https://www.youtube.com/watch?v=E7eoGHpg9Mk
Thomas Hughes, an analyst at MarketBeat, takes host Bridget Bennett through the five stocks he is watching in September: Nvidia, MongoDB, Snowflake, Okta and AeroVironment. He says the second-quarter earnings season answered the fear that AI would eat the software companies, and that Nvidia's report showed the market still does not understand what AI is worth. Bennett presses him on whether four of the five have already run too far, on why Nvidia's own return has lagged the rest of the AI trade, and on whether the fifth is a falling knife.
👤 Guest: Thomas Hughes, analyst at MarketBeat
🎙️ Host: Bridget Bennett, MarketBeat
📰 Published: 30 August 2026
🔴 YouTube | 🟢 Spotify | 🟣 Apple Podcasts | 🔗 Show notes | ⏱️ 23 min | ✅ Time saved: 8 min
Key Takeaways
September weakness is the entry point, not the risk
"So I'm looking at September as being a time to get into some good positions, some solid positions that will carry profits into the year end."
The second-quarter season answered the SaaS apocalypse question
Software companies are better at deploying AI than AI is at disrupting them
Nvidia's report was evidence the market is mispricing AI, not just a beat
"the market is not pricing in AI correctly"
A 70% growth rate is the wrong thing to be disappointed by
"this is 70% after it's grown 100% for 5 years in a row. That's insane."
The post-earnings rebound was the market signaling a buy rather than waiting for a correction
Nvidia's flat year was consolidation, and the valuation is still deep
The host put it to him that the stock is up about 22% over the year against 100% and 200% elsewhere in AI
MongoDB is the one name he would not buy at the start of the month
The trigger is a break above 470 that then holds as support
Snowflake is in a discovery phase with no visible ceiling on the price targets
Institutional buying set a record last quarter, which he says limits the downside
Okta is being paid for agentic traffic, because every bot is an identity to manage
"the rise of agentic traffic is just driving an exponential increase in need for their product"
He buys where the money is flowing, not where the value is hiding
"it could be Boot Barn and shoes for all I care if that's what the market was interested in"
AeroVironment is the one broken chart on the list, and he says it has stopped going down
The case is an enormous backlog and follow-on defense orders
September Sells Off, and That Is What Sets Up the Fourth Quarter
Bennett opened by asking whether the market would deliver its usual September slowdown. Hughes said the seasonal weakness is real but shallow, and that it is the reason to be buying rather than a reason to wait.
"historically speaking, September can be volatile" — investors come back from the summer and reposition for the fall and the winter, which produces selling, sideways trading and price corrections
He treats that as the setup for the rest of the year: "So I'm looking at September as being a time to get into some good positions, some solid positions that will carry profits into the year end."
Bennett noted the two of them had already discussed his year-end outlook in a video earlier in the week, and framed this conversation as the September-specific version of it
The Second-Quarter Season Buried the SaaS Apocalypse Fear
The last of the season's big reports had landed on the Wednesday. Hughes said the quarter was reassuring on the exact fears that had been holding software stocks down.
"There were a lot of fears about the SaaS apocalypse, AI disruption, the sustainability of growth." What the results showed instead, he said, is that "the AI bubble is still growing" and in some places accelerating
"It will be bigger than we have anticipated it to be and that it will continue to produce revenue and profits for companies for the foreseeable future."
Bennett agreed the market was popping on the end of earnings season, led by tech and AI, and made the case for the segment: there are winners and losers inside AI, and that is where an investor can get an edge
Nvidia: The Report Was Evidence the Market Has AI Fundamentally Wrong
The first pick is Nvidia, the source of the report that closed the season. Hughes's reaction was that the company has now done this so many times the superlatives have worn out — "I don't know what there is to say about it except wow again after four years of just mind-blowing wow."
He said the company outperformed by a wide margin and gave its first look at the coming fiscal year, which he described as "at least 2,000 basis points better than expected." The point he drew from it is not about the quarter.
The claim the whole segment rests on: "the market is not pricing in AI correctly" "It is a fundamental misunderstanding of what AI means to the global tech ecosystem and Nvidia is at the core of it."
The report also carried ecosystem updates showing the company reaching further into each layer of the AI stack
On the pick itself: he called the stock an execution story rather than a story that still needs proving
Nvidia and AMD Are Both Working, and That Is the Point About Market Size
Bennett noted that AMD is a regular on his monthly list and that he is known as an AMD bull, then asked whether there is room for both names to do well.
Hughes said there is, and that the metrics are what point to it: "AMD's strengths have been unfolding alongside Nvidia's strength."
The two happening at once is his evidence for the size of the opportunity: "To me, that's a sign of how big the market really is."
He tied it back to the same argument — that the market misunderstands what AI means for the ecosystem and for business over the long term
Demand Is Not a One-Time Thing, and 70% Growth Is the Wrong Thing to Be Disappointed By
Asked what the report showed about demand, Hughes said the results were far ahead of expectations after years of triple-digit growth, and that the more important part is the durability rather than the size.
"It's not just a one-time thing. It's been going on for years. It's going to continue into the next year."
The line he uses to deflate the deceleration argument: "So 70% growth is a down tick from 100% growth, but this is 70% after it's grown 100% for 5 years in a row. That's insane. It's just too much."
The guidance and the commentary hinted the company could do better still, because "there were some capacity constraints mentioned" Those constraints are what is driving the industry-wide push to add capacity, which he said feeds the earnings and revenue outlook for the whole ecosystem
The Market Bought the Report Instead of Selling It
Bennett asked whether an investor should buy immediately after a report that strong, given the stock and the wider AI complex jumped on the Thursday. Hughes said the shape of the reaction is itself the signal, because Nvidia normally sells off on good news.
"The market reaction after the release was to sell off initially but only marginally so went down to the near-term support target and then rebounded strongly." The stock was up 5% or 6% by the time they spoke
"To me, the market's signaling a buy." — the move told him buyers had been waiting for it, and "They're not waiting around for the market to correct anymore. They're just buying today."
The technicals, the valuation and the trend in analyst price targets all point the same way, to triple-digit gains "we are on track for more triple-digit gains with as much as 200 to 300% more to come over time based on the way the earnings outlook is unfolding" He was explicit that the path is not smooth: the price may move quickly or slowly, and there will be corrections along the way
A Flat Year in Nvidia Was Consolidation, Not the End of the Story
Bennett put the most common objection from viewers to him: that Nvidia is no longer a growth stock. "You look at their returns and they're only up about 22% in the last year for investors versus plenty of other AI stocks that are seeing 100, 200% returns", she said, and that is the kind of return retail investors are chasing.
Hughes said "the past year or so has been consolidation while the market figures out what's going on"
The report told him the boom, the upside potential and the valuation case are all intact — "the valuation is still very deep"
His answer to the question as asked: yes, September is a good time to get into Nvidia
MongoDB Is AI-Critical, but He Would Not Chase It Here
The second pick is MongoDB, a name he said comes up on the list less often than the chip stocks. The business is database services; the story is the Atlas platform, which he said helps clients deploy and operate the platform, build agentic workflows into their development process and automate back-end operations.
The results show traction, momentum, deeper penetration and more client usage, and he said the company is "just really emerging as an AI critical name" with that reflected in the share price
Why it matters twice over: the database platform is critical to developing AI now, and critical again later for inference, as the companies training models today give way to the companies deploying and using inference tomorrow
This is the one pick he would not buy at the start of the month: "right now you're not really going to want to chase MongoDB. You're going to want to wait for price pullbacks." The price is sitting very near a resistance point The trigger he is waiting for: "When we move above the 470 level and confirm it as support, that will show me the market is inflecting" and following through on the signals given since April
Bennett pointed out the stock is up about 45% in three months after bottoming during the software downturn, and asked whether a market-wide September slowdown could pull it back He would not rule it out, but "it looks like this market is consolidating at the top of its range and it's preparing to move higher. It's just a matter of time."
Asked about the analysts, he said sentiment has bottomed after the downgrades that came with the software sell-off: "Analysts are back into a more bullish posture and they're helping to provide support for the market."
Snowflake Sells Compliance, and the Results Ended the Software Argument
The third pick is Snowflake, another data business, and the differentiator he named is regulatory. It "enables businesses and highly governed industries, highly regulated industries to remain compliant while using AI behind these secure firewalls", and gives staff across an enterprise access to the same data without copying files or sending them to different places.
Bennett noted he has been bullish on the name for about two years through a great deal of volatility, and that the stock is now up over 80% in three months. Hughes said the market is still working out what the company is worth, and that the last two years explain why: a chief executive change, fears of slowing growth, and the same SaaS apocalypse fear that hit the rest of the group.
The finding that settled it, in his words: "these companies are better able to deploy AI than AI is able to disrupt their businesses"
Growth reaccelerated, results and guidance beat, and profitability is there — which he said is what has analysts raising price targets and leading the market higher
"So right now what we are in is a discovery phase." "the upside potential is virtually unlimited until the analysts start to see a top to the growth and a top to the price targets"
Institutional Buying Set a Record Last Quarter, Which He Says Caps the Downside
Bennett turned to ownership, noting "only about 65% institutional ownership which isn't a huge percentage for such a large company" and asking what the recent quarter of institutional buying signals.
Hughes said 65% is decent for a company that size, because the stock is widely held and carries a lot of retail interest
Inside that, institutions have been aggressively bullish, buying on a quarterly basis over the trailing 12 months
The detail he says matters: across the tech companies he covers, "their activity really spiked in Q3, set a noticeable record high underpinning the stock price movement" "To me, this limits downside risk because we can expect them to buy any selloffs and help amplify any upside movements."
Asked the same buy-now-or-watch-it question he had answered the other way on MongoDB, he said "Snowflake looks like a pretty good buy today." The stock is near the top of its range, but the price action is bullish "It's showing support at the 30-day moving average. What I see in play is a bull flag continuation pattern." Assuming it breaks to the upside: "We'll probably see 400 by early next year, if not higher."
Okta Is Getting Paid for Agentic Traffic, Because Every Bot Is an Identity
The fourth pick is Okta, which he described as a long-standing name on MarketBeat's radar that did well immediately after the pandemic and then struggled. "They've been wallowing near long-term lows for years since then, but have regained traction now, and that's due to AI."
The mechanism is specific. AI is creating agentic traffic — bots moving through corporate systems, some known and some not, trying to reach files and act on a business's behalf. Okta's business is helping companies identify and track those actors and control what they can reach.
"these bots equate to IDs and to employees", which for Okta is simply more business
"for Okta, the rise of agentic traffic is just driving an exponential increase in need for their product and that's being reflected in their results"
The growth still has years to run, on his reading, because the buildout has not happened yet: the data centers that will drive inference are not built, so this is an early-phase story
The stock was "up more than 22% following the release", confirming support at its prior resistance point, setting a new high and breaking out of its consolidation range "To me, this market's on track for 200, probably to retest its all-time highs near 260." He thinks this one could get there faster than any other name on the list "The SaaS apocalypse fears really had the market depressed and now it's kind of in a catch-up mode."
Why Okta's Analysts Took So Long to Catch Up
Bennett said the analyst price targets are the part that has not caught up, having been cut hard on the apocalypse fear, and made the broader point that targets are often wrong when AI enters a story — stocks fly above consensus and the analysts follow the price action afterwards. She asked why the recovery took as long as it did.
"Well, it just took time for the agents to get active." He said "the first part of the story was mostly OpenAI and chat bots", and that agentic traffic has only risen over the last year
That release triggered a lot of upward revisions, which he said put a bottom in the sentiment trend and reinvigorated it
Where the target range now sits: the revisions are "leading the stock back to the high end of the target range, which is now pegged at 200" — the same level he named as the next critical target
He Buys Where the Money Is Flowing, Not Where the Value Is Hiding
Before the last pick, Bennett stopped to ask about method: four of the five names have had strong 30- to 90-day runs and strong earnings reports, so why put stocks that have already moved on a buy list rather than waiting for a pullback.
"when it comes to the market and investing, you want to go to where the market's interested and where the market's flowing". The companies producing the strength are the ones the market wants to own
He is emphatic that the sector is incidental: "I mean, to me, the story is in Nvidia and AI, but it could be Boot Barn and shoes for all I care if that's what the market was interested in."
"So, what I'm following is the story in the news and the cash flow. And the cash flow is into AI." — into the companies monetizing it, which he said have the most robust outlook and the highest probability of reaching their price targets
Bennett pushed on what he actually watches, given that money can stay put while a market moves on headlines and fears: is it earnings, demand and backlogs? Those are the things that pique the market's interest and drive the stock, he said, but they are a gauge rather than the object — "But again, it's the market interest that I follow." "The news and the earnings especially is the most current, the most active news, the most important news." Those are the turning points he makes his judgments and predictions from
AeroVironment Is the Odd One Out: A Broken Chart and an Enormous Backlog
The fifth pick, AeroVironment, is the only name on the list without a strong chart. "They've struggled for the last couple years with a couple of issues, including a lost contract, integration issues with an acquisition." This year, he said, the company is at a bottom and showing a firm one, the results show the business has regained traction, and "it's pretty well positioned within the drone market, which we know is a pretty hot market."
What he is waiting on is execution against a backlog he called massive and growing, "which just means sustained growth, sustained profit, profitability, and capacity for capital returns down the road"
Bennett noted drone stocks have had a hard summer, and that the channel has covered why: investors are waiting to see where federal allocations go, who wins the contracts and how large they are. She asked whether that is the only catalyst Hughes said it could be, but "I think a lot of it has to do with just market conditions. A lot of stocks that were flying high got sold off over the summer." In AeroVironment's case the weakness runs back further than this summer, and profit-taking has been a headwind every time it has rebounded over the past year
On whether this is a falling knife, after Bennett noted the stock is down more than 40% over the last year: "I don't think that this one's got too much downside left in it. I think it's at the bottom and bouncing." It has solid support at these levels
The catalysts he expects: "the backlog is really enormous and it's going to be realized over the coming quarters", plus follow-on orders as the company expands into other segments of the defense industry "It's become more of a prime contractor than just a hardware provider."
Hughes's bottom line is that the AI trade is still early rather than late — the market is mispricing what AI is worth to the whole tech ecosystem, and the way to be positioned for that is to buy the companies the money is already moving into, using September's seasonal weakness to do it.
Products, Companies & Tools Mentioned
Nvidia (The first pick and the anchor of his argument: a wide beat, guidance he called "at least 2,000 basis points better than expected," and evidence that the market is not pricing AI correctly)
AMD (A regular on his monthly list; its strength unfolding at the same time as Nvidia's is his evidence for how big the AI market is)
MongoDB and its Atlas platform (The second pick — database services that clients use to build agentic workflows and automate back-end operations; critical to AI development now and to inference later. The one name he would not buy at the current price)
Snowflake (The third pick — lets highly regulated industries use AI behind secure firewalls and share one data source without copying files; growth reaccelerated and, he says, in a discovery phase)
Okta (The fourth pick — identity and access management, now being pulled up by agentic traffic, because every bot moving through a company's systems is another identity to track)
AeroVironment (The fifth pick and the only weak chart on the list — a lost contract and acquisition integration issues behind it, an enormous defense backlog ahead of it)
OpenAI (Named for where the AI story started: the first phase was mostly OpenAI and chatbots, before agentic traffic arrived)
Boot Barn (His example of how little the sector matters to him — he follows where the market's money is flowing, and today that is AI)
Books & Resources Mentioned
Get the latest market chatter as it happens: