Eleven employers with 2026 layoffs hold 7.7% of open US AI-titled postings
Open US postings on the Metix AI Platform on October 27, 2026 at eleven employers with publicly reported 2026 layoffs, and how many have an AI role in the title.
7.7%
Of 41,037 open US postings with an AI role in the title, 3,174 are at the eleven employers with 2026 layoff reports. The eleven hold at most 1.6% of all open US postings.
Among the eleven, the AI-titled share runs from 1.1% to 15.5%
Share of each employer's open US postings with an AI role in the title, bars scaled to the longest. Grey is the comparison employer, NVIDIA
- NVIDIA424 of 2,21819.1%
- Google524 of 3,39015.5%
- Microsoft154 of 1,13513.6%
- Meta281 of 2,30812.2%
- Salesforce79 of 72710.9%
- Dell24 of 22910.5%
- Cisco92 of 9799.4%
- IBM92 of 9989.2%
- Amazon1,709 of 22,0597.7%
- Oracle112 of 2,4164.6%
- PayPal18 of 4444.1%
- Intuit89 of 8,4251.1%
What it shows
The eleven have 43,110 open US postings, and 3,174 of them have an AI role in the title, 7.4%.
Those 3,174 are 7.7% of the 41,037 such postings in the US, while the eleven hold at most 1.6% of all open US postings. Among AI-titled postings they are at least 4.8 times over-represented.
Method and limits
The US total of open postings is above the Platform's banding threshold, so the market uses a floor (2,690,525) and every share of it is an upper bound.
Source: Metix AI Platform, jobs, 2026-10-27.
Queriesus-open.json · ai-title.json
{ "where": { "all": [ { "field": "location.country", "eq": "United States" }, { "field": "is_open", "eq": true } ] }, "size": 1}
{ "where": { "any": [ { "field": "title", "match": "ai" }, { "field": "title", "match": "ml" }, { "field": "title", "match": "machine learning" }, { "field": "title", "match": "artificial intelligence" }, { "field": "title", "match": "llm" }, { "field": "title", "match": "deep learning" }, { "field": "title", "match": "applied scientist" }, { "field": "title", "match": "nlp" }, { "field": "title", "match": "computer vision" } ] }, "size": 1}
Amazon alone has 1,709, 54% of the eleven
Open US postings with an AI role in the title at each employer; the outline behind each bar is the employer's whole open count
Amazon 1,709, Google 524, Meta 281, Microsoft 154, Oracle 112, Cisco 92, IBM 92, Intuit 89, Salesforce 79, Dell 24, PayPal 18
- Amazon53.8%1,709
- Google16.5%524
- Meta8.9%281
- Microsoft4.9%154
- Oracle3.5%112
- Cisco2.9%92
- IBM2.9%92
- Intuit2.8%89
- Salesforce2.5%79
- Dell0.8%24
- PayPal0.6%18
Source: Metix AI Platform, jobs, 2026-10-27.
Run it
Three ways in. Each says what it costs before you start.
1 API Credit buys 25 search results or 5 full records; $1 buys 30 API Credits.
Run it in your agent
150 to 170 API Credits$5.00 to $5.67More than the 100 free API Credits
Your agent stops and asks before spending more than 200 API Credits.
Your agent follows the prompt step by step: it reads the rules, runs the counts, checks the definitions the prompt asks it to check, and writes the files and charts. Use an agent that can write files, such as Claude Code or Codex.
Set up onceKey, connection, and a free check. Skip this if your agent already reaches the Platform.
1Get a key
Create a key on the Metix AI Platform →New accounts get 100 API Credits once, valid for 30 days. Set the key in the shell you start your agent from, or add the line to ~/.zshrc or ~/.bashrc so every new terminal has it:
Shellexport METIX_KEY=metix_xxxxxxxx
2Connect your agent
Claude Code
Registers the Platform for every project. Start claude in any folder and the twelve metix tools are there.
MCP setup guide →Shell: "${METIX_KEY:?run step 1 first}" && claude mcp add --scope user --transport http metix-ai-platform \ https://mira-api.metix.ai/mcp \ --header "Authorization: Bearer $METIX_KEY"Codex
Registers the same server. The key stays in your environment instead of the config file.
MCP setup guide →Shellcodex mcp add metix-ai-platform \ --url https://mira-api.metix.ai/mcp \ --bearer-token-env-var METIX_KEY
Skills
Four skills that teach any agent the Platform's endpoints and query rules, for agents without MCP. The installer starts with none ticked: press space on each, then enter.
Skills install guide →Shellnpx skills add MetixAI-Official/metix-skills
Other MCP
Point the client at this endpoint over streamable HTTP with both headers; without the Accept header the server answers 406. Older clients use /sse on the same host.
MCP setup guide →Endpoint and headershttps://mira-api.metix.ai/mcp Authorization: Bearer <your key> Accept: application/json, text/event-stream
3Check the setup
Ask this first. It reads your balance and the field list, runs no search, and costs nothing:
Prompt for your agentUse the Metix AI Platform to check my key status and read the contract; both are free. Then tell me my API Credit balance and which datasets I can query. Do not run any search.
4Paste the prompt
Start your agent in an empty folder, then paste. It writes its files there.
The question
Answer one question with the Metix AI Platform MCP: at employers with publicly reported 2026 layoffs, how many open US postings have an AI role in the title, and how does that compare with the whole US market? Work only through the public Platform (REST at https://mira-api.metix.ai, the MCP server, or the metix-skills) with the key in METIX_KEY, and never print the key.Answer one question with the metix-skills: at employers with publicly reported 2026 layoffs, how many open US postings have an AI role in the title, and how does that compare with the whole US market? Work only through the public Platform (REST at https://mira-api.metix.ai, the MCP server, or the metix-skills) with the key in METIX_KEY, and never print the key.Answer one question with the Metix AI Platform: at employers with publicly reported 2026 layoffs, how many open US postings have an AI role in the title, and how does that compare with the whole US market? Work only through the public Platform (REST at https://mira-api.metix.ai, the MCP server, or the metix-skills) with the key in METIX_KEY, and never print the key.
01Read before querying
Call GET /contract (free) and use only querySpecByEntity.job fields. A count with size 1 costs 1 API Credit. Check the balance with GET /auth/key/status (free) at the start and at the end, and stop and ask before the run passes 200 API Credits.
02Employers
Use the list in employers.json: eleven employers with a publicly reported 2026 layoff, and two comparison employers. Match on company.name with operator in, using the names as the Platform stores them. Count Amazon as the union of its two stored names so no posting is counted twice.
03Population
Open postings: location.country eq "United States" and is_open eq true.
04AI-titled
A posting is AI-titled when its title matches any of: ai, ml, machine learning, artificial intelligence, llm, deep learning, applied scientist, nlp, computer vision. Read about 15 titles for one employer and check that they are AI roles.
05Count
For each employer: open postings, AI-titled postings, and six title families (engineering, sales, support, people, marketing, finance) as in the queries folder. For the market: the AI-titled count for the whole US.
06Mind the band
A count above the threshold comes back as a string such as "100000+". Do not add it as a number. Count the US by state (the 50 states, DC and Puerto Rico), add the exact state counts, and add the threshold for each state that is itself banded. Publish that as a floor, and say that every share of the market is therefore an upper bound.
07Do not use posted_date as an age, and do not count descriptions that name AI
Both were tried and carry no signal here (postings are refreshed, and descriptions repeat boilerplate).
08Outputs
Write data/employer-postings.json with "unit": "jobs", the snapshot date, one row per employer and one market row, each count under a key ending in _count. Report the balance from GET /auth/key/status before and after as the cost.
09Chart
One bar per employer for the AI-titled share of its open postings, sorted, with the comparison employer drawn in a different color. Title it with the finding.
10Limits
Postings are demand, not headcount, and the card does not show that any role replaced a laid-off one. Postings repeated across locations count more than once. One day, not a trend.
1. Read before querying. Call GET /contract (free) and use only querySpecByEntity.job fields. A count with size 1 costs 1 API Credit. Check the balance with GET /auth/key/status (free) at the start and at the end, and stop and ask before the run passes 200 API Credits.
2. Employers. Use the list in employers.json: eleven employers with a publicly reported 2026 layoff, and two comparison employers. Match on company.name with operator in, using the names as the Platform stores them. Count Amazon as the union of its two stored names so no posting is counted twice.
3. Population. Open postings: location.country eq "United States" and is_open eq true.
4. AI-titled. A posting is AI-titled when its title matches any of: ai, ml, machine learning, artificial intelligence, llm, deep learning, applied scientist, nlp, computer vision. Read about 15 titles for one employer and check that they are AI roles.
5. Count. For each employer: open postings, AI-titled postings, and six title families (engineering, sales, support, people, marketing, finance) as in the queries folder. For the market: the AI-titled count for the whole US.
6. Mind the band. A count above the threshold comes back as a string such as "100000+". Do not add it as a number. Count the US by state (the 50 states, DC and Puerto Rico), add the exact state counts, and add the threshold for each state that is itself banded. Publish that as a floor, and say that every share of the market is therefore an upper bound.
7. Do not use posted_date as an age, and do not count descriptions that name AI. Both were tried and carry no signal here (postings are refreshed, and descriptions repeat boilerplate).
8. Outputs. Write data/employer-postings.json with "unit": "jobs", the snapshot date, one row per employer and one market row, each count under a key ending in _count. Report the balance from GET /auth/key/status before and after as the cost.
9. Chart. One bar per employer for the AI-titled share of its open postings, sorted, with the comparison employer drawn in a different color. Title it with the finding.
10. Limits. Postings are demand, not headcount, and the card does not show that any role replaced a laid-off one. Postings repeated across locations count more than once. One day, not a trend.1. Read before querying. Call GET /contract (free) and use only querySpecByEntity.job fields. A count with size 1 costs 1 API Credit. Check the balance with GET /auth/key/status (free) at the start and at the end, and stop and ask before the run passes 200 API Credits.
2. Employers. Use the list in employers.json: eleven employers with a publicly reported 2026 layoff, and two comparison employers. Match on company.name with operator in, using the names as the Platform stores them. Count Amazon as the union of its two stored names so no posting is counted twice.
3. Population. Open postings: location.country eq "United States" and is_open eq true.
4. AI-titled. A posting is AI-titled when its title matches any of: ai, ml, machine learning, artificial intelligence, llm, deep learning, applied scientist, nlp, computer vision. Read about 15 titles for one employer and check that they are AI roles.
5. Count. For each employer: open postings, AI-titled postings, and six title families (engineering, sales, support, people, marketing, finance) as in the queries folder. For the market: the AI-titled count for the whole US.
6. Mind the band. A count above the threshold comes back as a string such as "100000+". Do not add it as a number. Count the US by state (the 50 states, DC and Puerto Rico), add the exact state counts, and add the threshold for each state that is itself banded. Publish that as a floor, and say that every share of the market is therefore an upper bound.
7. Do not use posted_date as an age, and do not count descriptions that name AI. Both were tried and carry no signal here (postings are refreshed, and descriptions repeat boilerplate).
8. Outputs. Write data/employer-postings.json with "unit": "jobs", the snapshot date, one row per employer and one market row, each count under a key ending in _count. Report the balance from GET /auth/key/status before and after as the cost.
9. Chart. One bar per employer for the AI-titled share of its open postings, sorted, with the comparison employer drawn in a different color. Title it with the finding.
10. Limits. Postings are demand, not headcount, and the card does not show that any role replaced a laid-off one. Postings repeated across locations count more than once. One day, not a trend.1. Read before querying. Call GET /contract (free) and use only querySpecByEntity.job fields. A count with size 1 costs 1 API Credit. Check the balance with GET /auth/key/status (free) at the start and at the end, and stop and ask before the run passes 200 API Credits.
2. Employers. Use the list in employers.json: eleven employers with a publicly reported 2026 layoff, and two comparison employers. Match on company.name with operator in, using the names as the Platform stores them. Count Amazon as the union of its two stored names so no posting is counted twice.
3. Population. Open postings: location.country eq "United States" and is_open eq true.
4. AI-titled. A posting is AI-titled when its title matches any of: ai, ml, machine learning, artificial intelligence, llm, deep learning, applied scientist, nlp, computer vision. Read about 15 titles for one employer and check that they are AI roles.
5. Count. For each employer: open postings, AI-titled postings, and six title families (engineering, sales, support, people, marketing, finance) as in the queries folder. For the market: the AI-titled count for the whole US.
6. Mind the band. A count above the threshold comes back as a string such as "100000+". Do not add it as a number. Count the US by state (the 50 states, DC and Puerto Rico), add the exact state counts, and add the threshold for each state that is itself banded. Publish that as a floor, and say that every share of the market is therefore an upper bound.
7. Do not use posted_date as an age, and do not count descriptions that name AI. Both were tried and carry no signal here (postings are refreshed, and descriptions repeat boilerplate).
8. Outputs. Write data/employer-postings.json with "unit": "jobs", the snapshot date, one row per employer and one market row, each count under a key ending in _count. Report the balance from GET /auth/key/status before and after as the cost.
9. Chart. One bar per employer for the AI-titled share of its open postings, sorted, with the comparison employer drawn in a different color. Title it with the finding.
10. Limits. Postings are demand, not headcount, and the card does not show that any role replaced a laid-off one. Postings repeated across locations count more than once. One day, not a trend.
What you get
The aggregate files and the chart, a note on what the audits found and what they changed, and the API Credits the run spent, read from the balance before and after.
A call that returns 402 insufficient_quota means the key works and the balance is empty.
Reproduce the numbers
165 API Credits$5.50More than the 100 free API Credits
A short standard-library Python script sends the committed queries as counts and writes the aggregate files this page is built from. It needs Python and METIX_KEY set in the shell (step 1 of the agent path); your agent can run these lines for you as well.
curl -fsSL https://platform.metix.ai/casebook/source/ai-roles-at-employers-with-layoffs-2026.tar.gz | tar xz
cd ai-roles-at-employers-with-layoffs-2026
: "${METIX_KEY:?set METIX_KEY first}" && python3 cases/ai-roles-at-employers-with-layoffs-2026/fetch.pyWhat you get
data/*.json and data/receipt.json. Run git diff cases/ai-roles-at-employers-with-layoffs-2026/data to see what moved: the numbers should match, apart from what changed in the data since the snapshot.
Adapt it
Costs what your version reads. Write your own ceiling into step 1 of the prompt.
The prompt is the case. Change the parts in this table and your agent answers your question instead, with the same checks and the same way of reporting cost.
| To change | Edit | For example |
|---|---|---|
| The employers | Step 2 | Add the next company to announce layoffs |
| The role | Step 4 | Count data engineering titles instead of AI titles |
| The market | Steps 3 and 6 | Restrict to California, which is banded and needs a finer split |
What to ask before running it
- Which employers, and what public source says they cut staff?
- Which words make a title an AI role?
- The US, or one state?
- How many API Credits may the run spend before it stops?
Method and limits
How the population was defined, counted, and checked, and what the numbers cannot show.
Population
Open US postings on the Metix AI Platform on October 27, 2026. Employers are matched on the company name the Platform stores, listed in employers.json. The eleven are employers with a publicly reported 2026 layoff and a large visible posting count; NVIDIA is the comparison. Intuit is mostly tax and expert roles: 7,337 of its 8,425 postings, so its share is low by mix. Apple stays in the data file but is not used: it has 899 visible open US postings and 1 with an AI role in the title, which points to thin coverage rather than a hiring pattern.
AI-titled
A title with any of these words: ai, ml, machine learning, artificial intelligence, llm, deep learning, applied scientist, nlp, computer vision. The conditions are in queries/.
Counting
All counts, one API Credit each, and no posting is read; the last replay cost 165 API Credits. The US total is above the Platform's banding threshold, so it is the sum of the state counts with the threshold used for each of the 11 states that are themselves banded: a floor.
Limits
Postings are demand, not headcount, and they do not show that any role replaced a laid-off one. A posting repeated across locations counts more than once. Employers are matched on exact names, so counts are floors. This is one day, not a trend.
The last reproduction
What reproducing this case cost the last time the script ran, read from the Platform's own balance before and after.
- Ran on
- 2026-10-27
- Calls
- 158
- Search results
- 153
- Records read
- 0
- API Credits
- 165
Search results are IDs returned by searches, one per count query and one per match on a full search. Records are postings or profiles read in full: this case reads none.
Making this case cost about 700 API Credits more: the agent's audits, trial queries, and earlier runs that the published replay replaced. You do not pay that again.