Equity research, written by an analyst that never sleeps.

Drop a ticker. ResearchGPT reads the filings, transcripts, and news, then writes a cited investment memo with bull case, bear case, peers, and risks — in minutes, not days.

No spam, ever. Unsubscribe with one click.
Research request
Deep brief on NVDA — bull/bear, peers, risks
ResearchGPT working
03:41
  • Pulled latest 10-K, 4× 10-Q, 8 transcripts
  • Extracted 142 source spans · cross-referenced
  • Synthesizing bull case + counter-evidence
Memo · ready
NVIDIA Corp. · NVDA
Bull
Blackwell ramp + sovereign demand
Bear
Hyperscaler capex normalisation
14 citations·4m 23s·PDF · Excel · Notion
Built for analysts, RIAs, family offices & serious DIY investors
The flow

Three steps from ticker to memo.

01

Drop a ticker

Type any US-listed equity. Or paste a question — "How exposed is AAPL to China?" works too.

02

An agent reads everything

Filings, earnings transcripts, peer disclosures, recent news, sell-side commentary — pulled, parsed, and cross-referenced.

03

Get a cited memo

A structured memo lands in minutes. Bull, bear, peers, risks, management commentary — every claim sourced.

The deliverable

What you actually receive.

A real memo, not a chat transcript. Structured, sourced, and ready to drop into your IC pre-read or your own model.

Investment Memo
NVIDIA Corp. · NVDA
01

Executive summary

NVIDIA remains the structural beneficiary of accelerated compute, with data-center revenue compounding at +78% YoY and gross margins held at 74.6%. The thesis is now dependent on Blackwell ramp execution, hyperscaler capex durability, and competitive defense against custom silicon.[1] [2]

Bull case

  • 01Blackwell shipments tracking ahead of guidance with $11B+ in committed orders through FY27. [3][4]
  • 02Sovereign AI build-outs adding a structurally new buyer cohort outside the top-4 hyperscalers. [5]
  • 03CUDA + networking moat compresses competitor TCO advantage; switching costs remain prohibitive. [6]

Bear case

  • 01Hyperscaler capex normalization risk in FY27 as inference workloads shift to custom ASICs. [7][8]
  • 02Inventory days expanded 18 → 32 QoQ, suggesting demand-supply rebalancing. [9]
  • 03Concentration: top-4 customers represent 46% of data-center revenue. [10]
Risks flagged from latest 10-K
  • —Export controls on advanced compute to additional jurisdictions could materially impact revenue mix. [11]
  • —Customer concentration risk: a single hyperscaler represents 19% of FY revenue. [12]
  • —Supply-chain dependency on TSMC advanced-node capacity remains the binding constraint.
04

Management commentary

“Demand for Blackwell is extraordinary. We’re sold out for several quarters and are actively allocating supply.”

Jensen Huang, CEO  ·  FY Q3 earnings call [13]
05

Peer comparison

CompanyRev YoYGM %Fwd P/E
NVDA+78%74.632.1
AMD+24%51.829.6
AVGO+44%63.230.8
INTC−2%40.122.4

Figures derived from latest 10-Q filings and consensus estimates. [14]

14 numbered citations·Every claim links to source span·Exportable to PDF, Excel, Notion
What it does

Everything an analyst would, sourced.

Citations on every claim

Every sentence links to a source span — page in a 10-K, a quote from the earnings call, a peer disclosure. No hallucinated numbers.

Bull case + bear case

Both sides argued at full strength, with counter-evidence surfaced from the same corpus. Disagree with confidence.

Peer comparison, automatic

Top 3–5 peers selected by segment overlap, normalised on revenue growth, gross margin, and forward multiples.

Earnings-call signals

Tone shifts, hedged language, guidance changes, and Q&A pressure points — surfaced and quoted, not just summarised.

10-K risk extraction

We read the risk factors so you don't have to skim 70 pages — distilled, deduplicated, and ranked by recency vs. last cycle.

Export anywhere

PDF for the IC pre-read, Excel for the model, Notion for the running file. Or wire it to your stack via API.

Who it's for

For anyone who reads filings for a living — or for fun.

Independent analysts

Cover more names without the Bloomberg invoice.

RIAs & small funds

Get IC-ready memos without adding headcount.

Family offices

Diligence external recommendations in minutes.

Sophisticated retail

Research like a junior analyst would.

Questions

Things people ask before signing up.

Anything else? Reach us at hello@researchgpt.com.

  • Where does the data come from?

    Primary sources only — SEC filings (10-K, 10-Q, 8-K, S-1, proxy), official earnings transcripts, company IR materials, and reputable news. We log every source span so you can audit the trail.

  • How do you handle hallucinations?

    Every claim is grounded to a verbatim source span and a citation marker. If a claim can't be sourced, the agent declines to make it. You'll see the gaps explicitly — no fabrication.

  • What will it cost?

    Beta users will lock in early-supporter pricing for life. We're targeting a price point that sits between Perplexity Pro and AlphaSense — designed for individuals and small teams, not enterprise procurement.

  • Is my research private?

    Yes. Your tickers, queries, and downloaded memos are tied to your account and never used to train models. We'll publish a clear data-handling policy at GA.

  • When does the beta open?

    We're rolling out access in waves through 2026. Earlier requests get earlier seats — and a credit on your first month.

Be among the first to try it.

Drop your email. We’ll send you a seat the moment your wave opens — and a credit on your first month.

No spam, ever. Unsubscribe with one click.