technological change—but these companies are already positioned advantageously to ride that wave forward. They possess sufficient capital, scale, and execution capability to make this happen.
Join TechFlow official community to stay tuned
Telegram:https://t.me/TechFlowDaily
X (Twitter):https://x.com/TechFlowPost
X (Twitter) EN:https://x.com/BlockFlow_News
深潮TechFlow
