Dan Ives — How To Make Millions in AI Stocks Investment Thesis

Source: How To Make Millions Investing In AI Stocks | Dan Ives, From The Desk Of Anthony Pompliano, May 23, 2026.

The Framework: Third Inning of a 10-Year AI Supercycle

Ives' organizing lens: AI is year three of a 10-year buildout — "third inning, one out" in a nine-inning game. The trade is not one or two names; it is chips at the epicenter, then memory, infrastructure, software, cybersecurity, and second/third/fourth derivatives (utilities, space, physical AI). Asia fab trips confirm demand; spreadsheet bears in New York miss it.

Layer	Ives' View	How to Play
Chips	Nvidia fuels the revolution; supply-constrained puzzle for Meta/others	NVDA + memory (MU, Sandisk) + alternatives (AMD, INTC)
Infrastructure	Data centers are "hearts and lungs of AI" — without them, nothing works	Hyperscaler capex + colocation/power names
Mega-cap software	Distribution wins; MSFT and AAPL most mispriced large caps	MSFT ~$550–600 fair value vs ~$400
Space	New sector — 80% of SpaceX IPO is future orbital compute	SpaceX, Planet Labs, Rocket Lab, Voyager
Physical AI	Autonomous driving / robotaxis = first true consumer AI touchpoint	Tesla robotaxis + federal regulation catalyst

"We're in year three of a 10-year buildout of AI... If I see the demand in Asia, that's what makes me bullish."

Investment Thesis #1: SpaceX — Spatial Revolution, 80% Future, Tesla Merger by 2027

Argument: SpaceX defines a new sector — not space tourism, but data centers in space converging tech and a "spatial revolution." 80% of IPO value is future, not today's business; at $1.5–2 trillion Ives would not buy on current fundamentals alone — same as buying Intel on the US stake was a bet on future AI relevance, or Nvidia on enterprise before physical AI chips.

SpaceX is becoming a hyperscaler (Colossus, Anthropic/Google deals) with 80–85% probability of merging with Tesla by 2027 — converging data ("data is the new oil and gold") to narrow the model gap vs Anthropic/OpenAI. Vertical stack: cheap launch → orbital compute powered by Tesla solar → Starlink downlink → Optimus/robotaxis at the consumer edge. Tech giants cannot get to space without SpaceX-class launch capacity.

Contrarian element: Market treats SpaceX as a launch company; Ives views it as second/third/fourth derivative of the AI revolution, not a separate theme.

Trigger: Starship rapid reusability; orbital data-center economics proven; Tesla–SpaceX merger signals; Google/tech partnership depth.

Names: SpaceX (PRIVATE — IPO), Tesla (TSLA), Alphabet (GOOG) as compute/orbital partner.

Investment Thesis #2: Chips and Memory — Asia Demand Confirms the Supercycle

Argument: Everything starts at chips — Nvidia at the epicenter. Meta needs 10 chips; Nvidia supplies ~3; the rest comes from Google TPUs, AMD, Intel, Micron, Sandisk, and the full component puzzle. Ives' Asia trips (Taiwan fabs) show production, supply, and component bottlenecks firsthand — why he stays bullish through earnings season while Metro-North bears rely on narratives.

Street underestimates growth even when multiples look high — a memory supercycle. Micron up 600–800% and Marvell ~100% in six months still frighten investors; Ives cites March Iran/oil scare when Nvidia sold off — missing geopolitical dips means missing the multi-year move. Roundhill's DRAM ETF added ~$10B in two months — democratization of global memory exposure.

"Meta, for every 10 chips they need, Nvidia's going to give them three because of supply. Where do they get the others?"

Contrarian element: "Musical chairs" fear after huge runs; Ives: don't anchor to 6-month performance — triangulate Asia supply with hyperscaler budgets.

Trigger: Memory pricing and HBM guides; Taiwan fab utilization; hyperscaler capex sustaining through Iran/oil volatility.

Names: NVIDIA (NVDA), Micron (MU), Sandisk (SNDK), Marvell (MRVL), AMD, Intel (INTC).

Investment Thesis #3: Mispriced Mega-Caps — Microsoft and Apple Are the Large-Cap Anomalies

Argument: Microsoft runs every enterprise on Earth; Nadella is among the best CEOs. Copilot underwhelmed and OpenAI "training wheels" create noise — but at ~$400, Ives values MSFT at ~$550–600 (Ackman dislocation trade). Apple is "behind" on AI but owns 1.5B iPhones / 2.5B iOS devices — a toll collector on the consumer highway; ~20% of the world will access AI through Apple. WWDC Gemini launch is the strategy inflection; new CEO Turnus doubles down on services.

Google was "done" a year ago (DOJ, search death, Gemini nowhere) — now "victory parties." Narratives create opportunities for those who do the work.

Contrarian element: Consensus says mega-caps already won AI; Ives says MSFT and AAPL are the most mispriced large caps while semis rip.

Trigger: WWDC AI/Gemini integration; Microsoft Azure/Copilot enterprise monetization metrics; Apple services AI attach rate.

Names: Microsoft (MSFT), Apple (AAPL), Alphabet (GOOG).

Investment Thesis #4: Cybersecurity — Anthropic Won't "Eat" CrowdStrike or Palo Alto

Argument: At RSA Conference (March), the narrative was Anthropic/Mythos will eat cybersecurity and software. Ives talked to customers — wrong. CrowdStrike and Palo Alto stocks recovered sharply vs March lows. Agentic AI increases attack surface; cybersecurity spend rises with infrastructure budgets.

Same pattern as NYC cab drivers bearish on Alphabet — narrative vs ground truth.

Contrarian element: Model-layer fear trades sell CRWD/PANW; Ives sees mispriced security in an agentic world.

Trigger: Enterprise security budget surveys; CRWD/PANW earnings vs agentic adoption; Mythos cyber headlines without share loss.

Names: CrowdStrike (CRWD), Palo Alto Networks (PANW).

Investment Thesis #5: Physical AI and Utilities — Robotaxis First, Power as AI Derivative

Argument: Average consumers played with ChatGPT, but physical AI starts with autonomy — Tesla robotaxis as the first true AI interaction (Waymo pockets excepted). Federal preemption (Trump executive order) is the missing piece vs China's state-level lead. Separately, utility companies — long ignored — are re-rating as AI derivatives as power demand hits the grid.

Ives: US ahead of China on software/chips (Jensen/Nvidia; third-rate Nvidia H200 ~1.5–2 years ahead of Huawei); China leads robotics and nuclear power — but US wins models + silicon. Pro-innovation administration (Jensen/Cook/Musk China trip) vs Europe "building Blockbuster videos" under regulation.

Contrarian element: Investors chase model labs; Ives emphasizes electron path — power, robotaxis, and regulated autonomy before humanoids scale.

Trigger: Federal robotaxi framework; Tesla autonomy milestones; utility multiple expansion on data-center power offtake.

Names: Tesla (TSLA), GE Vernova (GEV) and utility AI-derivative names.

Investment Thesis #6: SaaS Disconnect — Fears Overstated, Winners Still Emerge

Argument: "SaaS apocalypse" hit ServiceNow, Salesforce, Oracle, Workday — Ives believes much of the selloff is disconnected from ultimate outcomes. It was a wake-up call for Benioff and legacy SaaS, but replacing Wall Street or corporate America stacks is easier said than done. As models commoditize, differentiation returns to people, engineering, and product — not the LLM alone.

70% of enterprise AI use cases are monetization and revenue enhancement, not cost-cutting; only ~20% of CEOs firing "because of AI" is real — ~80% is over-hiring cleanup. Ives' biggest fear is not SaaS — it is tech CEOs tripping on PR (Dario 20–30% job loss, Mustafa white-collar wipeout rhetoric) blocking data-center approvals and inviting regulation.

"Dario... 100% wrong... You start that type of fear, then all of a sudden, data centers don't get built."

Contrarian element: Buy quality SaaS on narrative dislocation while avoiding names with no adaptation path — not a blanket short.

Trigger: CRM/NOW/ORCL AI product revenue; data-center approval pace in US towns; enterprise AI spend mix (revenue vs cost).

Names: Salesforce (CRM), ServiceNow (NOW), Oracle (ORCL), Palantir (PLTR).

The Ecosystem Map

Ives AI 30: Wedbush research basket — 30 names across chips, software, infrastructure, cybersecurity; refreshed quarterly
Space derivatives: Planet Labs, Rocket Lab, Voyager (named alongside SpaceX)
US vs China: US wins chips/software/models; China wins robotics + nuclear power; "I need you, you need me" at Trump–Xi level
Bitcoin vs AI: Capital pendulum — AI is a once-in-100-year cycle diverting risk assets; crypto "not going anywhere" but rotates
FinTech automation: Figure Technologies (private) — HELOC/securitization automation; "software eating finance"
Macro overlay: Warsh Fed chair — volatile first 3–6 months but Ives can't believe he arrives as a hawk; don't lose sight of tech trade through oil/Iran noise
Enterprise AI data (Pomp): Sylvia portfolio AI — more usage correlates with faster net-worth growth (host cited, Ives agrees on monetization skew)

Key Risks

SpaceX valuation on today alone: Would not buy at $1.5–2T on current earnings — 80% future; low-probability full vision
Tech PR hubris: Job-cut rhetoric from Dario, Mustafa, others → protests, midterm politics, data centers not approved — Ives' #1 fear
Dario job-loss forecasts: Ives says 100% wrong — but if belief persists, regulatory drag slows buildout
Fed/Warsh volatility: New chair first 3–6 months historically rough; oil/Iran adds macro noise
SaaS disruption: Wake-up call is real even if overstated — wrong picks in software lose to AI-native entrants
Bitcoin capital diversion: Dollar allocation pendulum favors AI; wrong AI stock → regret vs crypto
Geopolitical dips: Iran/oil scares (March Nvidia) repeat — investors who tune out macro miss entries
Europe/regulatory lag: Blockbuster-risk if US copies European privacy/regulatory posture

Investment Opportunities at a Glance

Tier	Name / Category	Core Thesis	Conviction Signal
1	NVIDIA (NVDA)	Single chip fueling AI revolution; H200 1.5–2yr ahead of Huawei	"Godfather Jensen Nvidia" at epicenter
1	Micron (MU)	Memory supercycle; Asia fab demand; still cheap after 600–800% run	Triangulated Asia production vs street growth
2	Microsoft (MSFT)	Every enterprise on Azure; worth ~$550–600 vs ~$400	Ackman dislocation; Nadella + OpenAI partnership
2	Apple (AAPL)	Sleeping giant; 20% of world accesses AI via iOS; WWDC Gemini	"Toll collector" on consumer AI highway
2	Alphabet (GOOG)	Narrative flip from "done" to victory; Gemini + search resilience	NYC cab-driver bear case was wrong
2	Tesla (TSLA)	Robotaxis first consumer physical AI; 80–85% SpaceX merge by 2027	Federal autonomy regulation catalyst
2	Marvell (MRVL)	Component/memory puzzle beneficiary	Named alongside Micron shortage trade
3	CrowdStrike (CRWD)	Agentic AI increases security spend; Mythos fear overdone	Stock recovery post-RSA vs March lows
3	Palo Alto Networks (PANW)	Same cybersecurity contrarian as CRWD	Customer checks contradict "Anthropic eats security"
3	Sandisk (SNDK)	Memory/components in supply puzzle	Named with Micron in Meta chip shortfall
3	Intel (INTC) / AMD	Meta's remaining chip supply after Nvidia allocation	"Where do they get the others?"
3	GE Vernova (GEV)	Utilities re-rating as AI power derivatives	Multiple expansion on data-center demand
3	Palantir (PLTR)	Named in US AI stack alongside hyperscalers	US software/chip leadership thesis
3	Salesforce (CRM)	SaaS fear disconnected; Benioff wake-up call	Models commoditize → people/product win
4	SpaceX (PRIVATE)	Spatial revolution; orbital compute; 80% future value	Largest IPO; hyperscaler deals in 30–45 days
4	Rocket Lab (RKLB) / Planet Labs (PL)	Space-sector derivatives off AI revolution	Named as sector starters beyond SpaceX
4	Bitcoin (BTC)	Pendulum asset vs AI; not going away	Capital allocation tension with 100-year AI cycle

Monitoring Checklist

Bottom Line

You are in inning three of a 10-year AI buildout — missing geopolitical scares (March Nvidia/Iran) means missing the supercycle; Asia demand beats spreadsheet narratives.
SpaceX is 80% future and defines a new spatial-AI sector — but Ives would not pay on today's math alone; the Tesla merge + data convergence thesis is the 2027 catalyst to watch.
Memory and chips remain the epicenter — Micron's 600–800% move can still be "cheap" if the street underestimates exponential demand; don't fear the six-month chart.
MSFT and AAPL are the mispriced mega-caps — while Nvidia gets all attention; Google proves narrative reversals happen fast.
The biggest risk is not technology — it is tech's own PR — job-loss doom talk kills data-center approvals; Ives calls Dario 100% wrong on mass unemployment, but the politics are real.