if you missed palantir at $7... it’s happening again

Summary

Antonio Linares argues that the AI investment opportunity may be entering a second growth phase, with Palantir (PLTR) as his preferred example. He says he bought Palantir at $7 per share in 2022 and compares its current position to Nvidia (NVDA) in 2016.

His central investment measure is free cash flow per share (FCF/share). He believes stock prices ultimately track it, and that AI can drive durable FCF growth when it helps solve customer problems in ways competitors cannot easily replicate.

Linares rejects a simple comparison with the dot-com bubble. In his view, AI is already improving the core businesses of large companies, though he acknowledges that financial setbacks and sharp stock-price declines remain possible.

Investment Framework: “Ontological Velocity”

Linares’s framework centers on a reinforcing loop:

• A company has substantial proprietary data, customer engagement, or control of important workflows.
• It applies AI to that data to improve customer outcomes or internal efficiency.
• Better outcomes attract or retain more users, generating more data and further improving the platform.
• The resulting “ontology”—his term for a data-rich digital representation of a business or market—becomes harder to replicate.

He distinguishes between:

• AI builders, which invest heavily in computing infrastructure and bear significant capital-spending risk.
• AI “riders” or beneficiaries, which own data and workflows, use or rent models and compute, and may pass costs on to customers.

He prefers the latter layer, particularly businesses that can benefit from better models without having to fund all the infrastructure themselves. A potential thesis check for Palantir, he says, is whether customer retention and FCF/share continue to strengthen.

Company and Market Claims

Amazon, Meta, and Netflix

Linares says AI is improving the core value-creation processes of Amazon (AMZN), Meta (META), and Netflix (NFLX): shopping conversion at Amazon, engagement and ad performance at Meta, and content production at Netflix.

• Amazon: Its Rufus shopping assistant is associated with a claimed 40% increase in conversions when shoppers use Alexa to shop.
• Meta: The presenter says its apps have 3.6 billion daily users. He cites a 15.7% increase in Facebook conversions from AI-based ranking, though the subtitle’s wording on this figure is unclear.
• Netflix: AI has reportedly reduced production time and costs by 50% in some AI-assisted original productions.

Capital Spending and AWS

Amazon and Meta’s combined capex is said to have risen from $75 billion to $262 billion, while their combined FCF/share fell from $24 to $14.90. Linares views this as near-term pressure from investment, not necessarily evidence that the spending is wasted.

For Amazon Web Services (AWS), the speaker cites $169 billion in annual revenue, up 36.7% year over year, and a $496 billion backlog. He says AWS could reach $1 trillion in annual revenue. He also mentions more than $25 billion in annual chip revenue and Amazon’s potential to sell robotic-logistics capabilities to outside businesses.

Meta and Netflix

• Meta: Advantage+ is said to generate more than $75 billion in annual revenue, and over one million companies use Meta agents weekly. Linares argues Meta can deploy its compute for its own services or monetize it through offerings to businesses.
• Netflix: Recent capex is described as just under $1 billion, while FCF/share has risen about 70% since 2023.

Palantir

Linares calls Palantir a “singularity scaler.” He argues that customers can plug in improved commercial or open-weight models, making Palantir’s platform more capable without the company having to create every model.

He cites:

• Net dollar retention of 157%, up 700 basis points in one quarter.
• Quarterly FCF/share rising from $0.06 in September 2023 to $0.47 in June 2026.
• A stock-price increase of about threefold since an earlier forecast, making the original “100x growth potential” estimate less applicable.

Linares believes AI agents could expand Palantir from a productivity platform into something that effectively supplies digital labor. He presents this as a potentially very large opportunity, including the possibility of a trillion-dollar company. This is a speculative forecast, not a reported financial result.

Open-Weight Models and AMD

The presenter says an NVIDIA open model surpassed frontier models on five production tasks in 24 hours. He argues that capable open models could reduce customer dependence on leading AI labs such as OpenAI and Anthropic.

For AMD (AMD), Linares says he invested at $4.20 per share, made a bullish case when shares were at $75 in 2025, and still expects further growth. He describes AMD as a potential major AI company.

Other Companies Mentioned

Linares also identifies Spotify (SPOT), Duolingo (DUOL), and Hims & Hers (HIMS) as potential beneficiaries of proprietary data and AI.

• Spotify: Described as having an “ontology of tastes.” The presenter cites 777 million monthly active users, 300 million paying users, and revenue per employee doubling over three years with roughly the same staff.
• Duolingo: Described as having data on learning and user motivation. Linares says retention is at an all-time high and predicts daily active users could rise two to three times over the next 1.5–2 years. A separate subtitle reference to a 1% year-over-year retention change is unclear.
• Hims & Hers: Described as a consumer telemedicine platform with a vertically integrated peptide supply chain and a recent launch in Australia. The presenter says an AI doctor rollout in weight loss tripled engagement and messaging, and that AI resolves 80% of patient questions. The subtitle’s claim about projected revenue—“100 million three quarters after launch”—is unclear.

Risks, Cautions, and Disclosures

• Linares acknowledges that AI-related stocks could experience severe downturns, including hypothetical 80% or greater declines, and that an AI “winter” cannot be ruled out.
• He identifies falling Palantir net dollar retention and stagnating or declining FCF/share as evidence that could invalidate his thesis. He cautions against judging the thesis solely from one quarter of lower FCF if the decline reflects increased investment.
• He recognizes that heavy capex creates execution and financial risks, even while arguing that Amazon and Meta’s spending may build valuable platforms.
• He says nothing in the video is financial advice and that it is for educational and entertainment purposes only.

Tickers, assets, and sectors mentioned: PLTR, NVDA, AMZN, META, NFLX, AMD, SPOT, DUOL, HIMS; AI, cloud computing, semiconductors, advertising, e-commerce, streaming, logistics/robotics, and telemedicine. AWS, OpenAI, and Anthropic are also discussed.

Presenter/source: Antonio Linares, speaking on his investment channel or podcast. He also refers to his prior videos and social media posts on X.