This page tracks every stock BWB - Business With Brian has discussed on YouTube, extracted from AI transcription of their 7 most recent videos covering 35 tickers. They are currently bullish on CRDO, GLW, STRL, ORCL, BHTI, MOD and others. Bearish on CAT, OKLO, SMR. Each take below includes their stated reasoning and, where given, a price target.
No explicit macro commentary beyond noting Credo ships to all five largest U.S. hyperscalers, implying strong AI/data-center infrastructure demand.
| Ticker | Action | Why | Target | When |
|---|---|---|---|---|
CRDO●●● | buy | Credo's proprietary low-power chip technology bridges copper and optical interconnects inside and beyond the rack, with an end-to-end stack scaling to 1.6 terabit and shipping to all five largest U.S. hyperscalers. Revenue tripled from $135M to $437M per quarter in six quarters, gross margins run at ~68% (software-like for a hardware company), operating margins nearly doubled to 37%, and the company guides for over 80% revenue growth next year. | — | 6d ago |
STRL●●● | buy | Sterling is the go-to contractor for hyperscaler data center site work, with data center revenue growing 174% YoY and now ~70% of total revenue. It has $5.15B in backlog against $2.49B in revenue, locking in over two years of work, and trades at a PEG near 0.88. | — | 1w ago |
ORCL●●● | buy | Oracle's OCI cloud is the neutral AI compute provider, and while capex has exploded to ~$56B pushing free cash flow deeply negative, the buildout is backed by a $638B signed backlog (up 363% YoY). Revenue is set to roughly double to ~$140B within three years, and FCF should flip positive within two years. | — | 1w ago |
MOD●●● | buy | Modine builds full-stack liquid cooling systems for AI data center racks that now pull 50-130 kW, with data center revenue growing 73% last year to about a third of the company. Management targets ~$2B by 2028, the company is spinning off legacy auto parts to become a pure play, and it trades at a PEG near 0.6. | — | 1w ago |
AVGO●●● | buy | Broadcom designs custom AI accelerators for hyperscalers looking to reduce dependence on Nvidia, with AI revenue on track to nearly triple in a year and ~$73B in booked AI orders. It generates ~$27B in free cash flow, and while it looks expensive at 60x trailing earnings, forward earnings are ~24x with a PEG near 0.48. | — | 1w ago |
LITE●●● | buy | Lumentum makes 50-60% of the EML laser chips inside high-end AI optical transceivers and is the only supplier shipping 200GB per-lane versions at volume, designated for Nvidia's next-gen Rubin platform. Revenue climbed 90% YoY last quarter, operating margin swung to +22%, and Nvidia invested $2B into the company. | — | 1w ago |
CRDO●●● | buy | Credo created and owns ~88% of the active electrical cable market for AI clusters, with revenue more than tripling (up 206%) last year and operating margins swinging from -19% to +33%. At a 68% gross margin and a PEG well under 1, it remains cheap relative to its growth. | — | 1w ago |
SNDK●●● | buy | SanDisk pioneered high bandwidth flash (HBF) that matches HBM speed but holds 8-16x more data, co-writing the industry standard with SK Hynix while Samsung is still at concept stage. Revenue exploded 251% in a quarter, the company wiped out all debt, authorized a $6B buyback, and trades under 10x next year's earnings. | — | 1w ago |
TSEM●●● | buy | Tower is one of only ~3 companies making silicon photonics at scale, serving as the open neutral foundry for Marvell and Nvidia's photonic chips. Silicon photonics revenue more than doubled YoY, it has booked $1.3B of photonics orders for 2027 with $290M paid up front, and earnings are growing 50%+ per year with a forward PEG near 0.91. | — | 1w ago |
COHU●●● | buy | Cohu makes semiconductor test handlers and inspection systems for AI chips, with AI only ~2% of sales today but a $750M AI test pipeline not yet booked. Orders are already up 57% and the stock is priced for the cyclical downturn it's only now climbing out of, right as AI demand arrives. | — | 1w ago |
INOD●●● | buy | Innodata produces expert-labeled training and testing data for AI models, already serving 5 of the 7 largest tech companies. Revenue nearly tripled in two years to $252M with a swing to $32M in profit, and the agentic AI wave could turn lumpy project work into recurring revenue across the broader economy. | — | 1w ago |
ET●●● | buy | One of the largest natural gas pipeline operators with 130,000+ miles of pipe and ~90% fee-based contracts, making it the fastest path for data centers needing gas. Already piping fuel to Oracle's data centers and building new pipe in Texas specifically for data centers. Pays a 7% dividend covered twice over by cash flow, appealing to both dividend and growth investors. | — | 2w ago |
ETN●●● | buy | Power management company building the electrical backbone inside data centers—switchgear, breakers, and power distribution systems. A pure picks-and-shovels play since all electricity must run through Eaton's equipment regardless of source. Data center orders up 240% in a year, $14B backlog, and data centers now over 20% of sales. Partnered with Nvidia on 800V systems and acquired Boyd Thermal for liquid cooling. | — | 2w ago |
GEV●●● | buy | One of only a handful of companies that can build large frame gas turbines at scale, with factories capped at ~10 GW/year while demand has blown past that ceiling—order book sold out through 2030. This has handed enormous pricing power, with 10-20% price increases per new turbine. Also building one of the first small nuclear reactors in the western world, making it a win-no-matter-what name. Free cash flow swung from burning billions to nearly $4B. | — | 2w ago |
HWM●●● | buy | The pick behind the pick—makes single crystal turbine blade castings that survive temperatures beyond metal's melting point, and is essentially the only real at-scale choice besides a unit buried inside Berkshire Hathaway. EPS grew 6x since 2021, margins climbed from 17% to 26%, and the stock rose ~10x. A second growth wave is stacking on as its industrial turbine business is expected to double over the next few years. | — | 2w ago |
Credo's proprietary low-power chip technology bridges copper and optical interconnects inside and beyond the rack, with an end-to-end stack scaling to 1.6 terabit and shipping to all five largest U.S. hyperscalers. Revenue tripled from $135M to $437M per quarter in six quarters, gross margins run at ~68% (software-like for a hardware company), operating margins nearly doubled to 37%, and the company guides for over 80% revenue growth next year.
Corning is a 175-year-old materials company that draws optical fiber and holds the largest share of the world's fiber market at around 20%, positioning it to benefit from the shift to photonics in data centers.
Sterling is the go-to contractor for hyperscaler data center site work, with data center revenue growing 174% YoY and now ~70% of total revenue. It has $5.15B in backlog against $2.49B in revenue, locking in over two years of work, and trades at a PEG near 0.88.
Oracle's OCI cloud is the neutral AI compute provider, and while capex has exploded to ~$56B pushing free cash flow deeply negative, the buildout is backed by a $638B signed backlog (up 363% YoY). Revenue is set to roughly double to ~$140B within three years, and FCF should flip positive within two years.
Biostem is a regenerative medicine company using proprietary perinatal tissue allografts for chronic wound care, with clinical data showing 53% healing probability vs 31% for standard care. The company has filed a Form 10 with the SEC to uplist from OTC to NASDAQ, which would expand capital access and investor reach.
Modine builds full-stack liquid cooling systems for AI data center racks that now pull 50-130 kW, with data center revenue growing 73% last year to about a third of the company. Management targets ~$2B by 2028, the company is spinning off legacy auto parts to become a pure play, and it trades at a PEG near 0.6.
Broadcom designs custom AI accelerators for hyperscalers looking to reduce dependence on Nvidia, with AI revenue on track to nearly triple in a year and ~$73B in booked AI orders. It generates ~$27B in free cash flow, and while it looks expensive at 60x trailing earnings, forward earnings are ~24x with a PEG near 0.48.
Lumentum makes 50-60% of the EML laser chips inside high-end AI optical transceivers and is the only supplier shipping 200GB per-lane versions at volume, designated for Nvidia's next-gen Rubin platform. Revenue climbed 90% YoY last quarter, operating margin swung to +22%, and Nvidia invested $2B into the company.
SanDisk pioneered high bandwidth flash (HBF) that matches HBM speed but holds 8-16x more data, co-writing the industry standard with SK Hynix while Samsung is still at concept stage. Revenue exploded 251% in a quarter, the company wiped out all debt, authorized a $6B buyback, and trades under 10x next year's earnings.
Tower is one of only ~3 companies making silicon photonics at scale, serving as the open neutral foundry for Marvell and Nvidia's photonic chips. Silicon photonics revenue more than doubled YoY, it has booked $1.3B of photonics orders for 2027 with $290M paid up front, and earnings are growing 50%+ per year with a forward PEG near 0.91.
Cohu makes semiconductor test handlers and inspection systems for AI chips, with AI only ~2% of sales today but a $750M AI test pipeline not yet booked. Orders are already up 57% and the stock is priced for the cyclical downturn it's only now climbing out of, right as AI demand arrives.
Innodata produces expert-labeled training and testing data for AI models, already serving 5 of the 7 largest tech companies. Revenue nearly tripled in two years to $252M with a swing to $32M in profit, and the agentic AI wave could turn lumpy project work into recurring revenue across the broader economy.
One of the largest natural gas pipeline operators with 130,000+ miles of pipe and ~90% fee-based contracts, making it the fastest path for data centers needing gas. Already piping fuel to Oracle's data centers and building new pipe in Texas specifically for data centers. Pays a 7% dividend covered twice over by cash flow, appealing to both dividend and growth investors.
Power management company building the electrical backbone inside data centers—switchgear, breakers, and power distribution systems. A pure picks-and-shovels play since all electricity must run through Eaton's equipment regardless of source. Data center orders up 240% in a year, $14B backlog, and data centers now over 20% of sales. Partnered with Nvidia on 800V systems and acquired Boyd Thermal for liquid cooling.
Sponsored content. Sits at the intersection of aging demographics, GLP-1 users losing lean muscle, and 200 million unhappy protein drinkers. Its Rejuvenate product triggers muscle protein synthesis at 5 calories with zero dairy and 15-minute absorption, backed by 23 years of research. Just entered 4,500 CVS stores with 2,000 more locations coming, and the stock is near its cheapest ever with a tiny market cap.
One of only a handful of companies that can build large frame gas turbines at scale, with factories capped at ~10 GW/year while demand has blown past that ceiling—order book sold out through 2030. This has handed enormous pricing power, with 10-20% price increases per new turbine. Also building one of the first small nuclear reactors in the western world, making it a win-no-matter-what name. Free cash flow swung from burning billions to nearly $4B.
The pick behind the pick—makes single crystal turbine blade castings that survive temperatures beyond metal's melting point, and is essentially the only real at-scale choice besides a unit buried inside Berkshire Hathaway. EPS grew 6x since 2021, margins climbed from 17% to 26%, and the stock rose ~10x. A second growth wave is stacking on as its industrial turbine business is expected to double over the next few years.
Makes aeroderivative gas turbines—the lighter, faster-to-install turbines data centers grab when they can't wait for heavy frame machines. Booked ~$1B in data center power orders in a single quarter, matching an entire prior year. Moving to acquire Chart Industries to add cooling capabilities. The creator sees mispricing because Wall Street still treats it as an oil services company, making it the cheapest and least noticed name on the list with the most analyst upside at 35%.
Market underrates it because of its truck engine association, but its power generation arm builds large natural gas engines for data center microgrids. Doubled data center revenue to ~$3.5B, power generation sales grew 30% in a quarter, segment margins near 30%, and order book runs through 2028. Trades at ~23x forward earnings—a fraction of pure plays—with 15+ years of dividend raises, offering a quiet income-paying way to own the data center power buildout.
Pipeline operator transforming into a power company by building and owning on-site gas power plants at data centers, earning far more than just shipping fuel. Building a $2B 400MW plant for Meta on a long-term contract, with $5B+ committed to a fleet of such plants. Cash earnings grew from $4B in 2019 to over $7B today. The pricier, more growth-oriented way to play the gas side.
Largest natural gas producer in the country, sitting on the Marcellus Shale with the lowest cost structure. Locked an exclusive deal to supply gas for the 4.4GW Homer City AI campus. While pipeline names have run up, EQT is basically flat on the year and cheap, giving it the most torque in the group—any lift in gas demand from data centers drops almost straight to the bottom line. Analysts see 34%+ upside.
BWXT is the only profitable nuclear stock and sole manufacturer of US Navy submarine and carrier reactors, with a record $8.6B backlog up 77% YoY against under $4B in revenue. It combines a defense monopoly, small modular reactor components, TRISO fuel for Kairos/Google, and a cancer isotope business, with FCF surging from $46M to ~$300M. The only concern is entry price near 43x forward earnings, making pullbacks the preferred entry.
NVIDIA's real monopoly is CUDA software with 6M developers and over a billion chips deployed, creating lock-in no competitor can break. Revenue grew 8x in three years from $27B to $215B while the PE fell from over 100 to ~30, with a PEG near 0.5 indicating it's cheap relative to growth. The stock is oversold on its 200-day line after the pullback, which the creator sees as the ideal entry.
Alphabet runs four engines at different speeds: Search growth nearly doubled to 19% as AI mode crossed a billion users, Google Cloud revenue grew 63% with margins doubling to 33% and a $460B backlog, YouTube is now bigger than Netflix, and Waymo runs 400K paid rides weekly. It's the cheapest mega at ~24x forward earnings with a PEG near 1.3 and the most oversold name in the group after recent pullbacks.
Arista's single OS across all switches is why cloud giants standardize on it, and the industry shift from NVIDIA's InfiniBand to open Ethernet plays directly to its home turf. AI networking revenue doubled last year and is guided to double again, with $6B in deferred revenue, zero debt, and $12B net cash. However, at 44x forward earnings with a PEG near 2—the richest in the group—price matters more here, and the creator wants a deeper pullback toward the 50-day line for a better entry.
Rambus makes the traffic control chips (including the RCD conductor chip) needed in every high-speed AI memory module, getting paid on memory shipments regardless of which manufacturer wins. It transformed from a money-losing patent shell to a debt-free chip company with operating margins swinging from -38% to +37%. However, it's the highest risk name on the list at 60x earnings with a PEG above 2, facing a DOJ document request, a securities inquiry, and a CFO departure, so it's sized down and only bought on real pullbacks.
Fabrinet is the factory that physically builds optical transceivers for Cisco, NVIDIA, Coherent, and Lumentum, winning no matter whose brand is on the box, with high switching costs due to years-long clean room certification. Data center interconnect revenue grew 90% YoY and demand is running ahead of capacity, driving a new plant build. It's a thin-margin business (12% gross) but earns ~20% ROIC with no debt; the risks are small size, customer concentration (NVIDIA + Cisco ~50% of sales), and a PEG near 1.5 even after a 30% pullback.
Calix builds the software platform running on home Wi-Fi gateways for internet providers, positioning it at the front line of edge computing. After a decade of losses, it transitioned to a recurring-revenue software model, doubling revenue and lifting gross margins to ~57%, with every new ISP subscriber paying monthly.
Ambarella is the purest edge AI vision play, with ~80% of revenue from edge AI vision chips used in cameras, drones, and robots, and 46M+ chips shipped. Revenue re-accelerated ~37% with narrowing operating losses, though it is still investing hard for growth and big profits are ahead.
Renesas is the world's largest microcontroller maker, now embedding on-device AI into automotive silicon with its 3nm Arcar X5H chip delivering 400 TOPS at 35% lower power, with Bosch and ZF designing on it. The business is profitable with high-teens operating margins, and acquisitions of Altium and Pictoris give it full workflow ownership; a recent headline loss was a one-time write-down.
Mobileye is the purest automotive edge vision play, with its IQ chip already in 230M+ vehicles, making it one of the most widely deployed AI hardware platforms on Earth. Revenue is climbing again (~27% most recent quarter) as IQ volume rebounds and the company launches its own robo-taxi fleet, though Intel's large stake creates overhang risk.
Coherent sits at the testing gate for edge and automotive silicon, with orders jumping ~57% YoY as the chip cycle turns, and its Eclipse platform is winning test slots for high-performance AI processors. Recurring spares and services now comprise ~60% of business, providing sticky long-term revenue once handlers are installed.
Already up 60% year to date as an obvious data center power play, making it a crowded winner with risk baked in. Michael Burry is shorting it. The creator uses it as a cautionary example of where the easy money has already been made.
Oklo is a hyped nuclear pure play that fell ~30% in the spring selloff and is still years away from building its first reactor, unlike BWXT which is already building and getting paid. The creator prefers BWXT as the only nuclear stock he would actually buy on the pullback.
NuScale is another hyped nuclear pure play that fell ~30% in the selloff and remains years away from its first reactor, making it inferior to BWXT which already builds and earns revenue. The creator would not buy these pure plays on the pullback.
The creator highlights Credo Technology as a high-growth chip company serving hyperscalers, emphasizing its proprietary low-power technology, explosive revenue growth, software-like gross margins, and strong forward guidance.
The creator discusses the shift from copper to photonics in data centers due to power and distance constraints, highlighting Corning as a key beneficiary due to its dominant market share in optical fiber.
Brian argues that the recent semiconductor bear market sell-off has dragged down strong, undervalued AI-related companies alongside overvalued ones, creating buying opportunities. He highlights 11 companies across the AI infrastructure stack—from data center site preparation and cooling to custom chips, optics, memory, testing, and training data—that are still growing rapidly but trade at reasonable valuations on a PEG basis. He views the broader sell-off as a mistake by investors who sold, and sees these names as still early in their growth curves with significant booked backlogs.
The creator argues that data centers are being forced by new state legislation to build their own on-site power plants, with natural gas turbines being the only realistic near-term solution since small nuclear reactors won't arrive until ~2030. He identifies a picks-and-shovels opportunity across the natural gas power supply chain—from producers to pipeline operators, turbine makers, blade casters, and electrical equipment companies—that gets paid regardless of which AI giant wins. He contrasts these quieter names with crowded winners like Caterpillar, which Michael Burry is shorting.
The creator argues the June AI pullback was a buying opportunity, not a bubble bursting, citing Micron's blowout earnings as proof the AI buildout is accelerating. He presents 8 stocks across the AI infrastructure stack—from power (Eaton, BWX Technologies) to chips (NVIDIA, Broadcom, Rambus) to networking (Arista) to optics (Fabrinet) to megacap platforms (Alphabet)—sizing heavily in mega caps and scaling down into smaller, higher-risk small caps for an asymmetrical portfolio.
The video contains no stock or macro content; it is simply a brief July 4th holiday greeting.
The creator argues that edge AI—running intelligence on devices rather than in data centers—is the next major investment wave, growing faster than the data center buildout. He highlights recent market pullbacks as buying opportunities and walks through seven companies positioned across the edge AI stack: software platforms, robot brains, vision chips, automotive microcontrollers, custom ASICs, automotive vision, and chip testing equipment.
Based on AI analysis of this creator's recent videos. Theses are paraphrases of the creator's own words, not verbatim quotes. Not financial advice.