Understand the system.
Then price the stock.
The field guides explain who controls each layer of the AI stack and why the handoffs create bottlenecks. The company analyses then separate business quality from stock valuation, with assumptions, risks, and the evidence kept visible.
Field guides to the AI stack
Start with the system: chips, manufacturing, networking, power, powered campuses, cloud operators, and the customers that must turn all of it into economic value.
The Nvidia Story
It is much bigger than Nvidia.
Who are the companies involved between silicon and the final token?
AMD Doesn’t Need to Beat Nvidia
It just needs to become the second standard.
The AI market may be too large, expensive, and strategically important for one company to own.
The Next AI Bottleneck
It is not chips. It is electricity.
The chips are ready. The grid is not. Bloom Energy can provide speed, existing nuclear can provide scale, and advanced reactors may determine what comes next.
The AI Infrastructure Trade
A bottleneck is not a moat until it pays for itself.
Owning a bottleneck is not enough. You have to get paid for fixing it.
Company valuation research
Read one company at a time, then compare the assumptions. A low share price says nothing about operating risk, and an exceptional product does not automatically produce durable cash flow.
Use these collections to keep each company's role in the supply chain straight. The complete company library remains alphabetical below.
Optical Networking
The links, switches, lasers, fibers, and substrates that move data through an AI cluster.
Semiconductors
The foundries, equipment, process control, memory, storage, power, and custom silicon behind compute.
AI Infrastructure
The systems, power equipment, cooling, cloud capacity, and powered campuses that turn chips into usable compute.
Mega-cap Platforms
The consumer and cloud platforms funding, distributing, and monetizing the next computing cycle.
Software and Cybersecurity
The software platforms that convert data, automation, and security into recurring enterprise value.
Consumer, Fintech and Media
The networks, marketplaces, and brands competing for attention, transactions, and recurring demand.
Power and Nuclear
The electricity suppliers positioned behind rising data-center and industrial demand.
Space, Healthcare and Emerging Technology
Capital-intensive platforms pursuing new markets where execution matters more than the story.
ETFs, Bitcoin and Leveraged Vehicles
Portfolio vehicles whose structure, exposure, and path dependence matter as much as the underlying asset.
A to Z
AAOI at $101
The capacity bet is working, but shareholders have already paid for the factory.
Applied Optoelectronics may become one of the biggest beneficiaries of the optical-transceiver shortage. The harder question is how much of that future already belongs to today's shareholder.
AAPL at $341.46
The ecosystem is extraordinary. The price needs another growth curve.
Apple still owns the most valuable consumer technology ecosystem in the world. At this price, owning the ecosystem is not enough. It must keep turning loyalty into faster earnings growth.
Affirm at $71.50
The network is scaling. Credit still decides the outcome.
The payment network is compounding. The credit cycle still gets the final vote.
AMAT at $485.85
The factory is full. So is the valuation.
Applied Materials is selling the tools that make the AI buildout possible. The harder question is how much of that future a buyer is paying for today.
AMZN at $249.98
Three great machines. One enormous infrastructure bill.
Amazon owns three extraordinary machines: commerce, logistics, and cloud. The question is whether the newest infrastructure bill will make those machines more valuable or merely more expensive.
Arista at $206
The best business in AI networking may still be a difficult stock.
Arista has the margins, cash, software, and execution investors want. The valuation question is whether even an exceptional company can outrun exceptional expectations.
AST SpaceMobile at $61.81
The signal works. The economics still have to launch.
AST SpaceMobile is building a cellular network in orbit. The stock is already valuing the constellation as a global commercial system.
AVGO at $353.70
AI growth is extraordinary. So is the burden of proof.
Broadcom has become one of the most powerful toll collectors in AI. The stock now assumes that custom silicon, networking, and software can keep compounding together.
AXT at $75.90
The scarce wafer is real. So is the price.
AXT controls one of the least visible bottlenecks in optical networking. Investors still have to separate a genuine supply-chain advantage from a stock that already expects extraordinary execution.
Constellation at $263.40
Scarce clean power. A valuation that already sees it.
Constellation owns scarce clean generation. The investment question is how much of the coming power shortage the current price already captures.
Costco at $922.23
An exceptional business can still be an expensive stock.
The membership machine is exceptional. The valuation assumes exceptional becomes permanent.
Salesforce at $234.46
The customer record is valuable. AI must restart the growth curve.
Salesforce owns the customer record, but AI is forcing it to prove that a mature software empire can become a growth company again.
CrowdStrike at $252.88
The platform is winning. Owner earnings still have to catch up.
CrowdStrike is becoming the operating system for enterprise security. The stock still needs platform expansion to become owner earnings, not just adjusted earnings.
DELL at $563.21
The AI factory is shipping. Now the cash must follow.
Dell has become one of the fastest-growing ways to buy the AI factory. The stock now needs that factory to become a durable earnings engine, not a temporary order wave.
ETN at $442.23
The electrical bottleneck is real. The premium already knows.
Eaton sells the electrical nervous system behind the AI buildout. The business is exceptional. The stock assumes that exceptional growth lasts.
Corning at $154
AI has turned fiber into a growth story. Expectations have moved first.
Corning owns critical materials, manufacturing knowledge, and customer relationships across the AI data-center buildout. The question is whether a century-old industrial company should be valued like a perpetual AI compounder.
Alphabet at $344
Search funds the future. AI is making that future expensive.
Search is still the cash machine. Cloud is becoming a second one. The risk is that funding both requires more capital, debt, and dilution than Alphabet investors are used to.
GitLab at $47.06
More AI code creates more work to govern, secure, and ship.
AI can create more software and more chaos at the same time. GitLab wins if enterprises pay it to orchestrate both the humans and the agents building that software.
Robinhood at $119
The platform is broadening. The earnings remain market-sensitive.
The brokerage has become a financial platform. The earnings still move with the market's mood.
IBIT at $47.61
The wrapper is simple. The asset remains volatile.
IBIT removes the custody problem. It does not remove Bitcoin's valuation problem.
Illumina at $270
The sequencing leader needs a new growth curve.
Illumina still controls the sequencing ecosystem. The stock requires growth to return without sacrificing the margins the turnaround restored.
INTC at $122.95
The turnaround is real. The price assumes the foundry works.
Intel does not need another turnaround story. It needs evidence that manufacturing leadership can become profitable external foundry revenue.
IREN at $47
Is it a good buy at this price?
IREN has secured the power, customers and financing. Now it must prove that its AI ambitions can create value per share.
KLAC at $188.21
The tollbooth is exceptional. The stock already knows.
KLA owns one of the best tollbooths in semiconductor manufacturing. The price assumes that every harder chip will keep paying a larger toll.
Lumentum at $929
The optics boom is real. The valuation demands it be historic.
Lumentum has become one of the clearest beneficiaries of the AI optical bottleneck. The stock now requires that bottleneck to produce extraordinary revenue, margins, and per-share value for years.
Meta at $747.82
AI is lifting revenue while infrastructure consumes the cash.
Meta's AI is already making the advertising machine stronger. The problem is that building the next machine is consuming nearly all the cash the current one produces.
Marvell at $259
The AI interconnect thesis is right. The stock leaves little room to be early.
Marvell owns critical silicon across custom compute, switching, and optical connectivity. The investment question is whether that expanding role can grow quickly enough to satisfy a valuation already built around dominance.
Microsoft at $517.89
The distribution is unmatched. The economics still have to scale.
Microsoft has the enterprise distribution, cloud, and software stack to monetize AI repeatedly. The stock now asks whether that revenue can outrun the infrastructure required to deliver it.
Strategy at $158.92
Bitcoin exposure with a capital-markets engine attached.
Strategy offers Bitcoin exposure with a capital-markets engine attached. That engine can create Bitcoin per share - or senior claims faster than common equity value.
MU at $1,085.02
AI memory has changed the cycle. Has it ended it?
Micron is no longer selling ordinary memory into an ordinary cycle. The price assumes AI has changed the floor without eliminating the cycle.
Nebius at $227
Is it a good buy at this price?
Nebius may be building one of the best independent AI clouds in the world. The stock already expects it to succeed.
Netflix at $71.17
Streaming won. Disciplined compounding comes next.
Streaming won. Now Netflix must prove entertainment can keep compounding like software.
Nokia at $10.43
This is no longer just a telecom turnaround.
Nokia is combining optical networking, IP routing, mobile infrastructure, and an Nvidia partnership into a credible AI story. The harder question is how much AI economics a large telecom-equipment company can actually retain.
ServiceNow at $135.57
The AI control tower is real. The premium assumes it scales.
ServiceNow is no longer just digitizing workflows. It wants to become the control tower for AI agents across the enterprise, and the current price assumes that ambition produces premium economics.
NVTS at $12.16
The 800-volt option is real. The revenue is not yet.
Navitas may own valuable power technology for the 800-volt AI data center. Today, investors are paying for adoption that has not yet appeared in revenue.
Everpure at $126.50
Revenue is accelerating. Cash conversion must follow.
The storage company is becoming an AI data platform. The cash flow must catch the revenue acceleration.
Palo Alto at $374.60
Platform consolidation is working. The acquisition bill still counts.
Palo Alto Networks is buying and building its way toward a unified security platform. The strategy is working, but shareholders still have to count the acquisition bill and the stock issued along the way.
PLTR at $193
Is it a good buy at this price?
Palantir may be building one of the most important software platforms of the AI era. At $192.59, the stock already assumes that importance will become extraordinary per-share cash flow.
QQQ at $745.40
Exceptional businesses. Premium expectations.
QQQ owns many of the market's best businesses. It also concentrates investors in the expectations already attached to them.
Reddit at $150.03
Human conversation is scarce. The valuation knows it.
Human conversation is becoming scarcer. The stock assumes Reddit can monetize it without poisoning it.
Rocket Lab at $70
Is it a good buy at this price?
Rocket Lab is becoming far more than a rocket company. The problem is that investors are already paying for much of that future before Neutron has reached orbit and before the Iridium acquisition has closed.
Shopify at $142.25
The commerce engine is compounding. So are expectations.
The commerce operating system is compounding. The stock already prices it like the winner.
SMCI at $43.27
Revenue is surging. Cash conversion is the real test.
Supermicro is shipping AI systems at astonishing speed. The stock depends on whether revenue growth can outrun working-capital needs, financing costs, and dilution.
SNDK at $1,772.10
AI storage is booming. The price demands the new model.
Sandisk has turned AI storage demand into extraordinary earnings and long-term customer commitments. The stock now demands most of that new model before it has survived a full cycle.
Snowflake at $335.80
The data layer is accelerating. The shares paid to build it still count.
Snowflake is becoming the data layer for enterprise AI. The business is accelerating, but the stock still asks shareholders to ignore a very real bill paid in shares.
SoFi at $16.55
Is it a good buy at this price?
The business has crossed the profitability line. The stock still asks investors to underwrite credit risk, dilution, and a technology platform that has not yet reaccelerated.
SOXL at $151.75
The thesis can be right while the path is wrong.
SOXL gives you three times the daily semiconductor move. It does not give you three times the long-term semiconductor thesis.
SpaceX at $148.75
An extraordinary platform with an extraordinary burden of proof.
SpaceX has become a launch, connectivity, and AI infrastructure platform. The stock already assumes that the platform becomes one of the world's largest companies.
SPY at $772.04
The American earnings engine at a full price.
SPY owns the American earnings engine. The current price still determines how much of that compounding belongs to you.
Tempus at $84.57
The data flywheel is real. Owner earnings remain early.
Tempus is building a learning system around clinical data. The stock assumes that growth becomes durable profit before acquisitions and dilution absorb it.
Talen at $304.50
The megawatts are real. The contracts must follow.
Talen owns dispatchable generation beside the load. The stock assumes that location becomes a durable cash-flow advantage.
TQQQ at $79.85
Direction is only one of the three bets.
TQQQ magnifies the Nasdaq-100 for one day. Holding it longer turns direction, volatility, and timing into three separate bets.
Tesla at $372.59
The car company cannot justify the optionality price.
Tesla's car business does not explain the stock. Robotaxi, autonomy, energy, and robotics must do that - and several of those businesses are still being valued before they are proven.
TSM at $451.25
The world's best foundry is priced like the constraint never breaks.
TSMC is the most important manufacturer in AI. The valuation assumes its technological and geopolitical constraints remain profitable for a very long time.
Take-Two at $201.92
GTA VI is the catalyst. Recurring engagement is the value.
Grand Theft Auto VI can transform the earnings base. The stock is already waiting at the finish line.
VRT at $253.30
The thermal bottleneck is becoming an earnings machine.
Vertiv helps keep the AI factory powered and cool. The stock assumes that this physical bottleneck becomes a long-duration earnings machine.
Vistra at $138.76
The portfolio is balanced. Capital allocation must stay that way.
Vistra combines generation, retail customers, and disciplined hedging. The valuation depends on whether that machine stays balanced as AI reshapes power demand.
Workday at $189.75
Trusted workflows make AI useful. Customers still have to pay more.
Workday owns sensitive HR and finance workflows that enterprises cannot casually replace. AI can deepen that moat, but it still has to accelerate growth rather than become another included feature.
WULF at $15.73
The power is scarce. The construction bill is enormous.
TeraWulf has turned power access into long-term AI leases. Now it must build billions of dollars of infrastructure without giving away the economics.
Block at $76.63
Two ecosystems are scaling. Credit risk is scaling too.
Cash App and Square are compounding together. Credit risk is growing with them.
Zscaler at $193
The architecture fits the AI era. The valuation needs owner earnings.
Zscaler was built for a world where the internet replaced the corporate network. AI agents make that architecture more relevant and the valuation debate more demanding.
These are dated analytical scenarios, not live recommendations or personalized financial advice. Prices and estimates can change quickly. Read the linked filings, recheck current facts, and make your own decision.