Nitin Gupta retweeted
One of the stranger disconnects I've seen between public equity markets and the fundamentals of the underlying industry.
Market is increasingly concerned about an impending GPU glut, collapsing compute prices and the economics of the AI infrastructure buildout.
Meanwhile, $IREN just reported:
3-year contract pricing up 11% sequentially.
5-year contract pricing up 12%.
New H100 capacity contracted at 2x the pricing of a previous deal.
A short-term contract at $40M in annual revenue per MW.
That is not how customers typically behave when they expect abundant supply and rapidly falling prices.
To be clear, I'm not a buyer of $IREN and this is not an endorsement of the stock. Simply an observation about how disconnected the narratives in the equity market appear to be from actual customer behavior.
Nitin Gupta retweeted
Borrowed Conviction Ep.4 with @PasiHavia - #1 in the Citywire Global Equities category (2023-25 with his previous HCP Focus fund), now running UB Megatrends - how managing real money differs from being always up 100% on Twitter/X, how to deal with stress (a dog might help🐶), AI, US electrification, being burnt in China, why Estonia is a massive startup and unicorn powerhouse, $MU, $LLY, $ISRG and more.
Link in the first comment, as usual.
Nitin Gupta retweeted
Micron $MU will pay Netlist a total of $600M, $30M a quarter from Q4 2026 to Q3 2031, under patent license and settlement deals that end all pending litigation between them.
Nitin Gupta retweeted
Surprised $NLST isn't up more on this announcement:
"Under the cross-licensing agreement, Netlist will receive an upfront license fee of $200 million after withholding of Korean taxes, or $239 million gross. In addition, Netlist will receive 20 recurring quarterly license fees of up to $27.5 million, also net of tax, or $32.9 million gross, calculated in accordance with a revenue-based formula.
Over the five-year term, this represents aggregate net license fees of up to $750 million or gross licensed revenues of up to $897 million. Under the supply agreement, Samsung will supply Netlist with up to $300 million of memory products annually and up to $1.5 billion over the five-year term. This supply includes DRAM and NAND raw materials used to build Netlist proprietary solutions, as well as standard memory and storage products from Samsung for resale to our customer base."
They basically get almost their entire market cap in the next 5 years and supply to DRAM and NAND which they can resell + the validation and the investment from Samsung.
Nitin Gupta retweeted
Replying to @Mike10947310
Modular data centers could be a potential solution as well, especially in Tier-III cities. $DUOT's strategy is to set these up in small towns with excess or stranded energy capacity with no water usage which has been a huge point of contention amongst many communities kcbd.com/2026/05/28/new-data…
Nitin Gupta retweeted
Also worth mentioning this exchange for $PENG at Rambus' Rosenblatt conference. Figured this was worth a mention because we are still very early here, especially for a company sitting at only a $3.5b market cap with all these catalysts lining up. Kevin Cassidy explicitly called out their $500k, 11 PB kv cache appliance.
I think the key takeaway here is that HBM and DDR have to scale together as context windows blow up and spill over. Penguin has a the product that solves this exact problem.
Not to forget that Penguin's ability to procure supply is a major tailwind with close relationship with sk hynix (which the ceo just shared a post about on linkedin recently) gives them the massive edge to procure the silicon to build and ship these with massive ASPs.
Must read article on how agentic ai is completely changing the memory pooling landscape. Highly recommended for anyone in $PENG $ALAB or $RMBS semiengineering.com/agentic-…
Was very early to this trade (Jan 2026), @KairosPraxis can attest to it. 🙂
Been moving away for purely social arb but didn't feel like sharing thoughts on AI infra plays on here because I didn't think I bring anything special to the conversation and didn't want to add to the noise on this platform.
@AstutexAi convinced me that my thoughts are not as dumb as I think and I should share them regardless so going to try to do that more.
Nitin Gupta retweeted
Last week I sat down with @nitininvests for the first episode of a stream we plan to do regularly, called Borrowed Conviction.
We covered a lot of ground: process and performance, optics, shitcos, power semis, longevity, token pricing, luxury experiences, 800V DC, and plenty more. link in first reply $SVCO $RDDT $TTWO $PTON $LIND
Nitin Gupta retweeted
Replying to @nitininvests
This guy called this when sentiment was terrible.
$DUOT is now +71% in just 10 days.
Conviction > emotion.
substack.com/@alphametrics1/…
Nitin Gupta retweeted
Back in $ZVIA after the call on Wednesday. We traded this very well in Q4 2024 and were lucky to take profits on time. The call was extremely encouraging and the company is showing signs of a real inflection here:
1. Revenue & Volume:
This was Zevia’s best quarter since going public as they recorded first quarter growth of 21%. The growth was entirely volume driven (20.4%) which is a great sign.
2. Channel Expansion:
They executed a successful national Costco rotation and saw acceleration in velocity at Walmart. The e-commerce business is also outperforming expectations. They have expanded in Kroger as well adding incremental flavors along with regional players like HEB and Publix. This quote was particularly interesting to me about Costco:
“The advantage of the national rotation… strengthens our velocities in the markets in which we have permanent distribution as well as helps to spur discussions about future rotations… the hope is that we would get another national rotation in the balance of the year.” — Amy Taylor
3. Cardi B Campaign upside:
This is arguably the most exciting social arb development. Zevia signed Cardi B as a brand ambassador who has the 25th most followed Instagram account. Thisc ould be a step change in reach and awareness for Zevia which they have lacked for years.
This awareness seems to be translating already as the management said that between Cardi B and other marketing programs during the month of March, Zevia saw its highest ever organic social media reach and the highest level of social media engagement of any month since the brand’s launch.
4. Profitability:
Adjusted net loss was just -$0.1M in Q1 2026, compared to -$4.2M a year ago. Selling and marketing as a % of net sales improved dramatically from 40.3% → 31.5%, demonstrating strong operating leverage as revenues scale.
5. The Setup for the Rest of 2026
Net sales guidance raised to 7% growth at midpoint, which I believe is being sandbagged massively after the impressive Q1 results. Q2 comps might be a little harder to lap but comps will get easier in Q3 and Q4.
TLDR: Q1 2026 represents a clear inflection driven by and record revenue growth, a near break-even EBITDA result, broad channel distribution gains, and the launch of the highest-reach marketing campaign in the company’s history. The packaging refresh is still rolling out and the Cardi B summer campaign hasn’t fully hit yet, meaning the biggest catalysts are still ahead. Low conviction position due to the chart looking ugly but risk/reward appears asymmetric
$ZVIA Investment Thesis: Asymmetric Upside Potential Amid Emerging Trends
After recent analysis, I’ve shifted my stance on Zevia, due to the October Amazon data showing a significant surge in e-commerce sales. While the upcoming Q3 results may still underperform based on September data, I believe the guidance could surprise positively, creating an asymmetric risk-reward scenario.
Analysis based on Amazon sales data estimates their Amazon sales for the past month around $2.5M, which annualizes to roughly $30M. Their e-commerce sales are generally around $3M/quarter. With e-commerce making up close to 9% of Zevia’s total sales, this implies total quarterly sales could approximate $75M, and on an annualized basis, reach $300M (currently @ $160M). Given this growth trajectory, an emerging company in the Stevia beverage category should conservatively trade at 2x revenue, putting the potential market cap around $600M—representing an 8x upside from current levels. There are a lot of assumptions in the calculation which are not perfect but the margin of safety looks good due to the potential upside.
Management has also mentioned a $12M cost-cutting initiative and hinted at a major distribution partnership announcement. Sentiment on TikTok is starting to build, with the Stevia-based product gaining popularity among health-conscious consumers, positioning Zevia to grow alongside competitors like Olipop and Poppi.
This is a speculative play due to liquidity concerns and a potentially weak quarter coming up, but I am initiating a 2% position based on the substantial upside. The stock has regained compliance, and any good news could lead to outsized rewards.
Keeping an eye on Dr. Zevia (healthy version of the #2 soda - Dr. Pepper, Caramel Soda and Root beer flavours which have been driving the sales.
Nitin Gupta retweeted
Hey guys, sorry I'm late, it was a crazy busy day at work. I see you guys already shared some great insights on the $PENG Needham call!
For me, the biggest takeaways were the following:
1. If you look at the mid-point of the 12% guide and the segment guides from the Q2 call, my calculations imply an AI factory segment revenue of $1.02B for FY'27. However, with Kash explicitly saying this is well over a $1 billion business, it makes me think they are being conservative and should easily beat that 12% guide for the year.
2. With the guidance basically achieved, there is a very good chance their backlog for next year is incredibly strong. They mentioned it takes 3-6 months to convert backlog into revenue, so that pipeline is locked in. Also, "Nobody else in the market has a product like MemoryAI, which is helping our customers and helping us win deals." That CXL product is a massive differentiator right now and their partnership with $MRVL makes this a great opportunity (they also have partnership with $DELL, $AMD, $NVDA as we know).
3. SK Hynix being a strategic investor is incredibly bullish and honestly something I didn't fully realize until they mentioned it. This completely de-risks the supply issue. Procuring demand is a massive constraint for everyone else, but them saying we've been able to access it and serve as much demand as we can and that the relationship helps with margins is quite massive.
Maybe too much of a stretch but long-term, this could possibly lead to SK Hynix promoting PENG's products against competitors. It also gives PENG immense leverage with rack builders and larger enterprises because clients know PENG actually has the supply to deliver. That alone demands a huge premium and is a massive moat in my opinion.
Nitin Gupta retweeted
Surprised to see no movement in $BE after $BAM said this on their call today:
“The first deal we did in our AI infrastructure fund was our partnership with Bloom Energy. That was announced, I wanna say, 6 or 9 months ago, a $5 billion partnership. We are already in conversations to expand that partnership, not by percentages, but by multiples. That's very reflective of the opportunity set and the scale that we expect to play.”
Nitin Gupta retweeted
I usually do not brag about my performance as it feels counterproductive and puts on unnecessary pressure - I always expect my portfolio to immediately go into a drawdown right after. But it serves its purpose here - I made 25x my money in 6 years (though with three painful 35% drawdowns) which allowed me to quit my job and concentrate solely on managing my own money. And I did this almost exclusively with consumer stocks as I constantly lost money on tech (that changed only recently) and decided to concentrate where I thought I had an edge: alternative data in B2C businesses.
In 2020 I made 700% return like a lot of people and thought it was an easy game and then in 2021 i got burned really badly, my story is very similar to Kristjan Kullamägi’s described in Market Wizards (though with magnitudes less money, of course). These big drawdowns were actually the best lessons on risk management and position sizing – things you would never learn studying for the CFA. Once you think you are smart, the market punches you in the face immediately and hard. When Stan Druckenmiller talked about vomiting once or twice a week due to a drawdown, I knew that exact feeling. Eventually, you try to avoid that at all costs.
My biggest problem was always a lack of diversification, I was heavily concentrated in a few best ideas where I had conviction, but that resulted in larger drawdowns. And I still think it makes more sense than diversifying across 20 ‘meh’ ideas (I regularly ran with just 2-3 consumer companies which is clearly too low of a number). But you really have to build conviction in those ideas – only then it works. That is also very difficult with borrowed ideas – you cannot borrow conviction.
There are a few things that are important if you want to be successful in this game:
1. A constant flow of new ideas (making it much easier to drop the ones that aren't working)
2. Getting rid of losers quickly (but only if your thesis was violated or you have doubts)
3. Risk management and position sizing
I had to build a tool that would help me with those, at least with the first two. It matches my style – which is somewhere between a trader and investor – and it probably won’t fit everyone. There are multiple ways to make money and with AI coding (thanks @lovable, was a bit painful at times but works smoothly now) you can build anything that matches your exact workflow, personality or risk-taking appetite. But I genuinely believe a curious guy with a Claude subscription and a willingness to develop his skills can beat institutional money. I managed institutional money for quite a while (though clearly not a a top-tier of course, but similar in style to Polen which is having exactly the same issues) and these are very tough times for the industry, I have no doubt about it.
So check this out. It’s a very basic version right now, but more features and tiers are upcoming! I would be very grateful for any re-post as it would help to save on marketing and reach more eyeballs!
EquityDesk.ai is live:
- Stage Analysis for US, Europe and Asia
- Deep Research and Slide decks with standardized prompt for 3000+ global companies
- Earnings Call Ratings
- Industry Group Leaders/Laggards
- Alternative Data (Google Trends, TikTok, Reddit and Wikipedia) weekly
Nitin Gupta retweeted
Got a couple of questions regarding this in DMs so I think I might have not done the best job explaining this. Basically, the top-line is artificially understated, but the underlying earnings quality is actually the best it has been in years.
With $SIF's customers starting to supply their own raw materials, SIFCO no longer bills for the metal, so it doesn't hit their Revenue or their COGS. They only bill for their labor and services. This makes revenue look smaller, but it mathematically supercharges their Gross Margin %. SIFCO makes the exact same profit in dollar terms but on a leaner revenue base.
This is drastically reducing $SIF's working capital needs. They no longer have to tie up their own cash to finance the massive metal inventories. My supply chain data confirms this thesis. Historically, $SIF got burned by this. They had fixed-price contracts but were forced to buy their own materials. When commodity prices spiked or supply chains choked, SIFCO ate the costs, creating wild margin uncertainty and crushing profits.
This new dynamic fixes that historical overhang. By letting massive aerospace customers use their own purchasing power to buy the metal, $SIF completely offloads commodity price risk and supply chain bottlenecks. $SIF has essentially traded it's low-quality, high risk revenue for high quality, stable, and predictable margins. The top line is understated, but the fixed-cost absorption and margin quality appear to be inflecting. Very bullish long-term setup IMO.
$SIF appears to be offering a massive info arb right now. the more I research this, the more excited I get. Top-line revenue numbers appear to be optically compressed too with some customers now starting to supply their own raw materials. The value of these materials is being completely excluded from the reported net sales in Q1, something that was not present in their earlier annual filings.
While this depresses headline revenue, it simultaneously reduces COGS, meaning the impact on gross profit is entirely neutral. If SIFCO had procured these materials themselves, their reported top-line for the quarter would have been noticeably higher. The quality of these earnings is vastly superior to what the market is currently pricing in.
Furthermore, management explicitly noted that their recent margin surge is being driven by a strategic shift in sales mix and the exit of lower-margin business. As this transition continues, don't be surprised to see SIFCO's numbers continue to inflect with structurally higher margins as Mr. Market finally catches up to the story.
Nitin Gupta retweeted
Looking deep into $SIF (SIFCO Industries). It looks incredibly enticing optically with a highly relevant thematic tailwind and a great setup! Well positioned for the aerospace and defense recovery, driven by surging Boeing 737 MAX/787 production and record, stable F-35 contracts from Lockheed. The chart looks great with a low Float and Peter Abrahmson acquiring 6.4% of the float. The backlog is expanding rapidly and jumped $20 million from $119.2M at the end of Q4 2025 to $139.5M by Q1 2026.
Important to note that the earnings quality is high but is currently being masked by LIFO accounting; if FIFO had been used, inventory values would have been $11.1 million higher in Q1 2026, meaning the company's true earnings power is significantly depressed on paper.
Valuation appears cheap too with Enterprise Value below $100M. Given its recent return to profitability and a Q1 sales run-rate pacing near $96M annualized.
Could do really well here as throughput and margins continue to inflect!
Nitin Gupta retweeted
Of course not an expert on this market but $RMBS seems like the obvious long after that $AMD announcement today.
Rambus could really scale with this new TAM expansion for CPUs. Their traditional memory interface business is tied entirely to server volume, and the complexity of these new AI platforms acts as a huge multiplier for them.
Last year the CEO said, “The content on an MRDIMM is 4x the content of what we offer today on a standard DIMM, the timing is going to be linked to the rollout of platforms from mainly Intel and AMD.”
This 4x multiplier on content should scale with more CPUs sold. The CEO also said, “If you look at the types of architectures that inference requires, you clearly see that the ratio between CPUs and GPUs is changing in favor of CPUs. So overall, that’s a very good thing for us.”
They appear to be completely agnostic to the architecture. Every new CPU hitting the data center would need memory modules and buffer chips. It feels like a great way to play this server buildout because you can win regardless of which company ultimately wins the CPU war.
Nitin Gupta retweeted
Over the last several weeks, the market has started realizing the sheer importance of testing and validation companies within the AI and semiconductor supply chains. Over the past few days, I have been adding a some small-cap, pure-play exposure to this theme and initiated a position in Calnex Solutions $CLX.L. Please keep in mind that as a smaller cap, this stock will be volatile. But the fundamental setup here is incredibly asymmetric, and I believe the market is drastically mispricing their operating leverage inherent in their business model.
The Viavi/Spirent Read-Through & Growth Inflection:
Calnex is deeply integrated with $VIAV, as VIAVI white-labels Calnex's products. Historically, Calnex worked closely with Spirent (which made up nearly 70% of Calnex's business). When Keysight announced its acquisition of Spirent, Calnex was effectively boxed out and had to aggressively pivot to diversify its client base.
However, VIAVI recently closed the acquisition for Spirent's high-speed Ethernet and channel-emulation test lines (divested as part of the Keysight deal). This brought the Spirent division and its partnership back to VIAVI, meaning Calnex got their most critical partner back in November 2025.
VIAVI has been posting blockbuster numbers recently, which acts as a massive read-through for Calnex. Back in November, Calnex guided for low single-digit revenue growth for the upcoming years on their H1 earnings call but on their latest Spring 2026 trading update they are going to do 19% revenue growth with highly robust margins and profitability.
Multi-Theme Ecosystem & The Road to 1.6T: Just as VIAVI launched 1.6T test systems specifically for AI workloads, Calnex has their own 1.6G (1.6T) product launching next year.
Top-Tier Partnerships: Calnex's ecosystem includes titans like Nokia, NVIDIA, and Arista. As Nokia and NVIDIA scale their new AI-RAN and digital twin cloud environments, Calnex's emulation tools are critical for validating interoperability and performance.
Compounding High-Margin Revenue: Warranty and support revenues are scaling alongside their booming data center hardware business, layering on highly profitable, recurring cash flow.
Financials & Valuation: There is tremendous operating leverage embedded in this business model. My model currently puts them at a shockingly low ~5.5x EV/EBITDA based on the latest profitability updates.
Furthermore, the balance sheet is strong As of March 31, 2026, Calnex held £9.3m in cash, with absolutely zero debt. Additionally, due to the phasing of shipments through Q4, there is a further £3m in customer receipts expected to land in the early months of FY27.
Finally, the founder and CEO has deep skin in the game, owning 20% of the company.
TLDR
Calnex ($CLX.L) is a highly profitable, zero-debt network testing play trading at roughly 5.5 EV/EBITDA. They are deeply tethered to VIAVI's explosive AI datacenter growth. After guiding for low single digits, they are now tracking for 19% growth. They just released a breakthrough 400G AI validation tool, are getting repeat hyperscaler orders, and have 1.6T solutions coming in 2027.
Nitin Gupta retweeted
$SIF
+ Produces forged components for aerospace and defense industries which are experiencing significant structural tailwinds
+ Trading at 1x sales and 6x run rate EBITDA
+ No debt
SIFCO's core business is the aerospace and defense sector with the F-35 and Boeing 737/787. Currently, nearly every major commercial airplane and military helicopter in operation contains a structural component made by $SIF.
1) $LMT F-35: Proposed $1.5 trillion Pentagon budget includes 85 new F-35 jets. This will expand the base fleet and lock in decades of maintenance components.
2) $BA 737/787: Both the 737 and 787 are expected to ramp production by 20%+ throughout this year.
Because of this, the backlog is expanding rapidly and jumped $20 million from $119.2 million at the end of Q4 2025 to $139.5 million by Q1 2026.
They trade at 1x sales because their margins have historically been terrible. However, following their strategic restructuring, runoff of older contracts, and exit from European operations, $SIF has aggressively shed lower-margin business.
i) Q1 2026 gross profit jumped to $5.2 million vs. from $0.9 million.
ii) Adjusted EBITDA was $3.9 million vs. $(0.2) million in the first quarter of fiscal 2025.
If you run rate that adjusted EBITDA, that implies $SIF is trading at 6x EBITDA. Just taking the core business into account, the company should grow the top line 10%+.
Not financial advice. Do your own diligence.
h/t: @nitininvests for bringing to my attention




























