Modal growth playbook

Each Monday, give every account manager a list built from usage data: accounts whose spend grew or fell more than 30% week on week, accounts near their container or GPU concurrency limits, accounts that started a new product (Sandboxes, training, inference) for the first time, and accounts whose top workload moved to a different GPU type. Each line comes with a suggested reason to reach out.

Expansion in usage-based infrastructure follows workloads; catching a new workload or a limit early is the difference between growing with the customer and the customer splitting work to another cloud. The GTM engineering team already builds data systems and account-research to do this.

1. Spend up 30%+ week on week: offer an architecture review and check concurrency limits.
2. Within 20% of container or GPU concurrency limit: offer a limit increase before it bites.
3. First use of a new product (Sandboxes, training, inference): share the matching guide and a customer example.
4. Spend down 30%+: ask what changed; offer help with performance or a migration blocker.
5. Top workload moved to a new GPU type: share the pricing and performance notes for that GPU.

• Too many alerts

  Start with the top 50 accounts and two signals, then widen.

• Customers dislike being watched

  Lead with help (limits, cost, architecture), never with "we saw your usage drop".

• HiringAccount Manager (posted September 16, 2026): "We're hiring the first Account Managers at Modal... This function does not exist yet. There is no playbook, no territory map, no established motion." A revenue target "measured on retention and expansion"; customers are "engineers running GPU workloads, inference, and batch jobs in production"; experience at companies "consumption in nature" wanted.Source
• HiringGTM Engineer, Data Platform (posted October 13, 2026): build "an org map and people graph connecting companies, labs, teams, and individuals to identify researchers, engineers, and early-stage startups, including people beyond our existing users and CRM"; the team already builds "agents for account research". A Revenue Operations hire owns lead routing, attribution and territories.Source
• Your websiteSeries D post (October 14, 2026): usage has tripled since the Series C five months earlier; Sandboxes is "a $300M run rate product just on its own", with close to 3 billion sandboxes started and scaling to a million concurrent; the capacity pool spans "20+ clouds"; recent launches include multi-node training, Modal servers, Endpoint Candidates and virtual machines; Modal is "moving up the stack" into RL and post-training. Customers named include DoorDash, Cognition, Ramp, Linear, Hebbia, Runway, Physical Intelligence and Chai Discovery.Source
• Your websitePricing is per second (for example Nvidia B300 at $0.001972/s and H100 at $0.001097/s); Starter is $0 plus compute with $30 a month of free credits; Team is $250 a month plus compute with $100 of credits; Enterprise on request.Source

Remove or complete the robots.txt task and keep internal comments out of public files; drop /c-function from the sitemap (or redirect it to the Series D post); add canonical tags to the partners, startups and academics pages; fix "over 10,000+ teams" on the customers page and settle that figure against "tens of thousands of developers" on the partners page.

The Growth Engineer owns the site "as high-quality engineering systems"; a newly formed brand team is defining how Modal looks. These are one-hour fixes.

• Removing the group too early

  Check Search Console for the old URLs first.

• Your websiteChecked October 18, 2026: robots.txt still carries a public engineering comment, "TODO: delete this group on 2026-09-15, once the stale dashboard URLs have dropped out of Google and Bing". The sitemap lists /c-function, a "Modal — Series C" page with RSVP links for summer 2026 parties in San Francisco, New York and Stockholm. The partners, startups and academics pages have no canonical tag.Source
• Your websitePrograms: startup credits for companies that have "raised any amount of VC funding from our VC partner network" or more than $1M from any fund, one grant per tier; academic credits "up to $10k"; partner programs for integration partners, AI model partners ("day-zero deployment partner"), implementation partners (agencies and consultancies) and education partners. The customers page says Modal powers "over 10,000+ teams"; the partners page says "Tens of thousands of developers".Source
• HiringGrowth and docs roles: a Growth Engineer (posted May 18, 2026) owns the marketing site, docs, landing pages and microsites like the GPU Glossary and LLM Engine Advisor, including "agent-targeted, self-improving documentation"; a Documentation Engineer (September 21, 2026) because "docs [are] becoming even more important as agents increasingly deploy and operate Modal Apps".Source

Find the largest element on mobile (likely the hero or logo carousel), serve it smaller and earlier, defer the animated sections, and set a performance budget in CI so new campaigns cannot regress it.

The Growth Engineer is told to ship pages "without sacrificing performance"; a budget makes that automatic as the brand team adds richer work.

• Field data lags

  Track lab numbers weekly and field data monthly.

• programmable.infoChrome UX Report: mobile load (LCP) 3.07 s at the 75th percentile, 14% slower than a year ago. 709 URLs in the sitemap: 348 docs pages, 146 blog posts, 86 GPU Glossary pages, 54 model library pages. A Vanta trust center lists SOC 2 and HIPAA. GitHub: modal-examples has 1,282 stars; 166 Hacker News mentions.Source
• HiringGrowth and docs roles: a Growth Engineer (posted May 18, 2026) owns the marketing site, docs, landing pages and microsites like the GPU Glossary and LLM Engine Advisor, including "agent-targeted, self-improving documentation"; a Documentation Engineer (September 21, 2026) because "docs [are] becoming even more important as agents increasingly deploy and operate Modal Apps".Source
• HiringBrand (posted September 2, 2026): Modal's "newly formed Brand team"; "the brand that carries that story is still taking shape"; campaigns "from paid social to large-format outdoor", and editorial projects like the GPU Glossary "and forthcoming publications".Source

Set a standing process: when a major open model is released, publish a library page and a one-command deploy example within 24 hours, with measured latency and cost on Modal hardware, and post it where developers look (X, Reddit, Hacker News, the Slack community). Partner with the model makers where possible.

Speed is the content strategy here; Modal's own infrastructure makes a same-day benchmark possible where others need days.

• Benchmarks get disputed

  Publish the code and settings so anyone can reproduce them.

• Your websitePrograms: startup credits for companies that have "raised any amount of VC funding from our VC partner network" or more than $1M from any fund, one grant per tier; academic credits "up to $10k"; partner programs for integration partners, AI model partners ("day-zero deployment partner"), implementation partners (agencies and consultancies) and education partners. The customers page says Modal powers "over 10,000+ teams"; the partners page says "Tens of thousands of developers".Source
• HiringDeveloper Relations Engineer: "Modal builds AI infrastructure products that developers love. That's how we grew so quickly, and why word of mouth remains one of our most important channels today"; technical content, talks, community on X, LinkedIn, Reddit and Slack, and "joint marketing activities with other developer-focused companies".Source
• programmable.infoChrome UX Report: mobile load (LCP) 3.07 s at the 75th percentile, 14% slower than a year ago. 709 URLs in the sitemap: 348 docs pages, 146 blog posts, 86 GPU Glossary pages, 54 model library pages. A Vanta trust center lists SOC 2 and HIPAA. GitHub: modal-examples has 1,282 stars; 166 Hacker News mentions.Source
• Your websiteSeries D post (October 14, 2026): usage has tripled since the Series C five months earlier; Sandboxes is "a $300M run rate product just on its own", with close to 3 billion sandboxes started and scaling to a million concurrent; the capacity pool spans "20+ clouds"; recent launches include multi-node training, Modal servers, Endpoint Candidates and virtual machines; Modal is "moving up the stack" into RL and post-training. Customers named include DoorDash, Cognition, Ramp, Linear, Hebbia, Runway, Physical Intelligence and Chai Discovery.Source

List the partner funds on the startups page; add an accelerator and hackathon track (smaller credits, verified through the program); and route applicants who do not qualify to the free $30 a month tier with a follow-up when they raise.

The field marketer is asked to build "co-marketing programs with VCs, accelerators, and ecosystem partners"; a clear path for early founders is where word of mouth starts.

• Credit abuse

  Verify through the program and cap per company.

• Your websitePrograms: startup credits for companies that have "raised any amount of VC funding from our VC partner network" or more than $1M from any fund, one grant per tier; academic credits "up to $10k"; partner programs for integration partners, AI model partners ("day-zero deployment partner"), implementation partners (agencies and consultancies) and education partners. The customers page says Modal powers "over 10,000+ teams"; the partners page says "Tens of thousands of developers".Source
• HiringField Marketing Manager (posted October 5, 2026, San Francisco): account-based programs for named accounts; "Decide which industry conferences and sponsorships are worth investing in"; executive dinners, roundtables and technical workshops with AEs; "Build co-marketing programs with VCs, accelerators, and ecosystem partners to reach their portfolio companies".Source
• Your websitePricing is per second (for example Nvidia B300 at $0.001972/s and H100 at $0.001097/s); Starter is $0 plus compute with $30 a month of free credits; Team is $250 a month plus compute with $100 of credits; Enterprise on request.Source

Build a Sandboxes hub: what it is for (coding agents, RL rollouts, untrusted code), the scale numbers, pricing examples, comparison pages against the sandbox specialists, and a migration guide with code. Make it the landing page for sandbox search and ads.

Buyers of agent sandboxes search for sandboxes, not for "serverless AI infrastructure". A focused page captures that demand and shows a scale the specialists cannot.

• Comparisons date quickly

  Date every claim and review quarterly.

• Your websiteSeries D post (October 14, 2026): usage has tripled since the Series C five months earlier; Sandboxes is "a $300M run rate product just on its own", with close to 3 billion sandboxes started and scaling to a million concurrent; the capacity pool spans "20+ clouds"; recent launches include multi-node training, Modal servers, Endpoint Candidates and virtual machines; Modal is "moving up the stack" into RL and post-training. Customers named include DoorDash, Cognition, Ramp, Linear, Hebbia, Runway, Physical Intelligence and Chai Discovery.Source
• Competitor researchFrom rivals' homepages: Baseten ("the fastest, most reliable inference platform"), Together AI ("The AI Native Cloud"), RunPod ("The AI Developer Cloud"), Fireworks ("training and inference platform"), and for sandboxes E2B ("The Enterprise AI Agent Cloud... 100K+ teams, 2B+ sandboxes started") and Daytona ("Secure Infrastructure for Running AI-Generated Code", "90ms environment creation"). The hyperscalers (AWS, Google Cloud, Azure) are the incumbents Modal's solutions architects migrate customers from.See the research Source
• Your websitePricing is per second (for example Nvidia B300 at $0.001972/s and H100 at $0.001097/s); Starter is $0 plus compute with $30 a month of free credits; Team is $250 a month plus compute with $100 of credits; Enterprise on request.Source

Run monthly "GPU office hours" for one fund's portfolio at a time: a 30-minute talk on serving or fine-tuning, then 1:1 slots with Modal engineers, credits for attendees, and a follow-up from the AE for the most promising teams.

It reaches teams at the moment they choose infrastructure, with the fund's endorsement, and gives AEs warm leads.

• Engineer time

  Rotate FDEs and cap 1:1s.

• HiringField Marketing Manager (posted October 5, 2026, San Francisco): account-based programs for named accounts; "Decide which industry conferences and sponsorships are worth investing in"; executive dinners, roundtables and technical workshops with AEs; "Build co-marketing programs with VCs, accelerators, and ecosystem partners to reach their portfolio companies".Source
• Your websitePrograms: startup credits for companies that have "raised any amount of VC funding from our VC partner network" or more than $1M from any fund, one grant per tier; academic credits "up to $10k"; partner programs for integration partners, AI model partners ("day-zero deployment partner"), implementation partners (agencies and consultancies) and education partners. The customers page says Modal powers "over 10,000+ teams"; the partners page says "Tens of thousands of developers".Source
• HiringThe shared "About us" text on Modal's job posts: "We've passed $1B in annualized run rate, up 15x from a year ago. In October 2026 we raised a $750M Series D led by Accel at a $15.75B valuation." Named customers: Ramp, Cognition, DoorDash and Suno; the ML forward deployed team works with Suno, Lovable, Cognition and Meta. 43 open roles across New York, San Francisco, Stockholm and London.Source

Offer a structured two-week assessment: the prospect shares workload shapes (GPU types, utilization, startup times, idle time); the SA returns an architecture, a migration plan and a cost and latency comparison. It becomes the field marketer's and AE's lead offer for enterprise accounts.

Idle reserved GPUs are a common pain on the hyperscalers; per-second pricing with no idle cost is Modal's clearest argument, and an assessment makes it specific.

• Comparisons look biased

  Use the prospect's own numbers and show assumptions.

• HiringSolutions Architect (posted February 20, 2026): influence "7-8 figure infrastructure decisions"; "Architect end-to-end migration strategies from AWS, GCP, or Azure to Modal's serverless infrastructure"; "This role is not demo support."Source
• Your websitePricing is per second (for example Nvidia B300 at $0.001972/s and H100 at $0.001097/s); Starter is $0 plus compute with $30 a month of free credits; Team is $250 a month plus compute with $100 of credits; Enterprise on request.Source
• HiringAccount Executive, Enterprise (New York): "accelerate Modal's growth with the world's most innovative AI companies"; run proofs of concept and pilots with solutions and engineering; expand existing accounts; "contribute to the foundation of Modal's sales process, culture, and playbooks". A Regional Director post (July 30, 2026; $400K) describes a Manager of Enterprise Sales who builds playbooks, qualification frameworks and "top-of-funnel programs".Source
• Competitor researchFrom rivals' homepages: Baseten ("the fastest, most reliable inference platform"), Together AI ("The AI Native Cloud"), RunPod ("The AI Developer Cloud"), Fireworks ("training and inference platform"), and for sandboxes E2B ("The Enterprise AI Agent Cloud... 100K+ teams, 2B+ sandboxes started") and Daytona ("Secure Infrastructure for Running AI-Generated Code", "90ms environment creation"). The hyperscalers (AWS, Google Cloud, Azure) are the incumbents Modal's solutions architects migrate customers from.See the research Source

Use the graph to flag when an existing Modal user, an academic credit recipient or a well-known researcher changes company or founds one. Send the AE or DevRel lead a note with their history on Modal and an offer of credits for the new company.

They already know the product. A new AI company led by a past Modal user is the warmest lead there is.

• Feels intrusive

  Use public information only and lead with a helpful offer.

• HiringGTM Engineer, Data Platform (posted October 13, 2026): build "an org map and people graph connecting companies, labs, teams, and individuals to identify researchers, engineers, and early-stage startups, including people beyond our existing users and CRM"; the team already builds "agents for account research". A Revenue Operations hire owns lead routing, attribution and territories.Source
• Your websitePrograms: startup credits for companies that have "raised any amount of VC funding from our VC partner network" or more than $1M from any fund, one grant per tier; academic credits "up to $10k"; partner programs for integration partners, AI model partners ("day-zero deployment partner"), implementation partners (agencies and consultancies) and education partners. The customers page says Modal powers "over 10,000+ teams"; the partners page says "Tens of thousands of developers".Source
• HiringDeveloper Relations Engineer: "Modal builds AI infrastructure products that developers love. That's how we grew so quickly, and why word of mouth remains one of our most important channels today"; technical content, talks, community on X, LinkedIn, Reddit and Slack, and "joint marketing activities with other developer-focused companies".Source

Organise customer stories by workload (inference, training and RL, sandboxes for coding agents, batch, audio, biology, robotics), each with a number (latency, cost, time to production) and a short architecture diagram. Ask the Series D logos for one each.

Proof by workload is what the solutions architect needs in a 7-figure evaluation, and engineers trust architecture over adjectives.

• Customers protect architecture details

  Offer to anonymise numbers or show a simplified diagram.

• Your websiteSeries D post (October 14, 2026): usage has tripled since the Series C five months earlier; Sandboxes is "a $300M run rate product just on its own", with close to 3 billion sandboxes started and scaling to a million concurrent; the capacity pool spans "20+ clouds"; recent launches include multi-node training, Modal servers, Endpoint Candidates and virtual machines; Modal is "moving up the stack" into RL and post-training. Customers named include DoorDash, Cognition, Ramp, Linear, Hebbia, Runway, Physical Intelligence and Chai Discovery.Source
• Your websitePrograms: startup credits for companies that have "raised any amount of VC funding from our VC partner network" or more than $1M from any fund, one grant per tier; academic credits "up to $10k"; partner programs for integration partners, AI model partners ("day-zero deployment partner"), implementation partners (agencies and consultancies) and education partners. The customers page says Modal powers "over 10,000+ teams"; the partners page says "Tens of thousands of developers".Source
• HiringAccount Executive, Enterprise (New York): "accelerate Modal's growth with the world's most innovative AI companies"; run proofs of concept and pilots with solutions and engineering; expand existing accounts; "contribute to the foundation of Modal's sales process, culture, and playbooks". A Regional Director post (July 30, 2026; $400K) describes a Manager of Enterprise Sales who builds playbooks, qualification frameworks and "top-of-funnel programs".Source
• HiringSolutions Architect (posted February 20, 2026): influence "7-8 figure infrastructure decisions"; "Architect end-to-end migration strategies from AWS, GCP, or Azure to Modal's serverless infrastructure"; "This role is not demo support."Source

Build an internal benchmark: a set of common tasks (deploy an LLM, run a batch job, start a sandbox) given to the major coding agents with only Modal's docs. Track success rate, fix the docs where agents fail, and publish the benchmark and results.

Agents are becoming the ones who choose and use infrastructure; being the platform they succeed on is a growth channel no ad can buy.

• Agents change often

  Rerun monthly and report the trend.

• HiringGrowth and docs roles: a Growth Engineer (posted May 18, 2026) owns the marketing site, docs, landing pages and microsites like the GPU Glossary and LLM Engine Advisor, including "agent-targeted, self-improving documentation"; a Documentation Engineer (September 21, 2026) because "docs [are] becoming even more important as agents increasingly deploy and operate Modal Apps".Source
• Your websiteSeries D post (October 14, 2026): usage has tripled since the Series C five months earlier; Sandboxes is "a $300M run rate product just on its own", with close to 3 billion sandboxes started and scaling to a million concurrent; the capacity pool spans "20+ clouds"; recent launches include multi-node training, Modal servers, Endpoint Candidates and virtual machines; Modal is "moving up the stack" into RL and post-training. Customers named include DoorDash, Cognition, Ramp, Linear, Hebbia, Runway, Physical Intelligence and Chai Discovery.Source
• programmable.infoChrome UX Report: mobile load (LCP) 3.07 s at the 75th percentile, 14% slower than a year ago. 709 URLs in the sitemap: 348 docs pages, 146 blog posts, 86 GPU Glossary pages, 54 model library pages. A Vanta trust center lists SOC 2 and HIPAA. GitHub: modal-examples has 1,282 stars; 166 Hacker News mentions.Source

Add an enterprise page: security and compliance (trust center link), deployment options, support levels, procurement (contract and data processing terms), and a route straight to a solutions architect for technical evaluations. Track which assets buyers open.

Large buyers move faster when the paperwork is ready; the RevOps hire owns lead routing that should send these buyers to the right person on day one.

• Over-publishing terms

  Share summaries publicly, full terms on request.

• programmable.infoChrome UX Report: mobile load (LCP) 3.07 s at the 75th percentile, 14% slower than a year ago. 709 URLs in the sitemap: 348 docs pages, 146 blog posts, 86 GPU Glossary pages, 54 model library pages. A Vanta trust center lists SOC 2 and HIPAA. GitHub: modal-examples has 1,282 stars; 166 Hacker News mentions.Source
• Your websitePricing is per second (for example Nvidia B300 at $0.001972/s and H100 at $0.001097/s); Starter is $0 plus compute with $30 a month of free credits; Team is $250 a month plus compute with $100 of credits; Enterprise on request.Source
• HiringSolutions Architect (posted February 20, 2026): influence "7-8 figure infrastructure decisions"; "Architect end-to-end migration strategies from AWS, GCP, or Azure to Modal's serverless infrastructure"; "This role is not demo support."Source
• HiringAccount Executive, Enterprise (New York): "accelerate Modal's growth with the world's most innovative AI companies"; run proofs of concept and pilots with solutions and engineering; expand existing accounts; "contribute to the foundation of Modal's sales process, culture, and playbooks". A Regional Director post (July 30, 2026; $400K) describes a Manager of Enterprise Sales who builds playbooks, qualification frameworks and "top-of-funnel programs".Source

Launch a quarterly technical publication (online first, a printed edition for events): deep pieces on GPU performance, inference engineering and post-training, written by Modal engineers and guests, with the glossary as the reference layer. Offer an email subscription on every glossary page.

Owned audience compounds; it gives field marketing something to hand out and DevRel something to promote every quarter.

• Engineer time for writing

  Interview engineers and edit, rather than asking them to write.

• HiringBrand (posted September 2, 2026): Modal's "newly formed Brand team"; "the brand that carries that story is still taking shape"; campaigns "from paid social to large-format outdoor", and editorial projects like the GPU Glossary "and forthcoming publications".Source
• programmable.infoChrome UX Report: mobile load (LCP) 3.07 s at the 75th percentile, 14% slower than a year ago. 709 URLs in the sitemap: 348 docs pages, 146 blog posts, 86 GPU Glossary pages, 54 model library pages. A Vanta trust center lists SOC 2 and HIPAA. GitHub: modal-examples has 1,282 stars; 166 Hacker News mentions.Source
• HiringDeveloper Relations Engineer: "Modal builds AI infrastructure products that developers love. That's how we grew so quickly, and why word of mouth remains one of our most important channels today"; technical content, talks, community on X, LinkedIn, Reddit and Slack, and "joint marketing activities with other developer-focused companies".Source

Keep a light relationship with every academic credit recipient: a yearly check-in, invitations to events, and an automatic startup credit offer when they found or join a company. Highlight published papers that used Modal.

It costs little and catches founders before they choose infrastructure; it pairs with the people graph (idea 9).

• Academics move slowly

  Treat it as a long-term pipeline, measured yearly.

• Your websitePrograms: startup credits for companies that have "raised any amount of VC funding from our VC partner network" or more than $1M from any fund, one grant per tier; academic credits "up to $10k"; partner programs for integration partners, AI model partners ("day-zero deployment partner"), implementation partners (agencies and consultancies) and education partners. The customers page says Modal powers "over 10,000+ teams"; the partners page says "Tens of thousands of developers".Source
• HiringGTM Engineer, Data Platform (posted October 13, 2026): build "an org map and people graph connecting companies, labs, teams, and individuals to identify researchers, engineers, and early-stage startups, including people beyond our existing users and CRM"; the team already builds "agents for account research". A Revenue Operations hire owns lead routing, attribution and territories.Source
• HiringDeveloper Relations Engineer: "Modal builds AI infrastructure products that developers love. That's how we grew so quickly, and why word of mouth remains one of our most important channels today"; technical content, talks, community on X, LinkedIn, Reddit and Slack, and "joint marketing activities with other developer-focused companies".Source

A monthly meetup in each city: two short technical talks (one from a customer), demos, and time for engineers to meet. Record the talks for the blog and YouTube. Invite target accounts' engineers personally.

In-person events build the word of mouth Modal credits for its growth, and give AEs and DevRel a recurring place to meet target accounts.

• Attendance drops after novelty

  Keep talks sharp and rotate topics by workload.

• Your websiteChecked October 18, 2026: robots.txt still carries a public engineering comment, "TODO: delete this group on 2026-09-15, once the stale dashboard URLs have dropped out of Google and Bing". The sitemap lists /c-function, a "Modal — Series C" page with RSVP links for summer 2026 parties in San Francisco, New York and Stockholm. The partners, startups and academics pages have no canonical tag.Source
• HiringField Marketing Manager (posted October 5, 2026, San Francisco): account-based programs for named accounts; "Decide which industry conferences and sponsorships are worth investing in"; executive dinners, roundtables and technical workshops with AEs; "Build co-marketing programs with VCs, accelerators, and ecosystem partners to reach their portfolio companies".Source
• HiringDeveloper Relations Engineer: "Modal builds AI infrastructure products that developers love. That's how we grew so quickly, and why word of mouth remains one of our most important channels today"; technical content, talks, community on X, LinkedIn, Reddit and Slack, and "joint marketing activities with other developer-focused companies".Source
• HiringThe shared "About us" text on Modal's job posts: "We've passed $1B in annualized run rate, up 15x from a year ago. In October 2026 we raised a $750M Series D led by Accel at a $15.75B valuation." Named customers: Ramp, Cognition, DoorDash and Suno; the ML forward deployed team works with Suno, Lovable, Cognition and Meta. 43 open roles across New York, San Francisco, Stockholm and London.Source

Formalise the program: training and certification, a partner directory, deal registration and referral fees, co-selling with AEs on enterprise accounts, and credits for partner prototypes.

Agencies bring repeatable enterprise deployments without Modal adding headcount; certified partners also take load off forward deployed engineers.

• Weak partners hurt reputation

  Certify on real deployments, not just an exam.

• Your websitePrograms: startup credits for companies that have "raised any amount of VC funding from our VC partner network" or more than $1M from any fund, one grant per tier; academic credits "up to $10k"; partner programs for integration partners, AI model partners ("day-zero deployment partner"), implementation partners (agencies and consultancies) and education partners. The customers page says Modal powers "over 10,000+ teams"; the partners page says "Tens of thousands of developers".Source
• HiringAccount Executive, Enterprise (New York): "accelerate Modal's growth with the world's most innovative AI companies"; run proofs of concept and pilots with solutions and engineering; expand existing accounts; "contribute to the foundation of Modal's sales process, culture, and playbooks". A Regional Director post (July 30, 2026; $400K) describes a Manager of Enterprise Sales who builds playbooks, qualification frameworks and "top-of-funnel programs".Source
• HiringThe shared "About us" text on Modal's job posts: "We've passed $1B in annualized run rate, up 15x from a year ago. In October 2026 we raised a $750M Series D led by Accel at a $15.75B valuation." Named customers: Ramp, Cognition, DoorDash and Suno; the ML forward deployed team works with Suno, Lovable, Cognition and Meta. 43 open roles across New York, San Francisco, Stockholm and London.Source

For each of two or three industries: a dedicated page with customer proof and reference architectures, a forward deployed engineer who specialises in it, a presence at that industry's main events, and a quarterly industry briefing.

Specialist buyers trust specialist vendors; a dedicated track lets Modal win the next robotics or biology company before a general cloud does.

• Spreads the team thin

  Start with one industry and add only when it works.

• Your websiteSeries D post (October 14, 2026): usage has tripled since the Series C five months earlier; Sandboxes is "a $300M run rate product just on its own", with close to 3 billion sandboxes started and scaling to a million concurrent; the capacity pool spans "20+ clouds"; recent launches include multi-node training, Modal servers, Endpoint Candidates and virtual machines; Modal is "moving up the stack" into RL and post-training. Customers named include DoorDash, Cognition, Ramp, Linear, Hebbia, Runway, Physical Intelligence and Chai Discovery.Source
• HiringThe shared "About us" text on Modal's job posts: "We've passed $1B in annualized run rate, up 15x from a year ago. In October 2026 we raised a $750M Series D led by Accel at a $15.75B valuation." Named customers: Ramp, Cognition, DoorDash and Suno; the ML forward deployed team works with Suno, Lovable, Cognition and Meta. 43 open roles across New York, San Francisco, Stockholm and London.Source
• HiringField Marketing Manager (posted October 5, 2026, San Francisco): account-based programs for named accounts; "Decide which industry conferences and sponsorships are worth investing in"; executive dinners, roundtables and technical workshops with AEs; "Build co-marketing programs with VCs, accelerators, and ecosystem partners to reach their portfolio companies".Source

Publish a quarterly benchmark of popular open models across engines and GPUs, with all code and settings open so anyone can rerun them, including on other providers. Invite rivals and the community to submit runs.

Open methods turn a claim into a resource developers cite, and Modal's elasticity and pricing show in the numbers without Modal having to say so.

• Results do not always favour Modal

  Publish anyway and fix the product; credibility is the point.

• Competitor researchFrom rivals' homepages: Baseten ("the fastest, most reliable inference platform"), Together AI ("The AI Native Cloud"), RunPod ("The AI Developer Cloud"), Fireworks ("training and inference platform"), and for sandboxes E2B ("The Enterprise AI Agent Cloud... 100K+ teams, 2B+ sandboxes started") and Daytona ("Secure Infrastructure for Running AI-Generated Code", "90ms environment creation"). The hyperscalers (AWS, Google Cloud, Azure) are the incumbents Modal's solutions architects migrate customers from.See the research Source
• HiringGrowth and docs roles: a Growth Engineer (posted May 18, 2026) owns the marketing site, docs, landing pages and microsites like the GPU Glossary and LLM Engine Advisor, including "agent-targeted, self-improving documentation"; a Documentation Engineer (September 21, 2026) because "docs [are] becoming even more important as agents increasingly deploy and operate Modal Apps".Source
• Your websiteSeries D post (October 14, 2026): usage has tripled since the Series C five months earlier; Sandboxes is "a $300M run rate product just on its own", with close to 3 billion sandboxes started and scaling to a million concurrent; the capacity pool spans "20+ clouds"; recent launches include multi-node training, Modal servers, Endpoint Candidates and virtual machines; Modal is "moving up the stack" into RL and post-training. Customers named include DoorDash, Cognition, Ramp, Linear, Hebbia, Runway, Physical Intelligence and Chai Discovery.Source

Create a badge and a launch link that clones the repo into a Modal app with the right GPU, using a small config file. Offer it to research groups and open-source maintainers, starting with ones already using Modal.

It spreads through the places engineers already look, and each launch is a new user who has seen the product work.

• Abuse of free credits

  Small credits per launch, verified accounts.

• programmable.infoChrome UX Report: mobile load (LCP) 3.07 s at the 75th percentile, 14% slower than a year ago. 709 URLs in the sitemap: 348 docs pages, 146 blog posts, 86 GPU Glossary pages, 54 model library pages. A Vanta trust center lists SOC 2 and HIPAA. GitHub: modal-examples has 1,282 stars; 166 Hacker News mentions.Source
• HiringDeveloper Relations Engineer: "Modal builds AI infrastructure products that developers love. That's how we grew so quickly, and why word of mouth remains one of our most important channels today"; technical content, talks, community on X, LinkedIn, Reddit and Slack, and "joint marketing activities with other developer-focused companies".Source
• Your websitePrograms: startup credits for companies that have "raised any amount of VC funding from our VC partner network" or more than $1M from any fund, one grant per tier; academic credits "up to $10k"; partner programs for integration partners, AI model partners ("day-zero deployment partner"), implementation partners (agencies and consultancies) and education partners. The customers page says Modal powers "over 10,000+ teams"; the partners page says "Tens of thousands of developers".Source

A small, technical outdoor campaign near AI offices in San Francisco and New York and at the venues of the main AI conferences: one line an engineer understands (for example "Stop paying for GPUs you are not using."), a short URL, measured with branded search and the URL's visits.

For a company that grew on word of mouth, a sharp outdoor campaign starts conversations and recruits as well as sells.

• Hard to measure

  Use unique URLs and city-level branded search.

• HiringBrand (posted September 2, 2026): Modal's "newly formed Brand team"; "the brand that carries that story is still taking shape"; campaigns "from paid social to large-format outdoor", and editorial projects like the GPU Glossary "and forthcoming publications".Source
• HiringThe shared "About us" text on Modal's job posts: "We've passed $1B in annualized run rate, up 15x from a year ago. In October 2026 we raised a $750M Series D led by Accel at a $15.75B valuation." Named customers: Ramp, Cognition, DoorDash and Suno; the ML forward deployed team works with Suno, Lovable, Cognition and Meta. 43 open roles across New York, San Francisco, Stockholm and London.Source

Ship and promote the pieces agents need: a clear CLI and API, an MCP server, short machine-readable docs (llms.txt), and templates the major coding agents can use. Work with agent makers to include Modal as a deployment option, and measure deployments that start from agents.

It turns the agent shift Modal is already preparing for into distribution, and it compounds with the docs benchmark (idea 11).

• Agents change fast

  Keep the integration surface simple and standard.

• HiringGrowth and docs roles: a Growth Engineer (posted May 18, 2026) owns the marketing site, docs, landing pages and microsites like the GPU Glossary and LLM Engine Advisor, including "agent-targeted, self-improving documentation"; a Documentation Engineer (September 21, 2026) because "docs [are] becoming even more important as agents increasingly deploy and operate Modal Apps".Source
• Your websiteSeries D post (October 14, 2026): usage has tripled since the Series C five months earlier; Sandboxes is "a $300M run rate product just on its own", with close to 3 billion sandboxes started and scaling to a million concurrent; the capacity pool spans "20+ clouds"; recent launches include multi-node training, Modal servers, Endpoint Candidates and virtual machines; Modal is "moving up the stack" into RL and post-training. Customers named include DoorDash, Cognition, Ramp, Linear, Hebbia, Runway, Physical Intelligence and Chai Discovery.Source
• programmable.infoChrome UX Report: mobile load (LCP) 3.07 s at the 75th percentile, 14% slower than a year ago. 709 URLs in the sitemap: 348 docs pages, 146 blog posts, 86 GPU Glossary pages, 54 model library pages. A Vanta trust center lists SOC 2 and HIPAA. GitHub: modal-examples has 1,282 stars; 166 Hacker News mentions.Source