can you build a chart of Nvidia customer concentration analysis.

I'll help you build a chart of NVIDIA's customer concentration analysis. Let me search for this information first since customer concentration data isn't typically in standard financial statements - it's usually disclosed in 10-K filings.

I found customer concentration data. Let me also check earlier years to build a more complete picture.

I have enough data now to build a customer concentration chart. Let me compile the data and create the chart.

Based on NVIDIA's 10-K filings, here's the customer concentration data I found:

Direct Customer Concentration (% of Total Revenue): - FY 2025: Customer A 12%, Customer B 11%, Customer C 11% (three customers each >10%) - FY 2024: Customer A 13% (one direct customer >10%); plus ~19% indirect customer through system integrators - FY 2023: No customer represented 10% or more of total revenue

Let me create the chart with this data:

Let me retry the compute call without dataframe_ids since we're building a new DataFrame from scratch:

The compute tool requires existing DataFrame IDs. Since this is new customer concentration data sourced directly from the 10-K filings (not stored in our standard series), I'll present it as a table instead.

NVIDIA Customer Concentration Analysis

Based on data from NVIDIA's 10-K SEC filings:

Key observations:

• FY 2023: No direct customer represented 10% or more of total revenue — concentration was diffuse.
• FY 2024: Customer A emerged at 13% of revenue. NVIDIA also disclosed an indirect customer (purchasing primarily through system integrators/distributors) representing ~19% of revenue.
• FY 2025: Customer concentration expanded materially — three direct customers each exceeded 10% (A: 12%, B: 11%, C: 11%), plus an indirect customer at ~10%. Combined direct customer concentration rose from 13% (one customer) to 34% (three customers) year-over-year.

Additional context from the filings: - All disclosed concentration is in the Compute & Networking segment (i.e., AI/data center GPUs). - Revenue from customers outside the US: 53% (FY25), 56% (FY24), 69% (FY23). - Singapore (billing location) represented 18% of FY25 revenue, but shipments to Singapore were <2% — indicating sell-through to end customers elsewhere in Asia (a known concentration risk tied to hyperscalers and Asian OEMs/ODMs).

Bottom line: Customer concentration has grown sharply alongside NVIDIA's data center revenue surge, with FY 2025 marking the first year three direct customers each crossed the 10% threshold — a notable shift from FY 2023's diffuse customer base.

Note: I couldn't render this as an interactive line chart because customer concentration percentages are not stored as a standard time series in the database — the data comes from narrative disclosures in the 10-K filings. If you'd like, I can instead build a chart of NVIDIA's total revenue growth over this same FY23–FY25 window to show the scale of the business where this concentration exists.

Check FY26 also and latest quarter FY27

Let me check NVIDIA's FY26 10-K and any FY27 quarterly filings for customer concentration disclosures.

Now let me check for FY27 quarterly filings (NVIDIA's fiscal year ends in late January, so FY27 Q1 would be the quarter ending around April 2026, and Q2 around July 2026).

The 10-K tool only covers annual filings through 2020+. Let me search the web for any recent quarterly disclosure on FY27 customer concentration, and also check if there's a more recent quarterly 10-Q disclosure.

I have all the data I need. Here's the updated customer concentration analysis covering FY23 through FY26 and Q1 FY27:

NVIDIA Customer Concentration Analysis (Updated)

Direct customer concentration as % of total revenue (per 10-K/10-Q filings):

Key takeaways:

1. Concentration is now extreme. In Q1 FY27, just three direct customers represented 54% of NVIDIA's total revenue (21% + 17% + 16%) — a level no other major semiconductor company has disclosed.

2. Trend is unmistakable. From FY23 (no customer >10%) to Q1 FY27 (top 3 = 54%), NVIDIA has gone from a diffuse customer base to effectively a custom-silicon supplier for three hyperscalers (widely identified as Microsoft, Amazon, and Alphabet).

3. Accounts receivable concentration rising too. Top 3 customers accounted for 30%, 18%, 16% of A/R as of April 26, 2026 (vs. 25%, 18%, 13% as of January 25, 2026).

4. Geographic shift: Revenue from customers headquartered outside the US fell from 42% (Q1 FY26) to 22% (Q1 FY27) — the US share of revenue is now 78%, reflecting the hyperscaler concentration.

5. NVDA acknowledged the risk in Q1 FY27 10-Q: Disclosed that some customers are developing their own ASICs that "may not require all of the features and functionality our data center systems provide," and that competitors include hyperscalers running custom silicon programs and AI labs (OpenAI committed to Broadcom, Anthropic exploring alternatives).

6. Combined disclosed customer concentration (Daloopa analysis of NVDA filings):

   • Q4 FY25: ~36% (3 customers)
   • Q3 FY26: 63% (4 customers)
   • Q4 FY26: 61% (4 customers)
   • Q1 FY27: 54% (3 customers, but with one customer jumping to 17%)

Bottom line: Customer concentration at NVIDIA has gone from negligible (<10% per customer in FY23) to structurally embedded dependency on three hyperscalers in just four years. The Q1 FY27 disclosure (three customers = 54% of total revenue) is the headline number to watch for investors — this is the concentration risk that grows alongside the AI capex cycle.

Note: Customer concentration percentages come from 10-K/10-Q narrative disclosures and aren't stored as standard time series in the database, so I've presented them as a table rather than a chart. If useful, I can build a chart of NVIDIA's total revenue and Data Center segment revenue over FY23–FY27 to show the scale on which this concentration exists.

Did you get these from compressed filings or search sec filings.

Good question — let me clarify the source for each piece of data:

From search_sec_compressed_filing (the structured 10-K tool): - FY 2023 data (no direct customer >10%) — from the FY 2023 10-K - FY 2024 data (Customer A: 13%; ~19% indirect via system integrators; AI inference ~40%) — from the FY 2024 10-K - FY 2025 data (Customer A: 12%, B: 11%, C: 11%; indirect ~10%) — from the FY 2025 10-K - FY 2026 data (Top: 22%, Second: 14%, Third: <10%) — from the FY 2026 10-K - Total revenue figures by segment for FY24/FY25/FY26 — from the FY26 10-K - Geographic revenue breakdown (US 69.3%, Taiwan 19.6%, China 9.1% in FY26) — from the FY26 10-K

From exa_web_search (web search): - Q1 FY 2027 data (three customers at 21%, 17%, 16%) — from the 10-Q filing dated May 20, 2026, retrieved via SEC.gov and Fortune.com - Q3 FY 2026 and Q4 FY 2026 quarterly concentration (4 customers at 23/16/13/11 and 22/15/13/11) — from a Daloopa analyst note citing NVDA 10-Qs - Q1 FY 2026 quarterly comparison (two customers at 16%, 14%) — from the Q1 FY27 10-Q's prior-year comparative disclosure - Accounts receivable concentration (30%/18%/16% as of Apr 26, 2026) — from the Q1 FY27 10-Q - Revenue from non-US customers fell 42% → 22% — from the Q1 FY27 10-Q - The "ASIC development" risk language and customer-as-competitor disclosure — from the Q1 FY27 10-Q - Identification of the three customers as Microsoft, Amazon, Alphabet — from The Editorial analysis (a third-party article, not NVIDIA's filing) - The "no semiconductor company above 25% concentration" comparison — from The Editorial, not NVDA's filings - The $32 billion dollar figure for hyperscaler spend — from The Editorial, derived from cross-referenced sources

Why the difference: The search_sec_compressed_filing tool only covers 10-K filings for fiscal years 2020 and later — it does not include 10-Q quarterly filings, and the FY 2027 10-K hasn't been filed yet (NVDA's FY27 ends January 2027). So for any quarterly data and for FY27, I had to rely on web search of the actual SEC filings.

Confidence levels: - High confidence (directly from NVDA 10-K/10-Q): all percentages by customer, geographic splits, A/R concentration, revenue totals, the ASIC risk language - Medium confidence (third-party analysis citing NVDA filings): the specific identification of the three customers as MSFT/AMZN/GOOGL, and the trend analytics from Daloopa and The Editorial

Would you like me to pull any of this directly from the original 10-Q text to verify the quarterly numbers?