Goldman Just Said What the Market Isn’t Ready to Admit About AI — But They Missed One Explosive Detail



Everyone is still chasing chips like it’s 2024.

But the real trade for 2026? It’s shifting—fast.

Goldman Sachs’ Toshiya Hari just made a call most people will dismiss:

👉 Overweight hyperscalers, underweight semiconductors.

Sounds wrong… until you understand the setup.

The Market Is Pricing AI Backwards

Chips already had their run

Philadelphia Semiconductor Index up ~150%

Trading at premium valuations

Meanwhile:

$Amazon.com(AMZN)$  

$Microsoft(MSFT)$  

$Alphabet(GOOG)$  

$Meta Platforms, Inc.(META)$  

$Oracle(ORCL)$  

…are lagging because of one thing:

👉 They’re spending too much on AI.

But that’s exactly the opportunity.

The Asymmetry Nobody Sees

👉 Hyperscalers = priced for doubt

👉 Semiconductors = priced for perfection

That’s not balance.

That’s mispricing.

Now Here’s the Part the Market Is Underestimating: ORACLE

Everyone lumps hyperscalers together.

That’s a mistake.

Oracle is playing a different game:

👉 It’s using ~5x leverage to go all-in on AI infrastructure

That’s dangerous…

But also incredibly powerful.

Why This Changes Everything

If Oracle fails to reach positive cash flow:

👉 The downside is obvious (debt + pressure)

But if Oracle flips cash flow positive:

👉 That 5x leverage becomes a return amplifier

This is not linear upside.

This is convex payoff.

Translation (Simple)

Most companies:

Spend $1 → earn $1.20

Oracle (leveraged AI bet):

Spend $1 → if it works → earn multiples of that return

My Target Framework (Not Consensus)

If Oracle successfully crosses into sustainable AI-driven cash flow:

👉 Base case: $250–$350

👉 Bull case (AI infra dominance + operating leverage): $500+

Because at that point, the market stops asking:

“Can they afford AI?”

And starts asking:

“How big can this cash machine get?”

The Real 2026 Battlefield

We are entering the harshest phase of the AI cycle:

Not models

Not GPUs

Not hype

👉 Cash flow conversion

Capex is heading toward $700B.

If that doesn’t turn into real money:

Margins get crushed

Narratives collapse

Valuations reset

Final Thought

Goldman is right about the direction.

But the real alpha is in who survives the transition.

Semis won Phase 1.

Hyperscalers might win Phase 2.

But within hyperscalers?

👉 The biggest upside may come from the one taking the biggest risk.

And right now…

That’s Oracle.

@TigerObserver  @TigerPM  @Daily_Discussion  @MillionaireTiger  @TigerStars  

# Oracle Drops 4%: Does OpenAI's Miss Threaten Cloud Contract Delivery?

Modify on 2026-05-02 08:56

Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.

Report

Comment5

  • Top
  • Latest
  • Ah_Meng
    ·05-02
    TOP
    You are absolutely correct about this… I am living in fossil past, so I am always concerned about cost. $Oracle(ORCL)$ must run fast enough before the economy collapses. I sense a transfer of assets between some of the haves and have nots. With increased borrowing everywhere, when the economy goes under, debts will ballooned into demons, AI will still go on, but the players would be different. The pain of the Dot com bust (when you were still a toddler?) still hurts for those who experienced it…
    Reply
    Report
    Fold Replies
    • Shernice軒嬣 2000
      To sustain this explosion, we need serious AI infrastructure. That’s where Oracle comes in — positioned to become a true cash-generating machine as demand for compute reaches insane levels.
      05-02
      Reply
      Report
    • Shernice軒嬣 2000
      Sharing a post  on how he is using tokens to massively boost productivity — and this group is growing fast.
      As usage scales across commercial and personal levels, computational demand is going parabolic. This isn’t just hype anymore — it’s a structural shift.
      05-02
      Reply
      Report
    • Shernice軒嬣 2000Replying toShernice軒嬣 2000
      I’ve read about the dot-com bust. AI only emerged a few years ago, and we may be at a similar early-stage inflection point—where experience from the dot-com era could help in identifying the next dominant stocks.
      05-02
      Reply
      Report
  • Ah_Meng
    ·05-02
    The excitement around AI just feels so familiar to the dot com period… before its bursting…
    Reply
    Report